[go: up one dir, main page]

WO2017142360A1 - Procédé de codage et de décodage d'image, et codeur et décodeur d'image l'utilisant - Google Patents

Procédé de codage et de décodage d'image, et codeur et décodeur d'image l'utilisant Download PDF

Info

Publication number
WO2017142360A1
WO2017142360A1 PCT/KR2017/001800 KR2017001800W WO2017142360A1 WO 2017142360 A1 WO2017142360 A1 WO 2017142360A1 KR 2017001800 W KR2017001800 W KR 2017001800W WO 2017142360 A1 WO2017142360 A1 WO 2017142360A1
Authority
WO
WIPO (PCT)
Prior art keywords
dynamic range
range image
image
domain data
dct
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/KR2017/001800
Other languages
English (en)
Korean (ko)
Inventor
권오진
최승철
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Industry Academy Cooperation Foundation of Sejong University
Original Assignee
Industry Academy Cooperation Foundation of Sejong University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from KR1020170015839A external-priority patent/KR20170098163A/ko
Application filed by Industry Academy Cooperation Foundation of Sejong University filed Critical Industry Academy Cooperation Foundation of Sejong University
Priority to US15/999,734 priority Critical patent/US20190089955A1/en
Publication of WO2017142360A1 publication Critical patent/WO2017142360A1/fr
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
    • H04N19/17—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object
    • H04N19/172—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object the region being a picture, frame or field
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
    • H04N19/17—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object
    • H04N19/176—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object the region being a block, e.g. a macroblock
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
    • H04N19/179—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being a scene or a shot
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
    • H04N19/186—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being a colour or a chrominance component

Definitions

  • the present invention relates to an image encoding and decoding technique, and more particularly,
  • the present invention relates to a method of encoding and decoding HDR High Dynamic Range (JPEG) backward compatible, and to an image encoder and an image decoder using the same.
  • JPEG High Dynamic Range
  • An image can generally be represented by a limited number of bits that represent a limited range of values to represent a luminance signal.
  • the most common digital image formats currently in use use 24-bit (so-called 24-bit format) to store color and luminance information at each pixel in the image. For example, each value of red, green, and blue (Red, Green, and Blue) for a pixel may be stored in a range of 1 byte (8 bits). These images are called low dynamic range (LDR) images.
  • LDR low dynamic range
  • the brightness of light that can be detected by humans has a certain range.
  • the ratio of the darkest and the brightest light that can be detected is called the dynamic range.
  • the dynamic range In the 10-3 dynamic range of the brightness (luminance) that is easy for a person to recognize 10 cd / m 2 (candel a / m ') , while the conventional RGB color to use conventional expression by eight bits per
  • TMO tone-mapping operator
  • the legacy-JPEG (Legacy-JPEG) standard (ISO / IEC 10918) still dominates the photography market.
  • this standard does not support HDR images.
  • advanced image coding standards such as JPEG 2000 (IS0 / IEC 15444) or JPEG XR (ISO / IEC 29199) provide HDR image support, the adoption of HDR image coding by these standards is expected to be positive on the market. It is not becoming.
  • JPEG Commission (SC29WG1) recognizes that the main cause of this phenomenon is due to lack of backward compatibility with L-JPEG, which is already a tool chain in the market, a new image encoding called JPEG XT (ISO / IEC 18477). Standardization work has been initiated. Three profiles called profiles A, B and C have been proposed for JPEG XT.
  • An object of the present invention for solving the above problems is to provide an image encoding method compatible with backward JPEG.
  • Another object of the present invention for solving the above problems is to provide an image decoding method compatible with JPEG backwards.
  • Another object of the present invention for solving the above problems is to provide an image encoder that is backward compatible JPEG.
  • Another object of the present invention for solving the above problems is to provide an image decoder compatible with JPEG reverse direction.
  • An image encoding method for achieving the above object, the step of converting a first dynamic range image to a second dynamic range image, and encoding a second dynamic range image to generate a base layer code stream Deriving DC screte cosine transform (DCT) domain data for the second dynamic range image, deriving DCT domain data for the first dynamic range image, DCT for the second dynamic range image DCT diagram replacement paper for domain data and the first dynamic range image (rule 126) Deriving prediction coefficients related to the first dynamic range image from the main data, and using the DCT domain data for the second dynamic range image and the prediction coefficients related to the first dynamic range image, the prediction DCT for the first dynamic range image. Deriving domain data.
  • DCT DC screte cosine transform
  • the first dynamic range image may be an HDR image
  • the second dynamic range image may be an LDR image
  • Deriving the prediction coefficients related to the first dynamic range image may include using the correlation of the DCT domain data for the first dynamic range image with respect to the DCT domain data for the second dynamic range image.
  • the method may include calculating image related prediction coefficients.
  • the image encoding method uses at least one residual coefficient by using the DCT domain data for the second dynamic range image and the predictive DCT domain data for the dynamic range image derived from the prediction coefficients related to the thrust dynamic range image.
  • the method may further include generating and generating a residual layer codestream including the first dynamic range image related prediction coefficient and the at least one residual coefficient.
  • the generating of the base layer codestream may include converting the first layer dynamic range image into the second dynamic range image by performing a tone-mapping operation on the first dynamic range image.
  • Generating the base layer codestream also includes color transforming the second dynamic range image, DCT transforming the color transformed image, quantizing the DCT transformed image, and entropy the quantized image.
  • Alternative paper for encoding steps (Article 126) It may include.
  • the image quality coefficient used in the quantization of the DCT transformed image may be the same as the image quality coefficient used for quantization of the residual coefficient.
  • deriving DCT domain data for the second dynamic range image may include performing inverse quantization on the quantized DCT transformed image.
  • the AC coefficient of the DCT domain data for the first dynamic range image and the AC coefficient of the DCT domain data for the second dynamic range image are expressed as a function such as a polynomial, an exponential function, a logarithmic function, and a trigonometric function. It can have a correlation.
  • the DC coefficient of the DCT domain data for the first dynamic range image has a correlation represented by a prediction curve including a plurality of intervals with respect to the DC coefficient of the DCT domain data for the second dynamic range image, Each interval of the prediction curve may be defined by the same or different functions such as polynomials, exponential functions, logarithmic functions, trigonometric functions, and the like.
  • a method of decoding an image comprising: receiving a residual layer codestream including a first dynamic range image related prediction coefficient, receiving a base layer codestream, Decoding the received base layer code stream to generate a second dynamic range image; deriving DCT domain data for the second dynamic range image; deriving a residual DCT domain data; Image-related prediction coefficients and the second dynamic range already substituted (Article 126) Calculating predictive DCT domain data for the first dynamic range image from the DCT domain data for the digital image; adding the residual DCT domain data and the predictive DCT domain data for the first dynamic range image to add the DCT for the first dynamic range image. Reconstructing domain data, and decoding the DCT domain data for the first dynamic range image to generate a first dynamic range image.
  • the first dynamic range image may be an HDR image
  • the second dynamic range image may be an LDR image
  • Computing the predictive DCT domain data for the first dynamic range image may include deriving a first dynamic range image related prediction coefficient from the residual codestream and the DCT domain data for the second dynamic range image. 1 may include applying a function by the prediction coefficients related to the dynamic range image.
  • a method of decoding an image comprising: receiving a residual layer codestream including a first dynamic range image related prediction coefficient, receiving a base layer codestream, Deriving spatial domain data for a low 12 dynamic range image by performing an inverse -DCT lnverse Discrete Cosine Transform on the received base layer code stream, wherein the first dynamic range image ⁇ : series prediction coefficients and the first Calculating predictive spatial domain data for the first dynamic range image from spatial domain data for the dynamic range image, performing inverse-DCT transformation on the residual signal included in the residual layer codestream, and The spatial prediction data related to the first dynamic range image and the inverse-DCT transformed residual scene replacement paper (Rule 12) Article 6) Reconstructing the first dynamic range image from the call.
  • the first dynamic range image may be an HDR image
  • the second dynamic range image may be an LDR image
  • the step of calculating the predictive spatial domain data for the first dynamic range image may include deriving a first dynamic range image related prediction coefficient from the residual codestream and storing the predicted spatial domain data for the second dynamic range image. 1 1 may include applying a function by the prediction coefficients related to the dynamic range image.
  • an image encoder converts a first dynamic range image into a second dynamic range image and encodes a second dynamic range image to generate a base layer codestream.
  • a base layer processor to generate, an inverse quantizer for deriving DCT domain data by performing inverse quantization on the second dynamic range image quantized by the base layer processor, and deriving DCT domain data for the first dynamic range image
  • an enhancement layer processor for deriving a prediction coefficient related to the first dynamic range image from the DCT domain data for the second dynamic range image and the DCT domain data for the first dynamic range image.
  • the first dynamic range image may be an HDR image
  • the second dynamic range image may be an LDR image.
  • the enhancement layer processor is configured to utilize the correlation of the DCT domain data for the first dynamic range image with respect to the DCT domain data for the second dynamic range image to predict the first dynamic range image and the first dynamic range.
  • Example Alternative Paper for Images (Article 126) It may include a predictor for calculating side DCT domain data.
  • the enhancement layer processor may also generate at least one residual coefficient using the DCT domain data for the second dynamic range image and the prediction DCT domain data for the first dynamic range image derived from the prediction coefficients related to the first dynamic range image. And a residual layer codestream including the first dynamic range image related prediction coefficient and the at least one residual coefficient.
  • the image quality coefficient used for quantization of the second dynamic range image performed by the base layer processor may be the same as the image quality coefficient used for quantization of the residual coefficient performed by the enhancement layer processor.
  • the base layer processor may include a tone-mapping operator that performs a tone-mapping operation on a first dynamic range image and converts the second dynamic range image into a color. And a color converter for converting, a DCT converter for DCT converting the color converted image, a quantizer for quantizing the DCT converted image, and an entropy encoder for entropy encoding the quantized image.
  • an image decoder receives a base layer codestream, decodes the base layer codestream, and decodes DCT domain data for a second dynamic range image.
  • an image decoder receives a base layer codestream and performs entropy decoding, inverse-quantization, and inverse-DCT transform on the base layer codestream.
  • JPEG backward compatible HDR image encoding and decoding that utilizes the correlation of the HDR image to the tone-mapped LDR image in the DCT domain can be provided. It can improve the encoding and decoding performance.
  • FIG. 1 is a block diagram of a JPEG XT encoding system.
  • FIG. 2 is a block diagram of a block diagram of an HDR image encoder according to an embodiment of the present invention (rule 126) All.
  • FIG. 3 is a block diagram of an HDR image decoder according to an embodiment of the present invention.
  • FIG. 4 illustrates a plurality of image samples for explaining an experimental result of the HDR image encoding method and the decoding method according to an embodiment of the present invention.
  • FIG. 5 is an exemplary diagram illustrating a distribution of AC coefficients for various TM0s according to an embodiment of the present invention.
  • FIG. 6 is an exemplary diagram illustrating a distribution of DC coefficients for various TM0s according to an embodiment of the present invention.
  • FIG. 7 is a graph illustrating a concept of deriving a predictive HDR DC value according to another embodiment of the present invention.
  • FIG. 8 is a block diagram of an HDR image decoder according to another embodiment of the present invention.
  • FIG. 9 is a flowchart illustrating an encoding method according to an embodiment of the present invention.
  • FIG. 10 is an operation flowchart of a decoding method according to an embodiment of the present invention.
  • FIG. 11 is an operational flowchart of a decoding method according to another embodiment of the present invention. [Best form for implementation of the invention]
  • first, second, A, and B may be used to describe various components, but the components should not be limited by the terms. The terms are used only for the purpose of distinguishing one component from another.
  • the first component may be referred to as a second component without departing from the scope of the present invention, and similarly, the second component may also be referred to as the first component.
  • the JPEG XT encoding system shown in FIG. 1 can be applied for profiles A, B and C.
  • the JPEG XT encoding system includes a tone-mapping operator 10, an inverse -TMO (ll), a residual image generator 40, in addition to a legacy-JPEG encoder 20 and a legacy-JPEG decoder 30. And residual image encoder 50.
  • the JPEG XT encoding system including these detailed configurations outputs data of two layers, a base layer codestream and an enhancement layer, that is, a residual-layer codestream.
  • HDR images input to the JPEG XT encoding system are converted to tone-mapped LDR images by the tonemapping operator (10), color converters, DCT converters, and alternative paper (Article 126).
  • the enhancement layer (i.e., residual layer) codestream is a signal whose HDR image is output via the tonemapping operator (10), legacy -JPEG encoder (20), legacy -JPEG decoder (30), and inverse -TMO (ll). And a residual image generator 40 and a residual image encoder 50 that generate the residual image by using the HDR signal as an input.
  • the residual image encoder 50 and the quantizer, two image quality coefficients q and Q are used, respectively. Also, the choice of TM0 is given to the user, so any TM0 can be used with JPEG XT. Conversely, the TMO (ll) information may be included in the residual layer codestream used when reconstructing the HDR version of the residual layer decoder ⁇ LDR codestream.
  • 2 is a block diagram of an HDR image encoder according to an embodiment of the present invention.
  • An image encoder includes a tone mapping operator 100 and a legacy -JPEG encoder 200. Tone mapping operator 100 and legacy-JPEG encoder 200 may be referred to herein as a base layer processor.
  • the image encoder also does not include a legacy -JPEG decoder, unlike the JPEG XT encoder described with reference to FIG. 1, but instead includes a scaler 301, a color converter 310, a DCT converter 320. , Enhancement layer processor 300 and quantizer 330, entropy encoder 340, and HDR predictor 350, and de-quantization replacement sheet (Rule 126) And group 331.
  • a legacy -JPEG decoder unlike the JPEG XT encoder described with reference to FIG. 1, but instead includes a scaler 301, a color converter 310, a DCT converter 320. , Enhancement layer processor 300 and quantizer 330, entropy encoder 340, and HDR predictor 350, and de-quantization replacement sheet (Rule 126) And group 331.
  • An image encoder converts a first dynamic range image into a second dynamic range image, and encodes a second dynamic range image to generate a base layer code stream, wherein the base layer processor Inverse quantizer deriving DCT domain data by performing inverse quantization on the second dynamic range image quantized by D, and deriving DCT domain data for the first dynamic range image, and DCT for the second dynamic range image. And an enhancement layer processor for deriving a first dynamic range image related prediction coefficient from domain data and DCT domain data for the first dynamic range image.
  • the first dynamic range image may be represented using a larger amount of data than the second dynamic range image
  • the first dynamic range image may be an HDR image
  • the second dynamic range image may be an LDR image
  • the enhancement layer processor 300 is a predictor that calculates prediction coefficients related to the U dynamic range image by utilizing the correlation of the DCT domain data for the first dynamic range image with respect to the DCT domain data for the second dynamic range image.
  • the enhancement layer processor may also generate at least one residual coefficient using the DCT domain data for the first dynamic range image and the first dynamic range image related prediction coefficients, and the first dynamic range.
  • a residual layer codestream including the image related prediction coefficients and the at least one residual coefficient is generated.
  • Other JPEG backward-compatible HDR image coding may be implemented.
  • base layer encoding is applied in the same manner as the existing profile. That is, the HDR image input to the image encoder according to the present invention is tone-mapped by the tone mapping operator (TM0) 100 to be converted into an LDR image, a color converter 210, a DCT converter 220, a quantizer 230, constructs a base layer codestream compressed by legacy-JPEG encoder 200 including entropy encoder 240 and providing legacy-JPEG backward compatibility.
  • TM0 tone mapping operator
  • the tone mapping operator (100) may be referred to as dynamic range compression, and converts an image HDR image into an 8-bit LDR image by tone mapping an image HDR image without losing the features and details such as edge information from the original image. .
  • the color converter 210 converts the LDR image represented by RGB (Red-Green-Blue) to YCbCr.
  • DCT converter 220 performs an 8x8 block-based DCT transform on the image data represented by YCbCr.
  • DCT is one of techniques widely used for frequency conversion of an image and converts image data in the spatial domain into image data in the frequency domain using a cosine basis.
  • the resulting DC and AC components i.e., the DC coefficient and the AC coefficient are obtained.
  • the quantizer 230 receives the transform coefficient changed in the frequency domain by the DCT 220 as an input value and maps it to a discrete value. Data loss occurs during the quantization process, and continuous or large amounts of input data are mapped to a few discrete symbols after quantization.
  • entropy encoder 240 receives the output of quantizer 230 and performs entropy encoding.
  • entropy encoding alternative paper (Article 126) Is lossless compression, and is a process of minimizing the amount of data necessary for representation by variably allocating the length of a symbol according to the occurrence probability of the symbol.
  • the residual idling coding according to the exemplary embodiment of the present invention illustrated in FIG. 2 is clearly distinguished from the residual layer coding illustrated in FIG. 1.
  • the input HDR image is input to the scaler 301.
  • Scaler 301 scales the range of pixel values of the input HDR image to the LDR image range, where scaling is a uniform and reversible f loat ing-point scaling operation.
  • the color representation of the LDR image is converted into a YCbCr representation by the color converter 310 and an 8x8 block-based DCT is performed by the DCT converter 320.
  • One of the main features of HDR image coding proposed by the present invention is to perform HDR prediction based on the DCT coefficients of the tone-mapped LDR image encoded in the base layer, and each DCT coefficient of the input HDR image, and to estimate prediction coefficients and residuals. It is a configuration to generate a hierarchical code stream.
  • the HDR predictor 350 shown in FIG. 2 plays this role.
  • the HDR predictor 350 outputs the output of the inverse quantizer 331, which inversely quantizes the data output by the quantizer 230 in the encoder 200, that is, the DCT of the ton-mapped LDR image. Receive a coefficient as input. HDR predictor 350 also receives the DCT coefficients of the input HDR image as another input to derive the predictive HDR DCT coefficients and the prediction coefficients. The difference between the DCT coefficients of the input HDR image and the predictive HDR DCT coefficients forms a residual DCT coefficient. The residual DCT coefficients are quantized by quantizer 330 and entropy coded by entropy encoder 340.
  • quantizer 330 replacement paper (rule 126) Is the same as the quantizer 230 in the encoder, entropy encoder 340 and also performs the same role as the entropy encoder 240 in the encoder.
  • the image quality coefficient q used in the base layer may be used in the same way for the residual layer.
  • the finally generated residual layer codestream is composed of prediction coefficients estimated by the HDR predictor 350 and entropy coded residual DCT coefficients.
  • the DCT converter 220 of the base layer encoding process and the DCT converter 320 of the residual layer encoding process perform DCT conversion on the basis of blocks on the Y, Cb, and Cr color elements of the input image, and the resulting DCT coefficients. Are rearranged into one-dimensional vectors in zigzag order.
  • the k-th DCT coefficient in the first block of the input HDR image output by the DCT converter 320 is denoted by d, and the tone output by the de-quantizer 331.
  • the de-quantized DCT coefficients of the mapped LDR image are indicated by O). Also,
  • HDR predictor 350 is based on C ⁇ O) for each Y, Cb and Cr color element.
  • FIG. 1 is a block diagram of an HDR image decoder according to an embodiment of the present invention.
  • HDR image decoding which is a reverse operation of HDR image encoding, according to the present invention, can be described.
  • the HDR image decoder of the present invention may include a base layer decoder 400 and an enhancement layer decoder 500 that process legacy-JPEG compatible base layer codestreams.
  • the base layer decoder 400 receives the base layer codestream, decodes the base layer codestream, derives DCT domain data for the second dynamic range image, and generates a second dynamic range image.
  • the enhancement layer decoder 500 receives the residual layer codestream including the first dynamic range image related prediction coefficients, derives the first dynamic range image related prediction coefficients and the residual DCT domain data, and the first dynamic range image. in DCT-domain data for the relevant prediction coefficients and the shop second dynamic range image portion emitter said first calculates the predicted DCT domain data for the dynamic range image, the first dynamic range already and the predicted DCT domain data as to whether "residual DCT Sum domain data
  • Enhancement layer decoder 500 includes an HDR predictor 550, an entropy decoder 540, an inverse-quantizer 530, an inverse color transformer 510, and an inverse-scaler 501 that process a residual layer codestream. It may include.
  • base layer decoding is performed by a legacy legacy-JPEG decoder 400, and a legacy-JPEG decoder 400 includes an entropy decoder 410, an inverse-quantizer 420, and an inverse DCT converter 430. , And inverse-color converter 440.
  • the basic trade-off codestream input to the image decoder of FIG. 3 is converted into a quantized stream through an entropy decoder 410 and c DR j ⁇ transformed into a DCT domain via an inverse quantizer 420. .
  • Data in the DCT domain is converted into an LDR image, which is finally expressed in RGB, via an inverse DCT converter 430 and an inverse-color converter 440.
  • the prediction coefficients included in the residual layer codestream are input to the HDR predictor 550, and the residual tradeoff codestream passes through an entropy decoder 540 and an inverse-quantizer 530, which is a DCT coefficient of the residual signal ⁇ . Is converted to).
  • the HDR predictor 550 receives as input input prediction coefficients included in ⁇ ⁇ ) ( ⁇ ) and the residual layer codestream, and derives the predictive HDR DCT coefficients through HDR prediction.
  • the HDR predictor 550 located in the decoder uses a prediction coefficient included in the residual layer codestream to replace the encoder stage shown in FIG. 2 (rule 126).
  • the HDR predictor 350 derives the same HDR DCT coefficients as the predicted HDR DCT coefficients.
  • the DCT coefficients of the residual signal and the predicted HDR DCT coefficients c ⁇ M (A-) are summed to form f ⁇ in the form of a reconstructed HDR DCT coefficients, and the reconstructed HDR DCT coefficients are inverse2 "transformers (520), inverse- Finally, the image is finally restored to the HDR image through the color converter 510 and the inverse scaler 501. As described with reference to the embodiment of FIGS. From the point of view, there is a difference from the existing profile described in FIG.
  • profiles A and B generate their residual images in the form of an image divided by an HDR original image at each pixel in tone-mapped LDR images
  • profile C generates an HDR original image and a tone-mapped LDR image. Take the difference image as the residual image.
  • the present invention generates residual data in the DCT domain.
  • the L-JPEG decoding process is not required in the JPEG XT encoding according to the present invention, which means an effect of reducing the encoding time.
  • the existing prop ⁇ uses two quality factors.
  • Fig. 4 (a) shows the resultant image of uniformly quantizing the HDR sample image for display purposes and Fig. 5 (b) shows the tone-mapped LDR image using the TM0 technique proposed by Reinhard et al. .
  • TM0s were selected from among several selectable TM0 techniques, and the correlation between ⁇ and C ⁇ Ar) in the DC coefficient and AC coefficient was examined.
  • Alternative Paper (Article 126 of the Rules) 5 is an exemplary diagram illustrating a distribution of AC coefficients for various TM0s according to an embodiment of the present invention.
  • the five TM0 techniques used in FIG. 5 are expressed as "Reinhard02”, “Drago03”, “iCAM06”, “Mant iuk08” and "Mai ll”. Also, the image quality factor q was preset to 70. In experiments with different image quality factors, the same distribution was observed. Therefore, the effects of other image quality factors in designing the HDR predictor were negligible.
  • FIG. 5 shows the AC coefficient distribution for C ′ IDR (k) of c ⁇ O) for various TM0 techniques.
  • the horizontal axis means ⁇ O
  • the vertical axis means ⁇ '()
  • the ⁇ element is black
  • the Cb element is blue
  • the Ci- element is represented by red.
  • the correlation between the AC coefficients of the DCT domain data for the first dynamic range image and the AC coefficients of the DCT domain data for the second dynamic range image according to the present invention is not limited to the first-order polynomial function.
  • it may be expressed as a polynomial, an exponential function, a logarithmic function, a trigonometric function, and the like.
  • the AC coefficient related prediction may be performed for each of the Y, Cb, and Cr color elements, which may be defined by Equation 1 below.
  • aAC may mean a coefficient that minimizes a mean square error (MSE) between C ⁇ O) and C ⁇ A.
  • MSE mean square error
  • Alternative Paper (Article 126 of the Rules) 6 is an exemplary diagram illustrating a distribution of DC coefficients for various TM0s according to an embodiment of the present invention.
  • the five TM0 techniques used in FIG. 6 are “Reinhard02”,
  • the image quality factor q was also set to 70 in advance.
  • the DC coefficient of the image reflects the averaged pixel value in units of blocks and TM0 serves to improve the dynamic range of luminance.
  • the distribution can be interpreted as a global aspect of the reverse behavior of TM0 adopted for each image.
  • the distributions of Y, Cb and Cr for C ⁇ M (0) vs C ⁇ O) are different, they show a very high correlation of C ⁇ O) for c ⁇ A (0 ).
  • ce (O) for each color element of Y, Cb, and Cr is predicted by a cubic equation function of C ′ IDR (0) defined by Equation 2 below.
  • Equation 2 Alternative Paper (Article 126 of the Rules)
  • a) C, b, c and d may refer to coefficients that minimize the mean square error (MSE, mean square error) between c 'lmA and e ⁇ ). That is, HDR prediction according to an embodiment of the present invention may be performed by using a least square method.
  • MSE mean square error
  • FIG. 7 is a graph illustrating a concept of dividing a predicted HDR DC value into intervals according to another embodiment of the present invention.
  • the X axis is an LDR DC coefficient value
  • the y axis is a predictive HDR DC coefficient value
  • a range of values of the LDR DC coefficient is —1024 to 1023.
  • the coefficients ⁇ , b, c, and d of Equation 2 can be obtained using the least square method. From the point on the prediction curve defined by these coefficients, find the point (pl, p2) whose vertical distance from the point on the starting line and the end point of the prediction curve is the maximum in the positive and negative directions. Can be set as the reference point for dividing the section. If the cubic equation and the straight line do not meet, pl and p2 can be arbitrarily designated as -200, 200. However, it does not limit pl, p2 to -200 and 200.
  • the curve defined by the cubic equation is divided into three sections based on pl and p2, and the optimum prediction curve coefficients are extracted for each section.
  • Equation 3 One equation may be defined as Equation 3 below.
  • Equation 3 ⁇ ,, bi (C l , ⁇ are the coefficients for interval 1 (-1024 to pi) and a DCl are the coefficients for interval 2 (pi to p2), and in interval 3 (p2 to 1024) Coefficients.
  • FIG. 8 is a block diagram of an HDR image decoder according to another embodiment of the present invention.
  • FIG. 8 illustrates a decoder according to another embodiment different from the HDR image decoder according to the embodiment shown in FIG. 3, wherein the decoder shown in FIG. 8 spatially stores a residual layer codestream encoded using the HDR predictor in the DCT domain. Process on the domain. Therefore, the decoder according to the present embodiment replaces the residual data represented by the DCT domain with the spatial map (Article 26). Switch to main and perform HDR prediction in the spatial domain.
  • the HDR image decoder includes a decoder 400 for processing a legacy -JPEG compatible base layer codestream and a spatial domain predictor (for processing a residual layer codestream). 551, an enhancement layer decoder 500 including an entropy decoder 540, an inverse-quantizer 530, an inverse-color converter 510, and an inverse-scaler 501.
  • the base layer decoding is performed by the legacy legacy-JPEG decoder 400, and the legacy-JPEG decoder 400 includes an entropy decoder 410, an inverse-quantizer 420, and an inverse DCT converter ( 430, and an inverted-color converter 440.
  • the base layer codestream input to the image decoder of FIG. 8 is converted into a quantized stream through an entropy decoder 410 and an IDR (J) — transform represented by the DCT domain via an inverse quantizer 420.
  • c (A is converted to lDR (ik ⁇ ) via an inverse DCT converter 430 and converted to an LDR image that is finally expressed in RGB via an inverse-color converter 440.
  • the prediction coefficients included in the residual layer codestream are input to the spatial domain predictor 551, and the residual layer codestream is passed through an entropy decoder 540 and an inverse-quantizer 530, which is a low DCT coefficient of the residual signal. And the DCT coefficients of the residual signal are transformed to via an inverse-DCT converter 521. (w) is added to the output of the spatial domain predictor 551, input to the inverse-color converter 510, and finally reconstructed into an HDR image via the inverse-scaler 510.
  • the spatial domain predictor 551 receives the prediction coefficients included in the residual layer codestream received from the encoder and the inverse-DCT transformed base layer data ' lDR as input and performs HDR prediction.
  • Equation 4 IDCT ⁇ E '(ik ⁇ + IDCT ⁇ C! ⁇ () ⁇ where f Ot) is the residual signal and the DCT coefficient carried over the residual layer stream.
  • ⁇ () in Equation 4 an inverse -DCT operation may be performed on e i HD .
  • Equation 4 the second and third terms of Equation 4 are performed by performing inverse -DCT transform on HDR ov ⁇ , which is a signal predicted through the HDR predictor, and performing inverse -DCT transform. Expand to represent the value of the image. Therefore, the HDR image ⁇ reconstructed in the spatial domain may be generated as the sum of the reconstructed residual image and the reconstructed predictive HDR image.
  • Equation 4 is developed to reconstruct an HDR image in a spatial domain of a residual layer codestream encoded using an HDR predictor in a DCT domain.
  • Equation 5 _. ⁇ (0) —
  • Equation 4 the equation o ⁇ t) is decomposed into a DC component and an AC component as shown in Equation 6 below to calculate fflc ⁇ ) ⁇ .
  • nii means the i-th element of the 8x8 block containing m.
  • 1 ′ (input value) input to the inverse color converter 510 in FIG. 8 may be expressed by Equation 7 below.
  • ⁇ (, «) represents the result of performing IDCT on the residual signal
  • Equation 64 is a value calculated by the spatial domain predictor 551 as a result of predicting the HDR value by using the reconstructed LDR value mass (,?).
  • Equation 8 Equation 8
  • ⁇ 1 In one embodiment of the invention shown in FIG. 8, 1 is ("- ⁇ AC) and 5 is
  • FIG. 9 is a flowchart illustrating an encoding method according to an embodiment of the present invention.
  • An encoding method first converts a first dynamic range image into a second dynamic range image (S910), and encodes a second dynamic range image to generate a basic layer code stream (S920).
  • the first dynamic range image may be an HDR image
  • the second dynamic range image may be an LDR image.
  • step S920 of generating a base layer codestream may be performed by performing a tone-mapping operation on a first dynamic range image to convert the second dynamic range image to a second dynamic range image.
  • Color converting the range image DCT converting the color converted image, quantizing the DCT converted image, and entropy encoding the quantized image.
  • DCT domain data for the second dynamic range image is derived (S930), and DCT domain data for the first dynamic range image is derived (S940).
  • Deriving DCT domain data for the first dynamic range image (S940) includes scaling to the second dynamic range image data range, color converting the scaled image, and DCT converting the color converted image. It may include.
  • steps S930 and S930 are described as being sequentially executed, but two steps may be performed simultaneously, step S930 may be executed first, and step S940 may be performed later.
  • the two steps may be executed simultaneously or sequentially after the two steps may be changed depending on the characteristics of the steps described in FIG. 9.
  • a DCT domain data of the derived second dynamic range image and DCT domain data of the first dynamic range image are used to derive a prediction coefficient related to the first dynamic range image (S950).
  • Alternative Paper (Article 126 of the Rules)
  • the prediction coefficient related to the first dynamic range image may be calculated by using the correlation of the DCT domain data for the first dynamic range image with respect to the DCT domain data for the second dynamic range image.
  • the AC coefficients of the DCT domain data for the first dynamic range image and the AC coefficients of the DCT domain data for the second dynamic range image may include a linear polynomial, an exponential function, a logarithmic function, a trigonometric function, and the like. It can have a correlation expressed as a function.
  • the DC coefficient of the DCT domain data for the first dynamic range image and the DC coefficient of the DCT domain data for the second dynamic range image have a correlation represented by a prediction curve including a plurality of intervals.
  • Each interval of the prediction curve can be defined by the same or different functions such as polynomials, exponential functions, logarithmic functions, and trigonometric functions.
  • the prediction DCT domain data for the first dynamic range image is derived using the DCT domain data for the second dynamic range image and the prediction coefficients related to the first dynamic range image.
  • Generate one residual coefficient S960.
  • the residual coefficient may be a DCT coefficient
  • the first dynamic range may be defined as a difference value between the DCT domain data for the unknown and the predicted DCT domain data related to the first dynamic range image.
  • a residual layer codestream including a first dynamic range image-related prediction coefficient and the at least one residual coefficient is generated (S970).
  • the residual layer codestream may include prediction coefficients.
  • Alternative Paper (Article 126 of the Rules)
  • the base layer code stream and the residual layer code stream are transmitted to the decoder (S980). 10 is an operation flowchart of a decoding method according to an embodiment of the present invention.
  • the decoding method according to an embodiment of the present invention may be performed by the image decoder illustrated in FIG. 3, but the operation subject is not limited thereto.
  • the decoder receives a residual layer code stream including the first dynamic range image related prediction coefficients (S1010).
  • the decoder also receives a base layer codestream (S1020) and decodes the received base layer codestream to generate a second dynamic range image (S1030).
  • steps S1010 and S1020 are described as being sequentially executed, but two steps may be performed simultaneously, step S1020 may be executed first, and step S1010 may be performed later.
  • the two steps may be executed simultaneously or the order of two steps shown sequentially may be changed according to the characteristics of the steps described in FIG. 10.
  • the decoder derives DCT domain data for the second dynamic range image (S1040), and predicts the first dynamic range image from the prediction coefficients related to the first dynamic range image and the DCT domain data for the second dynamic range image. Derived DCT domain data (S1050).
  • the decoder finally reconstructs the first dynamic range image by converting the DCT domain data for the first dynamic range image (S1060).
  • DCT domain data for the first dynamic range image S1060
  • one dynamic alternative paper is processed through inverse-DCT conversion, inverse-color conversion, and inverse-scale. Convert DCT domain data for an image.
  • 11 is an operational flowchart of a decoding method according to another embodiment of the present invention.
  • the decoding method according to an embodiment of the present invention may be performed by the image decoder illustrated in FIG. 8, but the operation subject is not limited thereto.
  • the decoder receives a residual layer code stream including the first dynamic range image related prediction coefficients (S1110).
  • the decoder also receives a base layer codestream (S1120) and performs inverse-DCT transformation on the received base layer codestream to derive spatial domain data for the second dynamic range image (S1130).
  • steps S1110 and S1120 are described as being sequentially executed, but the two steps may be performed simultaneously, or step S1120 may be executed first and step S1110 may be performed later.
  • the two steps may be executed simultaneously or sequentially after the two steps may be changed according to the characteristics of the steps described in FIG. 11.
  • the decoder then calculates the first dynamic range image-related prediction spatial domain data from the first dynamic range image-related prediction coefficients and the spatial domain data for the second dynamic range image (S1140).
  • the decoder performs inverse -DCT transform on the residual signal included in the residual layer codestream (S1150), and extracts the first dynamic range image from the prediction space data for the first dynamic range image and the inverse -DCT transformed residual signal. Reconfigure (S1160).
  • the encoding performance can usually vary depending on the sample image and TM0 adopted.
  • three sample images of "01”, “02", and “03 '' and five TM0 techniques were selected, as shown in Figs. 5 and 6, and a total of 15 cases were tested. And the same conclusion was reached for the performance comparison.
  • Mantel et al. Evaluated the subjective image quality index of Signal_to-noi se rat io (SNR), Mean relative square error (MRSE), and HDR-VDP-2 to determine the subjective quality of HDR images compressed with JPEG XT profiles. And the result of the objective image quality comparison evaluation. This shows that the MRSE quality assessment index provides the most obvious results when using JPEG XT.
  • SNR Signal_to-noi se rat io
  • MRSE Mean relative square error
  • HDR-VDP-2 HDR-VDP-2
  • Cho i et al. Evaluated the performance of the JPEG XT profile by comparing the correlation between the coding performance and the various TM0 profiles using the PSNR quality evaluation index.
  • PSNR, SSIM, HDR-VDP-2 and MRSE-base SNR were selected for performance comparison, and the PSNR and MRSE-based SNR used in the present invention can be identified through the following notation and definition.
  • the program provided by the authors of each quality index was used for the evaluation when using SS IM and HDR-VDP-2.
  • M and N are the vertical and horizontal image size
  • an image encoding / decoding system to which the present invention is applicable is not limited to JPEG. That is, the present invention may be applicable to any system or apparatus as long as the MPEG system capable of encoding and decoding a video or the encoding or decoding system or apparatus for an image including a still image or a video. Operation of the encoding method and the decoding method according to an embodiment of the present invention can be implemented as a computer-readable program or code on a computer-readable recording medium.
  • Computer-readable recording media include all kinds of recording devices that store data that can be read by a computer system. Also computer readable
  • the computer-readable recording medium may include a hardware device specifically configured to store and execute program instructions, such as a ROM, a RAM, a flash memory, or the like.
  • Program instructions can include high-level language code that can be executed by a computer using an interpreter, as well as machine code such as that produced by a compiler.
  • While some aspects of the invention have been described in the context of a device, it may also represent a description according to the method in which the block or the device is characterized by a method step or a feature of the method step. Similarly, aspects described in the context of the method may also be characterized by the feature of the block or item being interacted with.
  • Some or all of the method steps may be performed by (or using) a hardware device such as, for example, a microprocessor, a programmable computer, or an electronic circuit. In some embodiments, one or more of the most significant method steps may be performed by such an apparatus.
  • a programmable logic device eg, a field programmable gate array
  • the field programmable gate array may operate in conjunction with a microprocessor to perform one of the methods described herein.
  • the methods are preferably performed by any hardware device.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Compression Or Coding Systems Of Tv Signals (AREA)

Abstract

L'invention concerne un codeur d'image comprenant : un processeur de couche de base permettant de convertir une première image à gamme dynamique en une seconde image à gamme dynamique, et de coder la seconde image à gamme dynamique de sorte à générer un flux de code de couche de base; un quantificateur inverse permettant de soumettre à une quantification inverse la seconde image à gamme dynamique quantifiée au moyen du processeur de couche de base, et de dériver des données de domaine DCT; et un processeur de couche d'amélioration permettant de dériver des données de domaine DCT pour la première image à gamme dynamique, et de dériver un coefficient de prédiction associé à la première image à gamme dynamique à partir des données de domaine DCT pour la seconde image à gamme dynamique et des données de domaine DCT pour la première image à gamme dynamique. La présente invention permet de mettre en œuvre le codage et le décodage, au moyen de la corrélation des données de première image à gamme dynamique et des données de seconde image à gamme dynamique, d'une image HDR présentant une rétrocompatibilité JPEG, de sorte à améliorer les performances de codage et de décodage.
PCT/KR2017/001800 2016-02-19 2017-02-17 Procédé de codage et de décodage d'image, et codeur et décodeur d'image l'utilisant Ceased WO2017142360A1 (fr)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US15/999,734 US20190089955A1 (en) 2016-02-19 2017-02-17 Image encoding method, and image encoder and image decoder using same

Applications Claiming Priority (6)

Application Number Priority Date Filing Date Title
KR20160019859 2016-02-19
KR10-2016-0019859 2016-02-19
KR10-2016-0064515 2016-05-26
KR20160064515 2016-05-26
KR10-2017-0015839 2017-02-03
KR1020170015839A KR20170098163A (ko) 2016-02-19 2017-02-03 이미지 부호화 및 복호화 방법, 이를 이용하는 이미지 부호화기 및 이미지 복호화기

Publications (1)

Publication Number Publication Date
WO2017142360A1 true WO2017142360A1 (fr) 2017-08-24

Family

ID=59626168

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/KR2017/001800 Ceased WO2017142360A1 (fr) 2016-02-19 2017-02-17 Procédé de codage et de décodage d'image, et codeur et décodeur d'image l'utilisant

Country Status (1)

Country Link
WO (1) WO2017142360A1 (fr)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20070026525A (ko) * 2004-04-23 2007-03-08 브라이트사이드 테크놀로지즈 아이엔씨. 높은 동적범위 (또는 고 명암비) 이미지의 엔코딩, 디코딩및 표시
US20080008235A1 (en) * 2006-07-10 2008-01-10 Segall Christopher A Methods and Systems for Conditional Transform-Domain Residual Accumulation
US20080193032A1 (en) * 2007-02-08 2008-08-14 Christopher Andrew Segall Methods and Systems for Coding Multiple Dynamic Range Images
KR20080107389A (ko) * 2006-01-23 2008-12-10 막스-플랑크-게젤샤프트 츄어 푀르더룽 데어 비쎈샤프텐 에.파우. 하이 다이나믹 레인지 코덱들
US20140205009A1 (en) * 2013-01-21 2014-07-24 The Regents Of The University Of California Method and apparatus for spatially scalable video compression and transmission

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20070026525A (ko) * 2004-04-23 2007-03-08 브라이트사이드 테크놀로지즈 아이엔씨. 높은 동적범위 (또는 고 명암비) 이미지의 엔코딩, 디코딩및 표시
KR20080107389A (ko) * 2006-01-23 2008-12-10 막스-플랑크-게젤샤프트 츄어 푀르더룽 데어 비쎈샤프텐 에.파우. 하이 다이나믹 레인지 코덱들
US20080008235A1 (en) * 2006-07-10 2008-01-10 Segall Christopher A Methods and Systems for Conditional Transform-Domain Residual Accumulation
US20080193032A1 (en) * 2007-02-08 2008-08-14 Christopher Andrew Segall Methods and Systems for Coding Multiple Dynamic Range Images
US20140205009A1 (en) * 2013-01-21 2014-07-24 The Regents Of The University Of California Method and apparatus for spatially scalable video compression and transmission

Similar Documents

Publication Publication Date Title
JP6245888B2 (ja) エンコーダおよび符号化方法
US8666186B1 (en) Lossy compression of high dynamic range video
US6987890B2 (en) Producing and encoding rate-distortion information allowing optimal transcoding of compressed digital image
Nageswara et al. Image compression using discrete cosine transform
JPH10327403A (ja) 圧縮条件の自動選択を伴う画像データの圧縮を最適化する方法
CN104702962A (zh) 帧内编解码方法、编码器和解码器
JP2016226001A (ja) デコーダおよび復号方法
US8396313B2 (en) Image compression and decompression using the PIXON method
US10148982B2 (en) Video compression using perceptual modeling
Jia et al. Layered image compression using scalable auto-encoder
CN102378991A (zh) 用于编码的数据中的压缩增益的压缩域系统和方法
Richter Backwards compatible coding of high dynamic range images with JPEG
Messaoudi et al. DCT-based color image compression algorithm using adaptive block scanning
JP5909149B2 (ja) 色変換装置、符号化器および復号器ならびにそれらのプログラム
US9787985B2 (en) Reduction of spatial predictors in video compression
RU2683614C2 (ru) Кодер, декодер и способ работы с использованием интерполяции
CN104282030A (zh) 图像压缩装置及方法
US20190089955A1 (en) Image encoding method, and image encoder and image decoder using same
Cui et al. Deep Network for Image Compressed Sensing Coding Using Local Structural Sampling
CN1937776A (zh) 图像压缩装置及方法
US10536697B2 (en) Method for re-encoding image and apparatus therefor
WO2025081929A1 (fr) Procédé de décodage d'image, procédé de codage d'image et appareil
WO2017142360A1 (fr) Procédé de codage et de décodage d'image, et codeur et décodeur d'image l'utilisant
Yu et al. Adaptive PQ: Adaptive perceptual quantizer for HEVC main 10 profile-based HDR video coding
JP4241517B2 (ja) 画像符号化装置及び画像復号装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 17753521

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 17753521

Country of ref document: EP

Kind code of ref document: A1