RU2846769C2 - Method and apparatus for filtering images with adaptive multiplier coefficients - Google Patents

Abstract

FIELD: physics.

SUBSTANCE: invention relates to video and/or still image encoding techniques. Disclosed is a device and a method for filtering reconstructed images, particularly video images, using adaptive multiplicative filters.

EFFECT: efficiency of the filtering operation is increased by limiting the allowable values of filter coefficients to those having only a limited number of “ones” in the binary representation.

19 cl, 11 dwg

Description

Field of technology to which the invention relates

Embodiments of the invention relate to image processing, such as video and/or still image coding. New methods and devices are provided for filtering images using a filter having adaptive filter multiplier coefficients.

State of the art

Video coding (video encoding and decoding) is used in a wide range of digital video applications, such as digital broadcast television, video transmission over the Internet and mobile networks, real-time interactive applications such as video chat, video conferencing, DVD and Blu-ray discs, video content acquisition and editing systems, and video cameras in security applications.

Since the development of the block-based hybrid video coding approach in 1990, the H.261 standard has introduced new video coding techniques and tools that have formed the basis for new video coding standards. One of the goals of most video coding standards has been to reduce the data rate compared to its predecessor without sacrificing image quality. Additional video coding standards include MPEG-1 video, MPEG-2 video, ITU-T H.262/MPEG-2, ITU-T H.263, ITU-T H.264/MPEG-4 Part 10, Advanced Video Coding (AVC), ITU-T H.265, High Efficiency Video Coding (HEVC), and extensions such as scalability and/or three-dimensional (3D) extensions to these standards.

Fig. 1 is a block diagram illustrating an embodiment of a coding system 300, which will be described in more detail below.

Fig. 2 is a block diagram showing an exemplary structure of a video encoder in which the invention may be implemented, and which will also be described in more detail below.

In particular, the illustrated encoder 100 includes a "loop filter" 120, in which the filtering operation according to the invention can be applied. However, more generally, the filtering operation is applicable elsewhere in the codec, for example, in an interpolation filter. Even more broadly, the invention is applicable not only to video, but also to the coding of static images.

Fig. 3 is a block diagram showing an exemplary structure of a video decoder in which the invention can be implemented, which will also be described in more detail below. In particular, the invention is applicable, for example, to loop filter 220.

Some background information about adaptive filtering will be summarized below.

Adaptive filtering for video coding minimizes the mean square error between original and decoded samples using an adaptive Wiener filter. Specifically, the proposed adaptive loop filter (ALF) is placed at the final stage of processing each image and can be considered a tool for detecting and correcting artifacts from previous stages. Suitable filter coefficients are determined by the encoder and explicitly communicated to the decoder.

For a general introduction to adaptive filtering, see “Adaptive Loop Filtering for Video Coding” by Chia-Yang Tsai, Ching-Yeh Chen, Tomoo Yamakage, In Suk Chong, Yu-Wen Huang, Chin-Ming Fu, Takayuki Itoh, Takayuki Watanabe, Takeshi Chujoh, Marta Karczewicz, and Shaw-Min Lei, published in IEEE Journal of Selected Topics in Signal Processing (Volume: 7, Issue: 6, Dec. 2013).

The description provided in the above document describes a specific implementation of a filtering operation with adaptive filter coefficients. The general operating principles can be described as follows.

Typically, the filtration equation looks like this:

Here R(i, j) is the sample in the image frame before filtering by coordinate (i, j).

R’(i, j) is a sample in the image frame after filtering, f (k, l) are the filter coefficients.

Figure 4 shows an example of a filter kernel. In this example, C20 is the central coordinate of the filter kernel (k = 0, l = 0), and L is equal to 8.

In this example, the filter kernel is symmetrical about the center. However, this statement is generally incorrect.

In the case of using integer arithmetic, the filtration equation can be written as:

Here, N is the output signal bit shift number, that is, the output signal is divided by the normalization factor. In particular, N can be predetermined. The "offset" is a scalar value used to compensate for losses in integer arithmetic. In the case of a bit shift by N, the offset can be 2 ^(N-1) . In the above equation, the filter coefficients f(k,l) can only have integer values, not fractional ones. Implementing the filter equation according to integer arithmetic is important to ensure accurate implementation in hardware. The right shift operation ">>N" has the effect of dividing by ^2N followed by rounding down.

Typically (but not necessarily), the following equation holds if no change in the average illuminance level is required.

In the encoder, the filter coefficients are estimated by minimizing the expected error value between the original and filtered pixels:

In the above equation, O(i,j) denotes the sample of the original image.

Figure 5 shows some typical example filter shapes for adaptive filters. The figure on the left shows a 5x5 diamond filter (a 13-tap filter with 7 unique coefficients), the middle figure shows a 7x7 diamond filter (a 25-tap filter with 13 unique coefficients), and the figure on the right shows a 9x9 diamond filter (a 41-tap filter with 21 unique coefficients).

The term "adaptive" filtering refers to the fact that the filtering process can be adjusted by the encoder. This applies, for example, to the filter shape, filter size, number of filter coefficients, and filter coefficient values. This data, also known as "filter hints," is passed to the decoder.

Adaptive filtering addresses the following technical challenges when applied to filtering implementations that involve multiplication, i.e., in which the filter coefficients are so-called multiplicative coefficients or multiplier coefficients. In other words, the following technical challenge, which the invention seeks to address, relates to filtering with adaptive filter coefficients, in which the filter coefficients used in the multiplication operation can be individually adapted (modified). In this regard, "individually" means for each image (image, frame) and/or for each pixel and/or for each coefficient.

The technical challenge is that implementing the multiplication operation is expensive, especially in specialized hardware implementations. Applying the filter requires a relatively large number of filter multiplication operations (e.g., 41 multiplications per pixel for a 9x9 diamond-shaped filter, as shown in Fig. 4).

This is illustrated in more detail below.

Suppose we want to multiply two unsigned eight-bit integers. The filter coefficient is C and the pixel sample is A.

The multiplication process can be decomposed into 8 single-bit multiplications, each of which can be implemented as a bit shift operation in binary arithmetic, and 7 addition operations, as shown below. Therefore, approximately 1 multiplication is equivalent to 7 additions.

The problem is that the multiplication process requires a large amount of computation. Therefore, implementing it on specialized hardware is very expensive.

Here, the 8-bit unsigned filter coefficient C is shown in binary representation, where C[0] is the least significant bit of the coefficient C, and C[7] is the most significant bit. Similarly, A[7], A[6], … A[0] are the bits corresponding from the most significant bit to the least significant bit in order. The operation P = C * A in binary arithmetic is represented, and the result is displayed in the bottom row.

In the example in Fig. 4, the filter kernel includes 41 filter taps, which means that 41 multiplication operations are required to process the pixel sample.

It should be noted that the invention and the technical problem it solves are specifically related to adaptive filtering with filter multiplier coefficients. The problem does not apply to fixed filters, and in particular, to filtering operations using multiple fixed filters.

An example of using multiple fixed filters is interpolation filtering for interpolation at fractional pixel positions in inter prediction, as illustrated in Fig. 6.

Many well-known codecs employ interpolation filtering using fixed interpolation filters. Although the filter coefficients are fixed for each filter, there are multiple filters for different fractional positions (half-pixel and quarter-pixel positions in the drawing). In this example, the entire filter set adapts based on the motion vector, but the filter coefficients do not adapt individually.

In the figure, the large circles correspond to actual sample positions in the image, and the smaller circles represent fractional positions generated by applying an interpolation filtering operation. In the specific example, there are three fractional positions (left quarter pixel, half pixel, and right quarter pixel) between two actual sample positions in the image. The left side of the figure shows the interpolation filter applied to interpolate half-pixel positions. The right side of the figure illustrates the interpolation filter to be used for quarter-pixel positions. Although these filters differ from each other, each interpolation filter is fixed. As noted, the example in Fig. 6 is provided for illustrative purposes only and is not part of the invention.

The invention aims to provide an improved concept of multiplicative adaptive filtering, which can simplify the multiplication operation and reduce the cost of the multiplication operation.

Disclosure of the essence of the invention

The embodiments of the invention are defined by the features of independent claims, and additional advantageous implementations of the embodiments are defined by the features of dependent claims.

According to the first aspect of the invention, a device for filtering a set of image samples is provided using a filter with adaptive multiplier coefficients represented by integers. The device comprises a processing circuit configured to determine the value of at least one filter multiplier coefficient such that it is within a set of permissible values, such that the binary representation of the absolute value of at least one multiplier coefficient with a predetermined number L of digits includes at least one "zero," and filters the set of image samples using the filter.

According to a second aspect of the invention, a method is provided for filtering a set of image samples using a filter with adaptive multiplier coefficients represented by integers. The method comprises the step of determining the value of at least one filter multiplier coefficient such that it is within a set of permissible values, such that the binary representation of the absolute value of at least one multiplier coefficient with a predetermined number L of digits includes at least one "zero," and the step of filtering the set of image samples using the filter.

According to the present disclosure, the set of image samples may, for example, be a sample of a video signal or a still image signal. The processing circuit may be implemented using any combination of software and/or hardware. The set of permissible values may, in particular, be a predetermined set of permissible values. Generally, the invention is also applicable to sets of signal samples other than images, such as signals including audio data.

A special approach of the invention consists of limiting the values that can be accepted by the filter coefficients of the adaptive multiplier filter, thereby simplifying the multiplication operation. Specifically, the permissible values of the filter coefficients are limited so that only a limited number of "ones" are permitted within a predetermined number of binary digits for expressing absolute values. This simplifies the multiplication operations for filtering, thus making the filtering operation more efficient.

As will be shown below, the smaller the number of permissible "1s" in a predetermined total number of binary digits, the higher the efficiency of the filtering operation. For example, the greatest efficiency gain can be achieved if any value that can be assumed by the coefficient values includes only up to one "1," that is, no more than one "1."

According to embodiments, the highest absolute value of the set of allowable values is limited to a predetermined maximum value N _max .

According to embodiments, the binary representation of the absolute value of at least one multiplier coefficient includes no more than two "ones." More specifically, the binary representation of the absolute value of at least one multiplier coefficient includes no more than one "one." As noted above and as will be described in detail below, the more zeros (and therefore fewer ones) in the binary representation of the permitted coefficient values, the greater the simplification of the multiplication operation for filtering, and thus the increased processing efficiency. Thus, the most efficient case is one in which there is only one "one," while, for example, two permitted "ones" still yield a good result. Of course, what is beneficial depends largely on the details of the situation, and, particularly for large filters, three or more "ones" may also be beneficial.

Typically, a set of acceptable values applies to at least one filter multiplier coefficient.

In accordance with embodiments, the set of permissible values applies to all coefficients of the filter multiplier.

According to alternative embodiments, the multiplier coefficients are further grouped into at least two groups, and the multiplier coefficients of one of the groups are limited to a set of permissible values. The multiplier coefficients in the other group or groups may, for example, take on all values within a predetermined range or may be limited in accordance with other predetermined rules. More specifically, the multiplier coefficients of another group, for example, may take on all values within a range defined by a predetermined maximum absolute value.

According to embodiments, a set of image samples refers to a set of video image samples. More specifically, the device may be configured to individually adapt the multiplier coefficients for each image and each pixel.

According to a further specific aspect of the invention, a device is provided for encoding a current set of image samples comprising a plurality of pixels. The device comprises an encoder and decoder for reconstructing the current set and a device according to the first aspect of the invention for filtering the reconstructed set.

According to embodiments, said encoding device further comprises a processing circuit that is configured to map the values of the multiplier coefficients to binary code words and add the code words to a bit stream for transmission to a decoding device.

More specifically, the length of the codewords depends on the number of different values of the multiplier coefficient. In other words, there are as many codewords as there are possible filter coefficient values. The mapping from a codeword to a value (which is a one-to-one mapping) can be a fixed mapping or can change depending on the transmitted additional information.

According to embodiments, the processing circuit is further configured to predict the filter multiplier coefficients and determine the residual coefficients of the multiplier by comparing the actually determined values with the predicted values obtained as a result of the prediction. A mapping of the binary codewords to the residual coefficients of the multiplier is then applied. In this case, prediction control information may be additionally included in the bitstream so that the decoding device receiving the bitstream knows about the applied prediction method and can reconstruct the filter multiplier coefficients from the encoded residual coefficients of the multiplier. Alternatively, the applied prediction method may be predetermined and therefore applied in the same manner in the encoder and decoder without any transmitted additional information. Possible prediction methods may include, but are not limited to, prediction using specified filter predictors and prediction based on previously reported filter coefficients. Since the values of the residual filter coefficients, expressing the difference between the actual filter coefficient and the corresponding predicted filter coefficient, are typically smaller in absolute value than the actual coefficients, the number and hence the size of code words can be smaller, which further reduces the amount of information signaled to the decoder.

Alternatively, mapping of the multiplier coefficients onto code words for inclusion in the bit stream may be performed on the multiplier coefficients determined according to the first aspect of the invention without performing prediction processing.

According to yet another aspect of the invention, a device is provided for decoding a current encoded set of image samples comprising a plurality of pixels. The device comprises a decoder for reconstructing the current set and a device according to the first aspect of the invention for filtering the reconstructed set.

According to embodiments, the processing circuit of the device according to the first aspect of the invention is further configured to obtain multiplier coefficients from binary code words included in the received bit stream by applying a mapping operation.

In particular, the obtained multiplier coefficients may be filter coefficients to be used for filtering. Alternatively, the obtained multiplier coefficients may be residual multiplier coefficients representing the difference between the actual coefficient values and the multiplier coefficients predicted according to a prediction scheme. The prediction scheme may be indicated by prediction control information additionally included in the received bitstream. In this case, the processing circuit is further configured to determine the filter coefficient values by reconstructing them from the obtained residual multiplier coefficients and the prediction control information. Alternatively, the prediction scheme (prediction method) may be predetermined and, therefore, applied in the same manner in the encoder and decoder without any transmitted prediction control information. The processing circuit then determines the filter coefficient values by reconstructing them from the obtained residual multiplier coefficients.

According to embodiments, the determining by the processing circuit further includes performing a determination whether a determined value of at least one multiplier coefficient, obtained directly from the received bit stream by means of a mapping operation or by reconstruction from the obtained residual coefficients of the multiplier, is in a set of valid values and, if not, converting the determined value to the nearest value that is within the set of valid values.

In this way, it is guaranteed that the filter coefficients that are applied to the samples of the reconstructed image obey the rules according to the invention.

The details of one or more embodiments are set forth in the accompanying drawings and the description below. Other features, objects, and advantages will be apparent from the description, drawings, and claims.

Brief description of the drawings

The embodiments of the invention are now described in more detail with reference to the accompanying drawings, in which:

Fig. 1 is a block diagram showing an example of a video coding system capable of implementing embodiments of the invention;

Fig. 2 is a block diagram showing an example of a video encoder configured to implement embodiments of the invention;

Fig. 3 is a block diagram showing an exemplary structure of a video decoder capable of implementing embodiments of the invention;

Fig. 4 shows an example of a filter core to which the invention can be applied;

Fig. 5 shows examples of typical filter shapes for adaptive filters to which the invention can be applied;

Fig.6 shows an example of a plurality of fixed filters to be applied in interpolation filtering, as a comparative example;

Fig. 7 illustrates a specific example of the implementation of an embodiment of the invention;

Fig. 8A illustrates exemplary encoder-side processing for encoding and signaling filter coefficients;

Fig. 8B illustrates exemplary decoder-side processing for decoding and reconstructing filter coefficients;

Fig. 9 illustrates a specific example of the implementation of another embodiment of the invention;

Fig. 10 illustrates a specific example of the implementation of another embodiment of the invention and serves to illustrate the obtained effect of the invention; and

Fig. 11 illustrates a further example of a filter core to which the invention can be applied.

In the drawings, identical reference numerals refer to identical or at least functionally equivalent features.

Implementation of the invention

In the following description, reference is made to the accompanying drawings, which form a part of the present invention and which show by way of illustration specific aspects of embodiments of the invention or specific aspects in which embodiments of the invention may be used. It is understood that embodiments of the invention may be used in other aspects and include structural or logical changes not shown in the drawings. Therefore, the following detailed description should not be considered in a limiting sense, and the scope of the invention is defined by the appended claims.

For example, it is understood that the described method of the present invention can also be applied to a corresponding device or system configured to perform the method and vice versa. For example, if one or a plurality of specific method steps are described, the corresponding device may include one or a plurality of blocks, for example, functional blocks, for performing the described one or a plurality of method steps (for example, one block performing one or a plurality of steps, or a plurality of blocks, each of which performs one or more of the plurality of steps), even if such one or more blocks are not explicitly described or illustrated in the drawings. On the other hand, for example, if a specific device is described based on one or a plurality of blocks, for example, functional blocks, the corresponding method may include one step for performing the functionality of one or a plurality of blocks (for example, one step performing the functionality of one or a plurality of blocks, or a plurality of steps, each of which performs the functionality of one or more of the plurality of blocks), even if such one or a plurality of steps are not explicitly described or illustrated in the drawings. Furthermore, it should be understood that features of the various exemplary embodiments and/or aspects described herein may be combined with one another unless otherwise specifically stated.

Video coding generally refers to the processing of a sequence of images that form a video or video sequence. Instead of the term "image," the terms "frame" or "picture" may be used interchangeably in the field of video coding. Video coding consists of two parts: video encoding and video decoding. Video coding is performed at the source end, typically involving processing (e.g., by compression) of the source video images to reduce the amount of data required to represent the video images (for more efficient storage and/or transmission). Video decoding is performed at the destination end and typically involves the inverse of the encoder's processing to reconstruct the video images. Embodiments that refer to the "coding" of video images (or images in general, as will be explained later) should be understood as referring to both the "coding" and "decoding" of video images. The combination of the encoding and decoding parts is also referred to as CODEC (COding and DECoding).

In lossless video coding, the original video images can be reconstructed, meaning the reconstructed video images have the same quality as the original video images (assuming no transmission losses or other data losses during storage or transmission). In lossy video coding, additional compression is performed, such as by quantization, to reduce the amount of data representing the video images that cannot be fully reconstructed at the decoder. This means the quality of the reconstructed video images is lower or worse than the quality of the original video images.

Several video coding standards, starting with H.261, belong to the group of "hybrid lossy video codecs" (i.e., they combine spatial and temporal prediction in the sample domain and two-dimensional transform coding to apply quantization in the transform domain). Each image of a video sequence is typically divided into a set of non-overlapping blocks, and coding is usually performed at the block level. In other words, in the encoder, video is typically processed, i.e., encoded, at the block level (video block), for example, using spatial (intra-image) prediction and temporal (inter-image) prediction to generate a prediction block, subtracting the prediction block from the current block (the block currently being/to be processed) to obtain a residual block, performing a transform on the residual block, and quantizing the residual block in the transform domain to reduce the amount of data that must be transmitted (compression). In the decoder, the inverse of the encoder's processing is applied to the coded or compressed block to reconstruct the current block for presentation. In addition, the encoder duplicates the decoder's processing loop so that both will generate identical predictions (e.g., intra- and inter-predictions) and/or reconstructions for processing, i.e., encoding, subsequent blocks.

While video image processing (also known as moving image processing) and still image processing (the term "processing" includes coding) share many common concepts, technologies, or tools, in the following description, the term "image" or "frame" and the equivalent term "image data" or "frame data" are used to refer to a video image in a video sequence (as explained above) and/or a still image. This avoids unnecessary repetition and unnecessary distinctions between video images and still images. If the description pertains only to still images (or still frames), the term "still image" should be used.

In the following embodiments, the encoder 100, decoder 200 and coding system 300 are described with reference to Figs. 1-3 (prior to a more detailed description of the embodiments of the invention based on Figs. 7-9).

Fig. 1 is a conceptual or schematic block diagram illustrating an embodiment of a coding system 300, such as an image coding system 300, in which the coding system 300 comprises a source device 310 configured to provide coded data 330, such as an encoded image 330, to a destination device 320 for decoding the coded data 330.

The source device 310 comprises an encoder 100 or an encoding unit 100 and may additionally, i.e. possibly, comprise an image source 312, a pre-processing unit 314, for example an image pre-processing unit 314, and a communication interface or a communication unit 318.

The image source 312 may comprise or be any type of image capture device, such as one for capturing a real image and/or any type of imaging device, such as a computer graphics processor for generating a computer animated image, or any type of device for receiving and/or providing a real image, a computer animated image (e.g., on-screen content, a virtual reality (VR) image) and/or any combination thereof (e.g., an augmented reality (AR) image). Hereinafter, all these types of images or pictures, as well as any other type of image or picture, will be referred to as an "image", "frame", "image data" or "frame data", unless specifically stated otherwise, while the previous explanations regarding the terms "image" or "frame", covering "video images" and "still images", still remain valid, unless explicitly stated otherwise.

A (digital) image is, or can be viewed as, a two-dimensional array or matrix of samples with intensity values. A sample in an array may also be called a pixel (short for picture element) or element. The number of samples in the horizontal and vertical directions (or along the axis) of an array or image determines the size and/or resolution of the image. Three color components are typically used to represent color, meaning an image can be represented by or include three arrays of samples. In RGB format or color space, an image contains a corresponding array of red, green, and blue samples. However, in video encoding, each pixel is typically represented in a luma/chroma format or color space, such as YCbCr, which contains a luma component, denoted Y (sometimes L is also used instead), and two chroma components, denoted Cb and Cr. The Y component of the luma (or luminance for short) represents brightness or grayscale intensity (e.g., as in a grayscale image), while the two chroma (or chrominance for short) components, Cb and Cr, represent the chrominance components or color information. Accordingly, an image in YCbCr format contains an array of luminance samples (Y) of luma samples and two arrays of chrominance samples (Cb and Cr) of chrominance values. RGB images can be converted or transformed to YCbCr format and vice versa, a process also known as color conversion or transformation. If the image is monochrome, it may contain only an array of luminance samples.

The image source 312 may be, for example, a camera for capturing an image, a memory, for example, an image memory containing or storing a previously captured or generated image, and/or an interface of any type (internal or external) for receiving or receiving an image. The camera may be, for example, a local or integrated camera integrated into the source device, the memory may be a local or integrated memory, for example, integrated into the source device. The interface may be, for example, an external interface for receiving an image from an external video source, for example, an external image capturing device, such as a camera, external memory, or an external device for generating an image, for example, an external processor of a computer graphics adapter, a computer or a server. The interface may be an interface of any type, for example, a wired or wireless interface, an optical interface in accordance with any proprietary or standardized interface protocol. The interface for receiving image data 313 may be the same interface as the communication interface 318, or a part thereof.

The interfaces between the units within each device include cable connections, USB interfaces, communication interfaces 318 and 322 between the source device 310 and the destination device 320 include cable connections, USB interfaces, radio interfaces.

In contrast to the pre-processing unit 314 and the processing performed by the pre-processing unit 314, the image or image data 313 may also be referred to as a raw image or raw image data 313.

The pre-processing unit 314 is configured to receive (raw) image data 313 and perform pre-processing of the image data 313 to obtain a pre-processed image 315 or pre-processed image data 315. The pre-processing performed by the pre-processing unit 314 may, for example, comprise cropping, color format conversion (for example, from RGB to YCbCr), color correction, or noise removal.

The encoder 100 is configured to receive pre-processed image data 315 and provide encoded image data 171 (further details will be described, for example, with reference to Fig. 2).

The communication interface 318 of the source device 310 may be configured to receive the encoded image data 171 and directly transmit it to another device, such as the destination device 320 or any other device, for storage or direct recovery or for processing the encoded image data 171 for, respectively, before storing the encoded data 330 and/or transmitting the encoded data 330 to another device, such as the destination device 320 or any other device for decoding or storing.

The destination device 320 comprises a decoder 200 or a decoding unit 200 and may additionally, i.e. possibly, comprise a communication interface or a communication unit 322, a post-processing unit 326 and a display device 328.

The communication interface 322 of the destination device 320 is configured to receive the encoded image data 171 or the encoded data 330, for example, directly from the source device 310 or from any other source, for example, a memory, for example, a memory of encoded image data.

The communication interface 318 and the communication interface 322 may be configured to transmit, respectively, receive the encoded image data 171 or the encoded data 330 via a direct communication link between the source device 310 and the destination device 320, such as a direct wired or wireless connection, including an optical connection, or via any kind of network, such as a wired or wireless network, or any combination thereof, or any kind of private and public network, or any combination thereof.

The communication interface 318 may, for example, be configured to package the encoded image data 171 into an appropriate format, such as packets, for transmission over a communication channel or communication network, and may further comprise protection against data loss.

The communication interface 322, which forms an analogue of the communication interface 318, may be, for example, configured to decompress the encoded data 330 to obtain the encoded image data 171, and may also be configured to perform data loss protection and data recovery after loss, for example, containing error concealment.

Both the communication interface 318 and the communication interface 322 may be configured as unidirectional communication interfaces, as shown by the arrow for the encoded image data 330 in Fig. 1, pointing from the source device 310 to the destination device 320, or as bidirectional communication interfaces, and may be configured to, for example, send and receive messages, such as to establish a connection, to confirm and/or resend lost or delayed data, including image data, and to exchange any other information related to the communication channel, and/or transmit data, such as transmitting encoded image data.

The decoder 200 is configured to receive the encoded image data 171 and provide decoded image data 231 or a decoded image 231 (further details will be described, for example, with reference to Fig. 9).

The post-processing processor 326 of the destination device 320 is configured to post-process the decoded image data 231, for example, the decoded image 231, to obtain the post-processed image data 327, for example, the post-processed image 327. The post-processing performed by the post-processing unit 326 may comprise, for example, a color format conversion (for example, from YCbCr to RGB), color correction, cropping or resampling, or any other processing, for example, to prepare the decoded image data 231 for display, for example, using the display device 328.

The display device 328 of the destination device 320 is configured to receive the post-processed image data 327 for displaying the image, for example, to a user or viewer. The display device 328 may be or comprise any type of display for presenting the reconstructed image, for example, an integrated or external display or monitor. Displays may, for example, comprise cathode ray tubes (CRTs), liquid crystal displays (LCDs), plasma displays, organic light emitting diodes (OLEDs) displays, or any other display, such as projectors, holographic displays, or hologram generating devices.

Although Fig. 1 depicts the source device 310 and the destination device 320 as separate devices, embodiments of the devices may also comprise both or both of the functionality of the source device 310 or the corresponding functionality of the destination device 320. In such embodiments, the source device 310 or the corresponding functionality and the destination device 320 or the corresponding functionality may be implemented using the same hardware and/or software or using separate hardware and/or software or any combination thereof.

As will be apparent to one skilled in the art based on the description, the presence and (exact) division of functionality of the various blocks or functions in the source device 310 and/or the destination device 320, as shown in Fig. 1, may vary depending on the actual device and application.

Several non-limiting examples will now be presented for the coding system 300, the source device 310 and/or the destination device 320.

Various electronic devices, such as a smartphone, tablet, or portable camera with a built-in display, can be considered examples of coding system 300. They comprise a display device 328, and most of them include an integrated camera, i.e., an image source 312. Image data captured by the integrated camera is processed and displayed. The processing may include internal encoding and decoding of the image data. Furthermore, the encoded image data may be stored in the integrated memory.

Alternatively, these electronic products may have wired or wireless interfaces for receiving image data from external sources such as the Internet or external cameras, or for transmitting encoded image data to external displays or storage devices.

On the other hand, set-top boxes do not contain a built-in camera or display, but instead process the received image data for display on an external display device. Such a box can be implemented, for example, as a chipset.

Alternatively, a device similar to a set-top box may be incorporated into a display device such as a television with a built-in display.

Surveillance cameras without a built-in display are another example of such products. They are a source device with an interface for transmitting captured and encoded image data to an external display or storage device.

In contrast, devices such as smart glasses or 3D glasses, such as those used for AR or VR, constitute a 320-device target. They receive encoded image data and display it.

Therefore, the source device 310 and the destination device 320 as shown in Fig. 1 are merely exemplary embodiments of the invention, and the embodiments of the invention are not limited to those shown in Fig. 1.

The source device 310 and the destination device 320 may comprise any of a wide range of devices, including any type of portable or stationary device, such as portable or wearable computers, mobile phones, smartphones, tablets or tablet computers, cameras, desktop computers, set-top boxes, televisions, display devices, digital media players, game consoles, video streaming devices, broadcast receiver devices, and the like. For large-scale professional encoding and decoding, the source device 310 and/or the destination device 320 may further comprise servers and workstations, which may be used in large networks. These devices may or may not use any type of operating system.

CODER AND CODING METHOD

2 shows a schematic/conceptual block diagram of an embodiment of an encoder 100, for example, an image encoder 100, which comprises an input 102, a residual calculator 104, a transform block 106, a quantizer 108, an inverse quantizer 110 and an inverse transform block 112, a reconstruction block 114, a buffer 116, a loop filter 120, a decoded picture buffer (DPB) 130, a prediction block 160, which includes an inter-estimator 142, an inter-prediction block 144, an intra-estimator 152, an intra-prediction block 154 and a mode selector 162, an entropy coding block 170 and an output 172. The video encoder 100 shown in Fig. 2 may also be called a hybrid video encoder or a video encoder according to a hybrid video codec. Each block may contain a processor and non-volatile memory to perform its processing steps by executing code stored in the non-volatile memory by the processor.

For example, the residual value calculating unit 104, the transform unit 106, the quantization unit 108, and the entropy encoding unit 170 form a forward signal path of the encoder 100, while, for example, the inverse quantization unit 110, the inverse transform unit 112, the reconstruction unit 114, the buffer 116, the loop filter 120, the decoded picture buffer (DPB) 130, the inter prediction unit 144, and the intra prediction unit 154 form a reverse signal path of the encoder, wherein the reverse signal path of the encoder corresponds to the signal path of the decoder to provide reverse processing for identical reconstruction and prediction (see decoder 200 in Fig. 3).

The encoder is configured to receive, for example, by input 102, an image 101 or a block 103 of an image 101, for example, an image of a sequence of images forming a video or a video sequence. The image block 103 may also be referred to as the current image block or the image block to be encoded, and the image 101 as the current image or the image to be encoded (in particular, when encoding video, to distinguish the current image from other images, for example, previously encoded and/or decoded images of the same video sequence, i.e., a video sequence that also includes the current image).

SEPARATION

Embodiments of the encoder 100 may comprise a partitioning unit (not shown in Fig. 2), for example, which may also be referred to as an image partitioning unit, configured to partition images 103 into a plurality of blocks, for example, blocks similar to block 103, typically into a plurality of non-overlapping blocks. The partitioning unit may be configured to use the same block size for all images of a video sequence and a corresponding grid defining the block size, or to vary the block size between images or subsets or groups of images, and to partition each image into corresponding blocks.

Each block in a block set may have square dimensions or, more commonly, rectangular dimensions. Blocks that are non-rectangular image areas may not be displayed.

Similar to image 101, block 103 is again or can be considered as a two-dimensional array or matrix of samples with intensity values (sample values), although of a smaller size than image 101. In other words, block 103 can contain, for example, one array of samples (e.g., a luminance array in the case of monochrome image 101) or three arrays of samples (e.g., luminance arrays and two chrominance arrays in the case of color image 101) or any other number and/or type of arrays depending on the color format used. The number of samples in the horizontal and vertical direction (or along the axis) of block 103 determines the size of block 103.

The encoder 100, as shown in Fig. 2, is configured to encode the image 101 block by block, for example, encoding and prediction is performed for each block 103.

CALCULATION OF THE RESIDUAL VALUE

The residual value calculating unit 104 is configured to calculate the residual block 105 based on the image block 103 and the prediction block 165 (further details about the prediction block 165 will be provided later), for example, by subtracting the sample values of the prediction block 165 from the sample values of the image block 103, sample by sample (pixel by pixel) to obtain the residual block 105 in the sample area.

TRANSFORMATION

The transform unit 106 is configured to apply a transform, such as a spatial-frequency transform or a linear spatial transform, such as a discrete cosine transform (DCT) or a discrete sine transform (DST), to the sample values of the residual block 105 to obtain transformed coefficients 107 in the transform domain. The transformed coefficients 107 may also be referred to as transformed residual coefficients and represent the residual block 105 in the transform domain.

Transformation unit 106 may be configured to apply integer DCT/DST approximations, such as the base transforms defined for HEVC/H.265. Compared to the orthonormal DCT transform, such integer approximations are typically scaled by a certain factor. To preserve the norm of the residual block processed by the forward and inverse transforms, additional scaling factors are applied as part of the transformation process. The scaling factors are typically selected based on certain constraints, such as scaling factors that are powers of two for the shift operation, the bit depth of the transformed coefficients, the tradeoff between accuracy and implementation cost, and so on. Specific scaling factors are, for example, specified for the inverse transform, for example, by the inverse transform block 212 in the decoder 200 (and the corresponding inverse transform, for example, by the inverse transform block 112 in the encoder 100) and corresponding scaling factors for the forward transform can be specified accordingly, for example, by the transform block 106 in the encoder 100.

QUANTIZATION

The quantization unit 108 is configured to quantize the transformed coefficients 107 to obtain quantized coefficients 109, for example, by applying scalar quantization or vector quantization. The quantized coefficients 109 may also be referred to as quantized residual coefficients 109. For example, for scalar quantization, another scaling may be applied to achieve finer or coarser quantization. Smaller quantization step sizes correspond to finer quantization, while larger quantization step sizes correspond to coarser quantization. The applicable quantization step size may be indicated by a quantization parameter (QP). The quantization parameter may, for example, be an index for a predetermined set of applicable quantization step sizes. For example, small quantization parameters may correspond to fine quantization (small quantization step sizes), and large quantization parameters may correspond to coarse quantization (large quantization step sizes), or vice versa. Quantization may include division by the quantization step size and corresponding or inverse dequantization, for example, by inverse quantization 110, which may include multiplication by the quantization step size. Embodiments according to High-Efficiency Video Coding (HEVC) may be configured to use a quantization parameter to determine the quantization step size. Typically, the quantization step size may be calculated based on the quantization parameter using a fixed-point approximation of the equation, which includes division. Additional scaling factors may be introduced for quantization and dequantization to restore the norm of the residual block, which may be changed due to the scaling used in the fixed-point approximation of the equation for the quantization step size and the quantization parameter. In one exemplary implementation, inverse transform scaling and dequantization may be combined. Alternatively, configured quantization tables may be used, and the encoder may transmit them to the decoder, for example, in a bitstream. Quantization is a lossy operation, with losses increasing as the size of the quantization steps increases.

Embodiments of encoder 100 (or, respectively, quantization unit 108) may be configured to output quantization settings including a quantization scheme and a quantization step size, for example, via a corresponding quantization parameter, so that decoder 200 can receive and apply the corresponding inverse quantization. Embodiments of encoder 100 (or quantization unit 108) may be configured to output quantization schemes and a quantization step size, for example, directly or entropy-encoded via entropy encoding unit 170 or any other entropy encoding unit.

The inverse quantization unit 110 is configured to apply the inverse quantization of the quantization unit 108 to the quantized coefficients to obtain dequantized coefficients 111, for example, by applying the inverse quantization scheme applied by the quantization unit 108, based on or using the same quantization stage size as the quantization unit 108. The dequantized coefficients 111 can also be referred to as dequantized residual coefficients 111 and correspond to, although usually not identical to, the transformed coefficients due to quantization losses, the transformed coefficients 108.

The inverse transform unit 112 is configured to apply the inverse transform applied by the transform unit 106, for example, the inverse discrete cosine transform (DCT) or the inverse discrete sine transform (DST), to obtain the inverse transform unit 113 in the sample domain. The inverse transform unit 113 may also be referred to as the dequantized inverse transform unit 113 or the residual inverse transform unit 113.

The reconstruction unit 114 is configured to combine the inverse transform unit 113 and the prediction unit 165 to obtain a reconstructed block 115 in the sample domain, for example, by sequentially adding the sample values of the decoded residual block 113 and the sample values of the prediction unit 165.

Buffer block 116 (or "buffer" 116 for short), for example, link buffer 116, is configured to buffer or store a reconstructed block and the corresponding sample values, for example, for intra-estimation and/or intra-prediction. In additional embodiments, the encoder may be configured to use unfiltered reconstructed blocks and/or the corresponding sample values stored in buffer block 116 for any type of estimation and/or prediction.

Embodiments of the encoder 100 may be configured such that, for example, the buffer block 116 is used not only to store the reconstructed blocks 115 for the intra-estimation 152 and/or intra-prediction 154, but also for the loop filter block 120 and/or, for example, the buffer block 116 and the decoded image buffer block 130 form a single buffer. Further embodiments may be configured to use the filtered blocks 121 and/or blocks or samples from the decoded image buffer 130 (both not shown in Fig. 2) as input data or the basis for the intra-estimation 152 and/or intra-prediction 154.

The loop filter block 120 (for short, the "loop filter" 120) is configured to filter the reconstructed block 115 to obtain a filtered block 121, for example, by applying a sample-adaptive offset (SAO) deblocking filter or other filters, such as sharpening or anti-aliasing filters or joint filters. The filtered block 121 may also be referred to as the filtered reconstructed block 121.

Embodiments of the loop filter unit 120 may comprise a filter analysis unit and an actual filter unit, wherein the filter analysis unit is configured to determine the loop filter parameters for the actual filter. The filter analysis unit may be configured to apply fixed predetermined filter parameters to the actual loop filter, adaptively select filter parameters from a set of predetermined filter parameters, or adaptively calculate filter parameters for the actual loop filter.

Embodiments of the loop filter unit 120 may comprise (not shown in Fig. 2) one or a plurality of filters (such as loop filter components and/or subfilters), such as one or more different kinds or types of filters, such as those connected in series or in parallel or in any combination thereof, wherein each of the filters may comprise, individually or together with other filters of the plurality of filters, a filter analysis unit for determining corresponding loop filter parameters, such as described in the preceding paragraph.

Embodiments of the encoder 100 (respectively, the loop filter unit 120) may be configured to output the loop filter parameters, for example, directly or entropy encoding via the entropy encoding unit 170 or any other entropy encoding unit, so that, for example, the decoder 200 can receive and apply the same loop filter parameters for decoding.

The decoded picture buffer 130 (DPB) is configured to receive and store the filtered block 121. The decoded picture buffer 130 may be further configured to store other previously filtered blocks, such as previously reconstructed and filtered blocks 121 of the same current picture or different pictures, such as previously reconstructed pictures, and may provide complete previously reconstructed, i.e., decoded pictures (and corresponding reference blocks and samples) and/or partially reconstructed current picture (and corresponding reference blocks and samples), such as for evaluation and/or inter prediction.

Further embodiments of the invention may also be configured to use previously filtered blocks and corresponding filtered sample values of the decoded picture buffer 130 for any type of estimation or prediction, such as intra-estimation and prediction, as well as inter-estimation and prediction.

The prediction unit 160, also called the block prediction unit 160, is configured to receive or obtain the image block 103 (the current image block 103 of the current image 101) and decoded or at least reconstructed image data, such as reference samples of the same (current) image from the buffer 116 and/or decoded image data 231 from one or a plurality of previously decoded images from the decoded image buffer 130, and to process such data for prediction, that is, to provide the prediction unit 165, which may be the inter prediction unit 145 or the intra prediction unit 155.

The mode selection unit 162 may be configured to select a prediction mode (e.g., an intra- or inter-prediction mode) and/or a corresponding prediction unit 145 or 155 that will be used as a prediction unit 165 for calculating the residual block 105 and for reconstructing the reconstructed block 115.

Embodiments of the mode selection unit 162 may be configured to select a prediction mode (e.g., from those supported by the prediction unit 160) that provides the best match, or in other words, the minimum residual value (the minimum residual value means the best compression for transmission or storage) or the minimum signaling overhead (the minimum signaling overhead means the best compression for transmission or storage), or takes into account or balances both. The mode selection unit 162 may be configured to determine the prediction mode based on rate distortion optimization (RDO), that is, to select a prediction mode that provides the minimum rate distortion optimization or whose associated rate distortion at least satisfies the prediction mode selection criterion.

The prediction processing (e.g., prediction block 160) and mode selection (e.g., mode selection block 162) performed by the exemplary encoder 100 will now be described in more detail.

As described above, encoder 100 is configured to determine or select the best or optimal prediction mode from a set of (predetermined) prediction modes. The set of prediction modes may comprise, for example, intra-prediction modes and/or inter-prediction modes.

The set of intra-prediction modes may contain 32 different intra-prediction modes, such as non-directional modes such as DC mode (or mean) and planar mode, or directional modes such as defined in H.264, or may contain 65 different intra-prediction modes, such as non-directional modes such as DC mode (or mean) and planar mode, or directional modes such as defined in H.265.

The set (or possible) of inter prediction modes depends on the available reference pictures (i.e., previous at least partially decoded pictures, for example, stored in DPB 230) and other inter prediction parameters, for example, whether the entire reference picture or only a part, for example, a search window of an area around the area of the current block of the reference picture is used to find the best match of the reference block and/or, for example, whether pixel interpolation, for example, half/half pixel and/or quarter pixel interpolation, is applied or not.

In addition to the above mentioned prediction modes, skip mode and/or direct mode may be applied.

The prediction unit 160 may be further configured to partition the block 103 into smaller blocks or sub-blocks, for example, iteratively using quadratic partitioning (quad-tree-partitioning, QT), binary partitioning (BT) or triple-tree-partitioning (TT) or any combination thereof, and to perform, for example, a prediction for each of the partitioned block or sub-blocks of the block, wherein the mode selection comprises selecting the tree structure of the partitioned block 103 and the prediction modes applied to each of the partitioned blocks or sub-blocks.

The external evaluation unit 142, also referred to as the external image evaluation unit 142, is configured to receive or obtain an image block 103 (the current image block 103 of the current image 101) and a decoded image 231 or at least one or a plurality of previously reconstructed blocks, for example, reconstructed blocks of one or a plurality of other/different previously decoded images 231, for external evaluation (or "external image evaluation"). For example, the video sequence may comprise the current image and the previously decoded images 231, or, in other words, the current image and the previously decoded images 231 may be part of or form a sequence of images that form the video sequence.

The encoder 100 may, for example, be configured to select (obtain/determine) a reference block from a plurality of reference blocks of the same or different images from a plurality of other images and provide the reference image (or index of the reference image, ...) and/or the offset (spatial offset) between the position (X, Y coordinates) of the reference block and the position of the current block as inter-estimation parameters 143 to the inter-prediction unit 144. The offset is also called a motion vector (MV). The inter-estimation is also called motion estimation (ME), and the inter-prediction is also called motion prediction (MP).

The inter prediction unit 144 is configured to receive, for example, the inter prediction parameter 143 and to perform inter prediction based on or using the inter prediction parameter 143 to obtain the inter prediction unit 145.

Although Fig. 2 shows two separate blocks (or stages) for inter coding, namely, inter evaluation 142 and inter prediction 152, both functions can be performed as one (inter evaluation typically requires/contains an inter prediction calculation/block, i.e., or "kind" of inter prediction 154), for example, by iteratively testing all possible or a predetermined subset of possible inter prediction modes, while storing the currently best inter prediction mode and the corresponding inter prediction block, and using the current best inter prediction mode and the corresponding inter prediction block as the (final) inter prediction parameter 143 and inter prediction block 145 without performing another inter prediction 144.

The intra-estimation unit 152 is configured to receive, for example, an image block 103 (the current image block) and one or a plurality of previously reconstructed blocks, for example, reconstructed neighboring blocks, of the same image for intra-estimation. The encoder 100 may, for example, be configured to select (obtain/determine) an intra-prediction mode from a plurality of intra-prediction modes and provide it as an intra-estimation parameter 153 to the intra-prediction unit 154.

Embodiments of the encoder 100 may be configured to select an intra-prediction mode based on an optimization criterion, such as a minimum residual (e.g., an intra-prediction mode that provides a prediction block 155 that is most similar to the current image block 103) or minimum distortion.

The intra-prediction unit 154 is configured to determine, based on the intra-prediction parameter 153, for example, the selected intra-prediction mode 153 of the intra-prediction unit 155.

Although Fig. 2 shows two separate blocks (or stages) for intra coding, namely, intra estimation 152 and intra prediction 154, both functions can be performed as one (intra estimation typically requires/contains the computation of an intra prediction block, i.e., or "kind" of intra prediction 154), for example, by iteratively testing all possible or a predetermined subset of possible intra prediction modes, storing the currently best intra prediction mode and the corresponding intra prediction block, and using the currently best intra prediction mode and the corresponding intra prediction block as the (final) intra prediction parameter 153 and the intra prediction block 155 without performing another intra prediction 154.

The entropy coding unit 170 is configured to apply an entropy coding algorithm or scheme (e.g., a variable length coding scheme (VLC), a context adaptive VLC scheme (CALVC), an arithmetic coding scheme, a context adaptive binary arithmetic coding scheme (CABAC)) on the quantized residual coefficients 109, the inter prediction parameters 143, the intra prediction parameter 153 and/or the loop filter parameters, individually or jointly (or nothing), to obtain coded image data 171, which can be output via the output 172, for example, in the form of an encoded bitstream 171.

DECODER

Fig. 3 shows an example video decoder 200 configured to receive encoded image data 171 (e.g., an encoded bitstream), such as encoded by encoder 100, to obtain a decoded image 231.

The decoder 200 comprises an input 202, an entropy decoding unit 204, an inverse quantization unit 210, an inverse transform unit 212, a restoration unit 214, a buffer 216, a loop filter 220, a decoded image buffer 230, a prediction unit 260 which includes an outer prediction unit 244, an inner prediction unit 254 and a mode selection unit 260 and an output 232.

The entropy decoding unit 204 is configured to perform entropy decoding of the encoded image data 171 to obtain, for example, quantized coefficients 209 and/or decoded coding parameters (not shown in Fig. 3), for example, (decoded) any or all of the prediction parameters 143, the intra prediction parameter 153 and/or the loop filter parameters.

In embodiments of decoder 200, inverse quantization unit 210, inverse transform unit 212, reconstruction unit 214, buffer 216, loop filter 220, decoded image buffer 230, prediction unit 260 and mode selection unit 260 are configured to perform inverse processing of encoder 100 (and corresponding functional units) for decoding coded image data 171.

In particular, the inverse quantization unit 210 may be identical in function to the inverse quantization unit 110, the inverse transform unit 212 may be identical in function to the inverse transform unit 112, the reconstruction unit 214 may be identical in function to the reconstruction unit 114, the buffer 216 may be identical in function to the buffer 116, the loop filter 220 may be identical in function to the loop filter 220 (with regard to the actual loop filter, since the loop filter 220 typically does not comprise a filter analysis unit for determining the filter parameters based on the original image 101 or the block 103, but receives (explicitly or implicitly) or obtains the filter parameters used for encoding, for example, from the entropy decoding unit 204), and the decoded image buffer 230 may be identical in function to the decoded image buffer 130.

The prediction unit 260 may comprise an inter-prediction unit 244 and an intra-prediction unit 254, wherein the inter-prediction unit 244 may be functionally identical to the inter-prediction unit 144 and the intra-prediction unit 254 may be functionally identical to the intra-prediction unit 154. The prediction unit 260 and the mode selection unit 262 are typically configured to perform block predictions and/or obtain a predicted block 265 only from the encoded data 171 (without any additional information about the original image 101) and to receive or obtain (explicitly or implicitly) the prediction parameters 143 or 153 and/or information about the selected prediction mode, for example, from the entropy decoding unit 204.

The decoder 200 is configured to output the decoded image 231, for example via the output 232, for presentation or viewing to a user.

Returning to Fig. 1, the decoded image 231 output from the decoder 200 may be post-processed in the post-processing processor 326. The resulting post-processed image 327 may be transmitted to an internal or external display device 328 and displayed.

DETAILS OF IMPLEMENTATION OPTIONS

The invention limits the values that can be used by the filter coefficients of an adaptive multiplicative filter so that the multiplication operation is simplified. When filtering a set of image signal samples, a filter with adaptive multiplier coefficients is used, where the multiplier coefficients are represented by integers. Given that the maximum absolute value of the coefficient C is N, the binary representation of N requires L = ceil (log ₂ (N)) binary digits. In other words, L binary digits can be expressed by the absolute values of the coefficients from zero (L "zeros") to 2 ^L - 1 (L "ones") (the sign of the coefficient is represented by a separate sign bit, which is not discussed here). In accordance with a specific approach of the invention, this set of values is limited so that any value that can be accepted by the coefficient C includes at most P<L "ones"("1") in the binary representation. For example, the case of all "ones" (L "ones") is excluded.

As will be shown below, the smaller the number P of permitted "ones," the higher the efficiency and performance of the filtering operation. For example, the greatest gain in efficiency can be achieved if any value that can be accepted by the coefficient C includes only up to one "1," that is, no more than one "1."

Specific embodiments of the invention will be described in detail below.

It should be noted that the exemplary values of the parameters given below are intended for illustrative purposes only, and a person skilled in the art knows that they can be replaced by any other possible values that are within the scope of the attached claims.

Filter coefficients are typically implemented using finite precision. A filter coefficient is represented using L bits, along with an optional sign bit. The number of bits, L, depends on the maximum absolute value of the coefficient. Specifically, given that the largest absolute value of coefficient C is N, the binary representation of N requires L = ceil(log ₂ (N)) binary digits.

The ceil(x) function, also denoted as or max(x), maps x to the smallest integer greater than or equal to x.

According to a first exemplary embodiment of the invention, at most one of the L bits (i.e., excluding the sign bit) of the filter coefficient may simultaneously be "one" ("1"). Other possibilities are not allowed.

For example:

Suppose L = 6, and one bit (the leftmost bit) is used to indicate the sign of the coefficient.

For example, the following filter coefficients are valid: 0 (0000000), 1 (0000001), -1 (1000001), 2 (0000010), -2 (1000010), 4 (0000100), -4 (1000100), 8 (0001000), -8 (1001000), 16 (0010000) …, -32 (1100000).

For example, the following filter coefficients are not allowed: 3 (0000011), -15 (1001111), 31 (0011111) …

In this case, an advantage is achieved because the constraint allows the multiplication to be implemented as a one-bit shift operation.

The bit shift operation can be mathematically represented as: f(X, M) = X * 2 ^M , where M is an integer greater than or equal to 0. According to a generalization of the above embodiment, no more than M of the L filter coefficient bits can be "1" simultaneously. Other possibilities are not allowed.

For example:

Assume L = 6, M = 2, and one bit is used to indicate the sign of the coefficient.

For example, the following filter coefficients are allowed: 0 (0000000), 3 (0000011), 9 (0001001), -4 (1000100), -9 (1001001), 18 (0010010), 33 (0100001) …

For example, the following filter coefficients are not allowed: 7 (0000111), -19 (1010011), 31 (0011111) …

In this case, the constraint allows the multiplication to be implemented using two bit shift operations and one addition operation.

In the more general case described above, for general M<L, an advantage is achieved because the constraint allows multiplication to be performed using M-bit shift and M-1-addition operations.

The examples above assume that a limited set of absolute values applies to all coefficients of the adaptive filter multiplication filter.

A more complex exemplary embodiment will now be described with reference to Fig. 7, in which the limitation according to the invention is applied, but not to all filter coefficients of the filter in question.

In the example, at the first stage, the coefficients are grouped into two groups. In the figure, the first group corresponds to the coefficient positions indicated by uncolored circles in the central part of the filter, and the second group corresponds to the coefficient positions indicated by filled black circles in the peripheral part of the filter.

The filter coefficients in the first group can take any value within a predefined range. In the illustrated example, it is assumed that this range corresponds to the set "S1", where S1 = [-511, …, 511]. This corresponds to a total number of bits (excluding the sign bit) of L = 9.

The filter coefficients in the second group can take any value in the set "S2", where S2 is a subset of S1. More specifically, in the example, the set S2 is defined as S2 = [-32, -16, -8, -4, -2, -1,0,1,2,4,8,16,32]. Accordingly, the permissible values in the set S2 are limited to those that can be represented by a single "1" in binary notation. In addition, the maximum permissible absolute value is limited to 32, that is, it is further assumed that the number L is limited to L = 6. In general, it should be noted that the number L can be set separately and differently for each group. Moreover, the specific grouping and definition of sets of permissible values may vary from image to image (from frame to frame). Alternatively, the grouping and definition of sets can be different for different filter shapes (e.g., 5x5 diamond, 7x7 diamond, 9x9 diamond, as shown in Fig. 5). Alternatively, the grouping and definitions can be predefined.

In this example, the advantage is that instead of a 9-bit multiplication, a 1-bit shift is used for the S2 set.

The corresponding data must be included in the bitstream in the encoder and communicated to the decoder so that the filter coefficients can also be correctly determined in the decoder. Of course, using a limited set of valid coefficient values leads to a reduction in the amount of signaling overhead and, thus, to more efficient coding, since fewer bits are required to represent the coefficients that must be signaled in the bitstream.

More specifically, the filter coefficient values applied by the encoder must be encoded and passed to the decoder. At the encoder, the filter coefficient values are converted into binary codewords (from filter value to codeword) via a mapping table or mapping function. The same mapping operation must be applied in the decoder (from codeword to filter coefficient value) to correctly interpret the filter coefficients.

The mapping function or table may differ for S1 and S2. Example mapping operations are given below for filter coefficient sets S1 and S2.

In the example below, S1 is given by {0,1, …, 511}, and S2 is given by {0,2,4,8,16,32} (only absolute values are considered).

S1 S2 Value of the filter coefficient Code word Value of the filter coefficient Code word 0 000000000 0 000 1 000000001 2 001 2 000000010 4 010 3 000000011 8 011 4 000000100 16 100 5 000000101 32 101 6 000000110 …

Forward (in the encoder) and reverse (in the decoder) transform operations must be used in the encoder and decoder to ensure that the decoder can correctly interpret the filter coefficient values. In the example above, the filter coefficient mapping operation differs for S2 and S1 because the number of distinct values in S2 is much smaller, and representing the S2 filter coefficients using the S1 mapping is wasteful. Therefore, the invention results in a reduction in signaling overhead and, thus, more efficient coding, since fewer bits are needed to represent the coefficients that must be signaled in the bitstream.

A general overview of the signaling of filter coefficients will now be given with reference to Fig. 8. Fig. 8A illustrates the processing on the encoder side, and Fig. 8B illustrates the processing on the decoder side.

In the encoder, the filter coefficients to be applied to the reconstructed samples are determined in accordance with the permissible values of the coefficients as defined by the specific approach of the invention (step S80).

The determined filter coefficients are used to filter the reconstructed image samples (step S82). According to the invention, the filter coefficients applied to the reconstructed image samples must comply with the rules set forth in accordance with the invention.

The next step of predicting filter coefficients (step S84) is optional. Filter coefficient prediction can be used additionally to reduce the amount of information transmitted to the decoder. Possible prediction methods include prediction using specified filter predictors and prediction based on previously reported filter coefficients. However, the prediction methods are not limited to those exemplified in this document, and any suitable prediction method known to a skilled person may be used.

In the next step (S86), the residual coefficients are mapped onto binary codewords. While the aforementioned prediction step S84 is possible, it should be noted that, as an alternative, the mapping can be applied directly to the filter coefficients determined in step S80.

More specifically, each integer filter coefficient (filter residual) is converted to a binary codeword before being included in the bitstream. There are as many codewords as there are possible filter coefficient values (filter residuals). The codeword-to-value mapping (which is a one-to-one mapping) can be a fixed mapping or can change depending on the additional information being signaled.

In the final step S88, the binarized filter coefficients (possibly residual ones), i.e., the codewords to which they were mapped, are included in the bitstream. If prediction is performed in step S84, it is additionally necessary to generate prediction control information and include it in the bitstream to signal the decoder the necessary prediction processing information to perform reconstruction.

Typically, the operations applied in the encoder are applied in reverse order in the decoder. This will be explained in more detail below with reference to Fig. 8B.

At the initial stage of S90, the received bitstream is parsed. The resulting binarized filter coefficients (i.e., the transmitted codewords) may represent residual filter coefficients (if prediction was applied at the encoder end). This is indicated by the additional acquisition of prediction control information from the parsed bitstream.

In any case, the binary code words are mapped using an inverse transform procedure (compared to the encoder) to filter coefficients (or residual filter coefficients) in step S92.

As a result, the filter coefficients are determined (reconstructed) at the decoder side (step S94). If prediction was applied such that the filter coefficients obtained in step S92 are residual filter coefficients, reconstruction further includes performing a prediction as specified in the prediction control information and adding the prediction result to the residual filter coefficients to obtain reconstructed filter coefficients.

After the filter coefficients are reconstructed (if applicable, by combining the predictor information and the filter residuals), they are applied to the samples of the reconstructed image (step S96).

According to the invention, the filter coefficients that are applied to the samples of the reconstructed image must obey the rules defined according to the invention.

Accordingly, if the filter coefficient obtained as a result of reconstruction (specifically, from combining the prediction results and residual results) does not have a valid filter coefficient value according to the rules of the invention (a filter coefficient value that is not included in the set of valid values), the filter coefficient reconstruction process additionally performs a rounding operation. Specifically, the rounding operation may transform the input filter coefficient value to the nearest valid coefficient value.

If filter coefficient prediction is used, the filter coefficients to be applied to the reconstructed image samples for filtering purposes are obtained by adding the prediction result (the "predictor") and the residual filter coefficients (as explained in the preceding paragraphs for the encoder and decoder, respectively). Obviously, it is possible for the residual filter coefficients to be absent (equal to 0), especially if the prediction is close to perfect (the filter coefficients to be predicted are very similar to the predictor). In this case, according to the invention, one of the following two options is used:

1. The coefficient values obtained by prediction must comply with the rules defined in accordance with the invention. For example, in the case of prediction based on predetermined filters, the filter coefficients of the predetermined filters must comply with the rules defined in accordance with the invention.

2. The filter coefficients obtained after prediction must be rounded to the nearest acceptable coefficient value.

It should further be noted that the division into several two groups has been explained here only for the sake of simplicity, but more than two groups are also possible, in which, for at least one group, the set of permissible values is determined according to the invention, i.e. only a limited number of "ones" within a predetermined total number of binary digits.

For example, Fig. 9 illustrates the case where the filter coefficients are grouped into three groups.

The first group of coefficients, located close to the center of the filter kernel, determines the values of the filter coefficients in the set S1 = [-511, …, 511].

The second group of filter coefficients, located at the periphery of the core and indicated by dashed circles, allows the values of the filter coefficients to lie within a modified bounded set S2, in which S2 here S2 = [-128, -64, -32, -16, -8, -4, -2, -1,0,1,2,4,8,16,32,64,128]. This is the set of all coefficient values that can be represented by L = 8 binary digits, with only one "one".

The third group of filtration coefficients, located between the first and second groups and indicated by filled circles, allows the values of the filtration coefficients to be within another limited set S3, in which

S3 = [-64, -48, -40, …, 0,1,2,3,4,5,6,8,9,10,12,16,17,18,20,24,32,33, 34,36,40,48,64].

In other words, the set S3 is the set of all coefficients that can be represented with L = 7 binary digits, in which at most two bits are equal to "one" in the absolute value of the coefficient, and the additional constraint that the maximum absolute value is set to 64 applies. (Otherwise, for example, the absolute value 96 would also have to be allowed, since it can be expressed with two leading "ones" in 7 binary digits.)

Next, with reference to Fig. 10, another exemplary embodiment of a specific advantage of the invention is shown.

In the example in Fig. 10, the grouping is performed in the same way as in Fig. 7.

The filter coefficients in the first group can take any values with nine full-range bits and a sign bit, that is, the above set S1 = [-511, 511].

The filter coefficients in the second group can take on a limited set of values S2, where S2 here S2 = [-128, -64, -32, -16, -8, -4, -2, -1, 0, 1, 2, 4, 8, 16, 32, 64, 128]. This corresponds to the values that can be represented by a single "1" in binary notation. Moreover, the maximum allowable absolute value is limited to 128, i.e., it is additionally assumed that the number L is limited to L = 8.

In other words, the filter size corresponds to that shown in Fig. 4, that is, a 9 × 9 diamond-shaped filter. As noted in the Background section, 41 multiplications with 9-bit filter coefficients are typically required. Since one multiplication is equivalent to 8 binary additions, as mentioned in the Background section, the number of additional operations per pixel is 48 * 8 = 328 addition operations.

According to the invention, the peripheral 28 coefficients (i.e., the coefficients in the second group) can be implemented as a single bit shift operation. The implementation of the bit shift operation has very little hardware complexity and can thus be omitted from the calculation.

Thirteen multiplication operations with 9-bit coefficients equal 13 * 8 = 104 additions per pixel. The number of operations per pixel is reduced by 68%.

The numbers above are rough estimates, and the exact complexity reduction will depend on the actual implementation.

An additional benefit of implementations using at least two coefficient groups is explained below.

According to the invention, not all filter coefficients are coarsely quantized, and the filter coefficients in the first group have a more precise quantization.

Typically, coarse quantization of filter coefficients causes coding loss. However, the presence of a first group of filter coefficients that accepts a wide range of values can be used to compensate for coding loss by the encoder.

A possible implementation of the encoder is as follows. The following description uses the filter coefficient labels shown in Fig. 11, which may differ from those previously used in connection with other drawings:

Step 1: Obtain all filter coefficients (C ₀ , …, C ₂₀ ) using the least squares method, assuming no restrictions on the coefficient values.

Step 2: Set the constraint by rounding the coefficients (C ₇ , …, C ₂₀ ) to the nearest acceptable value.

This stage introduces quantization noise into the filter coefficients and thus reduces the coding gain.

Step 3: Re-estimate the freely selectable filter coefficients ( _C0 , …, _C6 ) to compensate for the quantization error. This third step allows us to recover most of the coding loss introduced in Step 2.

In more detail:

In the first step, the equation below is solved for a 41-tap filter (with 21 unique coefficients):

The above equation is called the least squares equation and is used to find the Cx filter coefficients in the encoder.

The element X _x,y represents the expected value R(i + k, j + l) * R(i + m, j + n), the correlation between the two reconstructed samples before filtering. The indices k, l, m, and n are chosen according to the shape of the filter applied.

The element P _x denotes the expected value R(i + k, j + l) * O(i, j).

At the second stage, for the filter coefficients from C ₇ to C _20, the closest approximate coefficients are found that satisfy the constraints:

The coefficients from _C7 ' to _C20 ' obey the rules established by the invention. Note that the f() function described above introduces quantization noise into the filter coefficients from _C7 to _C20 , which were previously obtained by solving the least-squares equation.

The quantization noise introduced in the second stage is expected to reduce the performance of the filtering operation. Filtering effectiveness is typically measured by the peak signal-to-noise ratio (PSNR), so after stage 2, the PSNR of the filtered image will be reduced.

In the third step, the equation below is solved for a 13-tap filter (with 7 unique coefficients):

In the third stage, the filter coefficients from C ₀ to C ₇ are recalculated, taking into account the quantization noise introduced in the second stage. This stage primarily reduces the reduction in filtering efficiency caused by the application of stage 2.

It should be noted that, in general, the application of the filtering operation with the coefficients of the adaptive multiplicative filter is not limited to samples of reconstructed images. As described in Fig. 2 and Fig. 3, the reconstructed block typically corresponds to the image block that is obtained after the combination of the inverse transform block and the prediction block. As is obvious to those skilled in the art, the filtering operation with the adaptive filter coefficients can also be applied at other stages of the encoding and decoding operations, for example, to the prediction block (265, 165), the inverse transform block (213, 113), the quantized coefficients (209, 109), the dequantized coefficients (111, 211), or the decoded image (231). In this case, the invention is applied to the filter coefficients of the filtering operation.

Thus, the invention relates to an improved device and method for filtering reconstructed images, particularly video images, using adaptive multiplicative filters. The efficiency of the filtering operation is increased by limiting the permissible values of the filter coefficients to those that have only a limited number of "ones" in their binary representation.

Note that this specification provides explanations for pictures (frames), but fields are replaced by pictures in the case of an interlaced image signal.

Although embodiments of the invention have been primarily described in terms of video coding, it should be noted that embodiments of encoder 100 and decoder 200 (and, accordingly, system 300) may also be configured to process or code still images, that is, to process or code an individual image independently of any preceding or following image, as in video coding. In general, if the image processing coding is limited to one image 101, then in this case only the inter-estimator 142, inter-prediction 144, 242 are not available. Most, if not all, of the other functions (also called tools or technologies) of the video encoder 100 and the video decoder 200 can be equally used for still images, such as for splitting, transforming (scaling) 106, quantizing 108, inverse quantizing 110, inverse transforming 112, intra-estimator 142, intra-prediction 154, 254 and/or loop filter 120, 220 and entropy coding 170 and entropy decoding 204.

Wherever the embodiments and description refer to the term "memory," the term "memory" shall be understood to include and/or include a magnetic disk, an optical disk, a solid state drive (SSD), a read-only memory (ROM), a random access memory (RAM), a USB flash drive, or any other suitable type of memory, unless otherwise explicitly stated.

Wherever the embodiments and description refer to the term "network," the term "network" shall be understood to mean and/or shall include any type of wireless or wired network, such as a Local Area Network (LAN), a Wide Area Network (WLAN), a Wide Area Network (WAN), Ethernet, the Internet, mobile networks, etc., unless otherwise explicitly stated.

One skilled in the art will understand that the "blocks" ("blocks" or "modules") in the various drawings (methods and devices) represent or describe the functionality of embodiments of the invention (and not necessarily individual "blocks" in hardware or software) and thus equally describe the functions or features of embodiments of the device as well as embodiments of the method (block = step).

The terminology "blocks" is used merely for illustrative purposes of the functionality of the embodiments of the encoder/decoder and is not intended to limit the present invention.

In several embodiments presented in the present invention, it should be understood that the disclosed system, device, and method may be implemented in other ways. For example, the described embodiment of the device is merely exemplary. For example, the division into blocks is simply a division of logical functions and may be different in an actual implementation. For example, multiple blocks or components may be combined or integrated into another system, or some functions may be ignored or not performed. Additionally, the displayed or discussed interrelationships or direct connections or communication links may be implemented using certain interfaces. Indirect connections or communication links between devices or blocks may be implemented in electronic, mechanical, or other forms.

Blocks described as separate parts may or may not be physically separate, and parts depicted as blocks may or may not be physical blocks, may be located in a single location, or may be distributed across multiple network blocks. Some or all of the blocks may be selected according to actual needs to achieve the objectives of the solutions of the embodiments.

Additionally, the functional blocks in embodiments of the invention may be integrated into a single processing block, or each of the blocks may exist separately physically, or two or more blocks may be combined into a single block.

Embodiments of the invention may further comprise a device, such as an encoder and/or decoder, that comprises a processing circuit configured to perform any of the methods and/or processes described herein.

Embodiments of the encoder 100 and/or decoder 200 may be implemented as hardware, firmware, software, or any combination thereof. For example, the encoder/encoding or decoder/decoding functionality may be performed by a processing circuit with or without firmware or software, such as a processor, microcontroller, digital signal processor (DSP), field programmable gate array (FPGA), application-specific integrated circuit (ASIC), or the like.

The functionality of the encoder 100 (and the corresponding encoding method 100) and/or the decoder 200 (and the corresponding decoding method 200) may be implemented using software instructions stored on a machine-readable medium. The software instructions, when executed, cause the processing circuit, computer, processor, etc., to perform the steps of the encoding and/or decoding methods. The machine-readable medium may be any medium, including a non-volatile medium on which a program is stored, such as a Blu-ray disc, DVD, CD, USB (flash drive), hard drive, server storage accessible via a network, etc.

An embodiment of the invention comprises or is a computer program containing program code for performing any of the methods described herein when executed on a computer.

An embodiment of the invention comprises or is a computer-readable medium containing program code that, when executed by a processor, causes a computer system to perform any of the methods described herein.

An embodiment of the invention comprises or is a set of chips that performs any of the methods described herein.

LIST OF REFERENCE POSITIONS

100 coder

102 input (e.g. input port, input interface)

103 image block

104 residual value calculation [block or step]

105 residual block

106 transformation (e.g. additionally containing scaling) [block or step]

107 transformed coefficients

108 quantization [block or stage]

109 quantized coefficients

110 inverse quantization [block or stage]

111 dequantized coefficients

112 inverse transform (e.g. additionally containing scaling) [block or step]

113 reverse transformed block

114 recovery [block or stage]

115 restored block

116 (linear) buffer [block or stage]

117 reference samples

120 loop filter [block or stage]

121 filtered blocks

130 Decoded picture buffer (DPB) [block or stage]

142 external evaluation (or external image evaluation) [block or stage]

143 external evaluation parameters (e.g., reference image/reference image index, motion vector/offset)

144 inter prediction (or inter image prediction) [block or stage]

145 external prediction block

152 Internal Evaluation (or Internal Image Evaluation) [block or stage]

153 intra-prediction parameters (e.g. intra-prediction mode)

154 intra prediction (frame/picture intra prediction) [block or stage]

155 intra prediction block

162 mode selection [block or stage]

165 prediction block (either inter-prediction block 145 or intra-prediction block 155)

170 entropy coding [block or stage]

171 encoded image data (e.g. bitstream)

172 pin (output port, output interface)

200 decoder

202 input (port/interface)

204 entropy decoding

209 quantized coefficients

210 inverse quantization

211 dequantized coefficients

212 inverse transform (scaling)

213 reverse-transformed block

214 Recovery (Block)

215 restored block

216 (linear) buffer

217 reference samples

220 contour filter (in contour filter)

221 filtered blocks

230 decoded picture buffer (DPB)

231 decoded images

232 pin (port/interface)

244 inter prediction (image/frame inter prediction)

245 external prediction block

254 intra prediction (intra image/frame prediction)

255 intra prediction block

260 mode selection

265 prediction block (inter prediction block 245 or intra prediction block 255)

300 coding system

310 source device

312 source image

313 (raw) image data

314 preprocessor/preprocessing unit

315 pre-processed image data

318 communication block/interface

320 destination device

322 communication block/interface

326 post processor/post processing unit

327 post-processed image data

328 device/display unit

330 transmitted/received/messaged (encoded) image data.

Claims (19)

1. A device for filtering a set of image samples (115; 215) using a loop filter (120; 220) with adaptive multiplier coefficients represented by integers, comprising a processing circuit configured to determine the value of each multiplier coefficient of the filter (120; 220) within a set of permissible values, wherein the set of permissible values consists of [-64, -32, -16, -8, -4, -2, -1, 0, 1, 2, 4, 8, 16, 32, 64], obtaining a set of image samples (115; 215) and filtering the set of image samples (115; 215) using the filter (120; 220). 2. The device according to claim 1, in which the maximum absolute value of the set of permissible values is limited by a given maximum value N _max . 3. The device according to claim 1 or 2, in which all coefficients of the loop filter multiplier (120; 220) are determined to be within the limits of the said set of permissible values. 4. A device according to any one of claims 1 to 3, in which the processing circuit is further configured to group the multiplier coefficients of said loop filter (120; 220) into at least two groups, and the multiplier coefficients of one of the groups are limited by said set of permissible values. 5. The device according to claim 4, in which the coefficients of the multiplier of another one of the mentioned groups can take all values within the range determined by the specified maximum absolute value. 6. A device according to any one of claims 1 to 5, wherein the set of image samples is a set of samples (115; 215) of a video image. 7. The device according to claim 6, wherein said device is configured to individually adapt the multiplier coefficients for each image and each pixel. 8. A method for filtering a set of image samples (115; 215) using a loop filter (120; 220) with adaptive multiplier coefficients represented by integers, comprising the steps of determining (S80; S92, S94) the value of each multiplier coefficient of the filter (120; 220) within a set of permissible values, wherein the set of permissible values consists of [-64, -32, -16, -8, -4, -2, -1, 0, 1, 2, 4, 8, 16, 32, 64], obtaining a set of image sample values (115; 215) and filtering (S82; S96) the set of image samples (115; 215) using the loop filter (120; 220). 9. A device for encoding a current set of samples (103) of an image, including a plurality of pixels, comprising an encoder with a decoder (110, 112, 114) for reconstructing the current set and a device according to any one of claims 1-7 for filtering the reconstructed set (115). 10. The device according to claim 9, further comprising a processing circuit configured to map the value of the multiplier coefficients to binary code words and add the code words to a bit stream (330) for transmission to the decoding device (200). 11. The device according to claim 10, wherein the length of said code words depends on the number of different values of the multiplier coefficient. 12. The device according to claim 10 or 11, wherein the processing circuit is further configured to perform a prediction of the coefficients of the loop filter multiplier (120) and to determine the coefficients of the residual multiplier by comparing the actually determined values with the predicted values obtained as a result of the prediction, wherein the mapping onto binary code words is applied to the residual coefficients of the multiplier. 13. The apparatus according to claim 12, in which the processing circuit is further configured to generate prediction control information and add the prediction control information to the bitstream (330). 14. A device for decoding an encoded current set of samples (171) of an image, including a plurality of pixels, comprising a decoder (204, 210, 212, 214) for reconstructing the current set and a device according to any one of claims 1-7 for filtering the reconstructed set (215). 15. The device according to claim 14, wherein said processing circuit is further configured to obtain multiplier coefficients from binary code words included in the received bit stream (330) by applying a mapping operation. 16. The device according to claim 15, wherein said obtained multiplier coefficients are residual multiplier coefficients representing the difference between the actual values of the coefficients and the multiplier coefficients predicted according to the prediction circuit, and wherein the processing circuit is configured to determine the value of the multiplier coefficients of the loop filter (220) by reconstructing them from the obtained residual multiplier coefficients. 17. The apparatus of claim 16, wherein the prediction scheme is indicated by prediction control information further included in the received bitstream (330), and wherein the processing circuit is further configured to use the prediction control information in reconstruction. 18. The device according to any one of claims 15 to 17, in which the step of determining, using the processing circuit, further includes performing a determination whether the determined value of each multiplier coefficient, obtained directly from the received bit stream (330), by means of a mapping operation or by means of reconstruction from the obtained residual multiplier coefficients, is within the set of permissible values, and, if not, converting the determined value to the nearest value located in the set of permissible values. 19. A non-volatile machine-readable storage medium storing program instructions that, when executed by a processor, cause the execution of a method for filtering a set of image samples (115; 215) using a loop filter (120; 220) with adaptive multiplier coefficients represented by integers, containing steps in which the value of each filter multiplier coefficient is determined within a set of permissible values, wherein the set of permissible values consists of [-64, -32, -16, -8, -4, -2, -1, 0, 1, 2, 4, 8, 16, 32, 64], a set of image sample values is obtained, and the set of image samples is filtered using the loop filter.