JP2536684B2 - Image coding device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION (Industrial field of application) In a recording, transmission and display device for processing an image signal, a high-efficiency coding method for digitizing an image with a smaller amount of data, particularly adaptive quantization. The present invention relates to an image encoding device used.

(Prior Art) In high-efficiency image coding, there is adaptive quantization in which the step size of quantization is changed in block units according to its activity (activity and degree of change).

A block with high activity is difficult to detect even if the quantization error is large, and a block with low activity is likely to detect the quantization error.

Therefore, a block having high activity of the block has a wide quantization step width, and a block having low activity has a narrowing step width of the quantization.

As a result, the quantization is adapted to the visual characteristics and unnecessary fine quantization is eliminated, so that the amount of generated data can be reduced.

In encoding using orthogonal transformation, the processing block is generally 8 × 8 pixels or the like according to the size.

In such adaptive quantization, the information needs to be transmitted to the decoding side. Therefore, if there are many types of changes due to activity, the information will increase.

Therefore, it is general to set the types of changes by the adaptive processing to about four types as the quantization class.

This quantization class is a value in the intermediate stage where the quantization step is determined from the activity.

FIG. 7 is a block diagram showing a conventional image coding apparatus.

In FIG. 7, the image signal input from the image input terminal 1 is supplied to the orthogonal transformer 2.

The orthogonal transformer 2 orthogonally transforms the input signal for each block of 8 × 8 pixels by a technique such as DCT (discrete cosine transform).

One of the transform coefficients output from the orthogonal transformer 2 is a DC coefficient indicating the average value of the block and the other is an AC coefficient indicating the state of change, and each coefficient is supplied to the adaptive quantizer 3.

The adaptive quantizer 3 quantizes each coefficient with a set step width by a method described later and supplies it to the variable length encoder 4.

Adaptive quantization changes the quantization method depending on the nature of the image, but here, the step size of the quantization is changed by the block activity of the orthogonal transform coefficient.

The variable-length encoder 4 variable-length-encodes the quantized coefficient and outputs it as data from the data output terminal 5 to the decoding device.

Here, since the AC coefficients are concentrated near 0, the data length can be reduced by making the variable-length code of the AC coefficients have the shortest code length of 0 and the longer the absolute value, the longer the code length. .

On the other hand, the transform coefficient which is the output signal of the orthogonal transformer 2 is also supplied to the activity detector 6.

The activity detector 6 obtains the sum of absolute values of the AC coefficients of each block, and supplies it to the quantization class determiner 7 as an activity value A for each block.

 FIG. 8 is a diagram showing the characteristics of the quantization class determiner.

As shown in FIG. 8, the quantization class determiner 7 determines the quantization class value C based on the activity value A.

There are four class values C from 0 to 3, and the class value C increases by 1 every time A is doubled. The larger the number of classes, the more desirable it is in terms of characteristics, but the number of classes cannot be increased because the information needs to be transmitted.

The class information output from the quantization class determiner 7 is transmitted from the class information output terminal 8 to the decoding device side and also supplied to the adaptive quantizer 3.

 FIG. 9 is a waveform diagram showing a state of adaptive quantization.

9A shows the signal waveform before conversion, FIG. 9B shows the activity A, and FIG. 9C shows the quantization step Sq in the conventional example.

The signal waveform image before conversion is a transition between a flat portion and a changed portion as shown in FIG. 9 (A).

In activity A, the flat portion is small and the changed portion is large, so that the boundary block becomes large as shown in FIG. 9 (B).

If adaptive quantization is performed as it is, the quantization step Sq greatly changes for each block, and as shown in FIG. 9C, the boundary block has a large Sq change.

(Problem to be Solved by the Invention) In encoding using orthogonal exchange, a quantization error diffuses in a block when inversely transformed by a decoding device.

Then, in the edge portion of the image, the quantization error spreads to the periphery of the edge. This is called mosquito noise and causes visual deterioration in image quality.

When adaptive quantization is performed, the edge portion has high activity (activity), so that the quantization becomes rough, the quantization noise increases, and the mosquito noise increases.

Therefore, it has been considered that it is difficult to apply adaptive quantization in the conventional transform coding.

The smaller the adaptive quantization block is, the more appropriate processing is performed, but since the quantization class information needs to be transmitted for each block, there is a problem that the amount of data increases if the block is made smaller. .

Also, there is a problem that the number of kinds of quantization classes is similar and cannot be too large.

The present invention has been made by paying attention to the above points, and (1) LPF is used to correlate the quantization class value determined by the detected activity between blocks.
, The activity is reduced by the adjacent low activity block corresponding to the flat part in the edge part, the quantization is not excessively coarse in the adaptive quantization, and the quantum which has the visual characteristic without the increase of the mosquito noise. Image quality deterioration at the edge portion is improved, and (2) the class value is thinned out, and the class value of the block not transmitted is interpolated from the transmitted one, and the quantization step width is determined by the class value. Therefore, it is an object of the present invention to provide an image coding apparatus and a decoding apparatus that can reduce the amount of data to be transmitted because the information of the quantization step is thinned out.

(Means for Solving the Problems) In order to solve the above problems, the present invention detects (1) activity in each block unit in high-efficiency coding of an image in which the quantization step size is changed in block units. Activity detection means, quantization class determination means for determining a quantization class value from the activity, filter means for filtering the quantization class value over the values of surrounding blocks, and the filtered class value To provide a means for determining a quantization step width from a block, and (2) activity of each block unit in high efficiency coding of an image in which the quantization step size is changed in block units. And a quantity for determining the class value of quantization from the activity. Decimation class determination means, filtering means for filtering the quantized class value over the values of surrounding blocks, thinning means for thinning out the filtered class value in block units, and output signal of the thinning means And an interpolating means for outputting a signal obtained by interpolating the class value of the block which has been thinned out, and a means for determining the quantization step width from the output signal of the interpolating means. Is provided.

(Operation) The quantization class is judged not in each block unit but in a wider range so that the quantization does not become rough at the edge part.

Specifically, the quantization class determined on a block-by-block basis is passed through an LPF (low pass filter) in order to provide correlation between blocks.

Information of the quantization class to be transmitted is thinned out, and the quantization class of the block that is not transmitted is created by interpolating from the transmitted one.

Since the quantization class is correlated with the neighboring block, the class is lowered by the block of the lower class corresponding to the adjacent flat portion at the edge portion, and the quantization is not coarsened by the adaptive quantization.

Therefore, quantization suitable for visual characteristics can be performed without increasing mosquito noise.

Since the quantization step information is decimated, the amount of data that has to be transmitted is reduced. The number of classes increases due to filtering, which results in smooth changes.

(Embodiment) FIG. 1 is a block diagram showing a first embodiment of the image coding apparatus of the present invention. The same parts as those in FIG. 7 are designated by the same reference numerals.

In FIG. 1, the difference from FIG. 7 is that the class information, which is the output signal of the quantization class determiner 7, is mainly input to the adaptive quantizer 10 via the LPF (low pass filter) 9. And the control operation in the adaptive quantizer 10 is changed.

In FIG. 1, the class information which is the output signal of the quantization class determiner 7 is input to the LFP 9 and smoothed.

The process in LPF9 is a process in which the process for a normal LPF pixel value is replaced with a process for a class value for each block.

FIG. 2 shows the two-dimensional representation of each LPF tap coefficient.

With this tap coefficient, a squared cosine frequency characteristic is obtained both vertically and horizontally.

Here, the class value C input to the LPF 9 is four-valued, but the filtered class value C ′ is further multivalued and supplied to the adaptive quantizer 10.

The adaptive quantizer 10 quantizes the coefficient by multiplying the quantization step Sq determined by the class value C ′ by the control coefficient k.

k is a coefficient for controlling the amount of data, and is determined from the outside in order to obtain the desired amount of data. Therefore, it can be said that Sq is the relative value of quantization.

FIG. 3 is a diagram showing conversion characteristics of the class value C ′ and the quantization step Sq.

In FIG. 3, each time the class value C'increases by 1, the quantization step Sq is It increases one by one.

That is, each time the activity A doubles, the quantization step width becomes become.

 FIG. 9 is a waveform diagram showing a state of adaptive quantization.

FIG. 9 (A) is the signal waveform before conversion, FIG. 9 (B) is the activity A, FIG. 9 (C) is the quantization step Sq in the conventional example, and FIG. 9 (D) is the quantum in this embodiment. Represents the conversion step Sq.

The signal waveform image before conversion is a transition between a flat portion and a changed portion as shown in FIG. 9 (A).

In activity A, the flat portion is small and the changed portion is large, so that the boundary block becomes large as shown in FIG. 9 (B).

If adaptive quantization is performed as it is as in the conventional example of FIG.
The quantization step Sq greatly changes for each block, and a large Sq change occurs at the boundary block as shown in FIG. 9 (C).

What is filtered as in the present embodiment gradually changes, and as shown in FIG. 9D, the Sq change of the boundary block also becomes fine to some extent.

FIG. 4 is a block diagram showing a first embodiment of the image decoding apparatus of the present invention.

In FIG. 4, the compressed data transmitted from the data output terminal 5 of the encoding apparatus of FIG. 1 is supplied to the variable length decoder 12 via the data input terminal 11.

The variable length decoder 12 converts the variable length code into a normal code and supplies it to the inverse quantizer 13.

On the other hand, the class information input signal transmitted from the class information output terminal 8 of the encoder of FIG. 1 is an LP which operates in the same manner as the class information input terminal 14 and the LPF 9 of FIG.
It is supplied to the inverse quantizer 13 via F17.

The inverse quantizer 13 replaces this code with a representative value of quantization and supplies it to the orthogonal inverse transformer 15.

The step size of this replacement is determined by the class value C ′ input from the class information input terminal 14 as in the case of the adaptive quantizer 10 in the encoding device, according to the characteristic shown in FIG.

The quadrature inverse converter 15 performs inverse DCT conversion on the input signal, obtains a reproduced image signal, and outputs it through the image output terminal 16.

In addition, since the quantization class in the present invention has correlation between blocks by filtering, even if the information of all blocks is not transmitted, the result obtained by interpolation is not so different.

That is, the quantization class information may be thinned out, and the lost class may be interpolated from the remaining class.

FIG. 5 is a block diagram showing a second embodiment of the image coding apparatus according to the present invention. The same parts as those in FIG. 1 are designated by the same reference numerals.

5 is different from FIG. 1 mainly in that the processing in the LPF 18 is improved and the output signal of the LPF 18 is reduced by the decimator 19.
Is supplied to the class information output terminal 8 and the interpolator 20 via
The output signal of the interpolator 20 is supplied to the adaptive quantizer 10, and only different points will be described.

LPF18 is basically the same as LPF9 in FIG. 1,
The output class value C ′ is rounded into 8 types (3 bits) and supplied to the decimator 19.

The decimator 19 decimates the class value every other block in the horizontal and vertical directions, transmits it from the class information output terminal 8 to the decoding device side, and also supplies it to the interpolator 20.

Every other block in the horizontal and vertical directions, so 16
One class will be transmitted for each 16 pixels, 8 ×
This is 1/4 of the case of 8 pixels, and the data amount is 3/8 as compared with the conventional example and the first embodiment.

The interpolator 20 interpolates and creates the class value C ′ of the lost block.

Depending on the position of the block, the interpolation is 1/2 when C'is above or below or to the left or right, and is 1 when there are 4 diagonal directions.
Add / 4. The coefficient of the interpolation filter at that time is the second
It will be four times the one in the figure.

FIG. 6 is a block diagram showing a second embodiment of the image decoding apparatus of the present invention. The same parts as those in FIG. 4 are designated by the same reference numerals.

This is a decoding device corresponding to the coding device in the case of thinning out the classes shown in FIG. 5, and the difference from FIG.
The point is that the class value C ′ input from the class information input terminal 14 is supplied to the inverse quantizer 13 via the interpolator 20.

It goes without saying that the interpolator 20 performs the same interpolation operation as that of the encoding device shown in FIG.

(Effects of the Invention) The image encoding device and the decoding device of the present invention have extremely excellent effects as described below.

(1) LPF in order to correlate the quantization class value determined by the detected activity between blocks
In the edge part, the activity is reduced by the adjacent blocks with low activity corresponding to the flat part, the adaptive quantization does not make the quantization unnecessarily coarse, and the quantum which has the visual characteristic without the increase of mosquito noise. The image quality deterioration at the edge portion is improved.

Also, (2) the class value is thinned out, and the class value of the block that is not transmitted is interpolated from the transmitted one, and the quantization step width is determined by the class value, so the information of the quantization step is thinned out and transmitted. The amount of data that needs to be reduced can be reduced.

[Brief description of drawings]

FIG. 1 is a block diagram showing a first embodiment of the image coding apparatus of the present invention, FIG. 2 is a two-dimensional representation of each LPF tap coefficient, and FIG. 3 is a class value C'and quantum. Step Sq
FIG. 4 is a block diagram showing the first embodiment of the image decoding apparatus of the present invention, and FIG. 5 is a block diagram showing the second embodiment of the image encoding apparatus of the present invention. FIG. 6 is a block diagram showing a second embodiment of the image decoding apparatus of the present invention, FIG. 7 is a block diagram showing a conventional image encoding apparatus, and FIG. 8 shows the characteristics of the quantization class determiner. Shown, No. 9
The figure is a waveform diagram showing a state of adaptive quantization. 1 ... Image input terminal, 2 ... Orthogonal transformer, 3, 10 ... Adaptive quantizer, 4 ... Variable length encoder, 5 ... Data output terminal, 6 ... Activity detector, 7 ... Quantum Classifier, 8 …… Class information output terminal, 9,18 …… LPF, 11
...... Data input terminal, 12 …… Variable length decoder, 13 …… Inverse quantizer, 14 …… Class information input terminal, 15 …… Orthogonal inverse transformer, 16 …… Image output terminal, 19 …… Decimation Container, 20 ... Interpolator.

Claims (2)

(57) [Claims] 1. High-efficiency coding of an image in which a quantization step size is changed in block units. Activity detection means for detecting activity in each block unit, and a quantization class for determining a quantization class value from the activity. Image coding, comprising: determining means, filter means for filtering the quantization class value over the values of surrounding blocks, and means for determining a quantization step width from the filtered class value. apparatus. 2. In high-efficiency coding of an image in which the quantization step width is changed in block units, activity detection means for detecting activity in each block unit, and a quantization class for determining a quantization class value from the activity. Determination means, filtering means for filtering the quantization class value over the values of surrounding blocks, thinning means for thinning out the filtered class value in block units, and inputting the output signal of the thinning means However, the image coding apparatus has an interpolating unit that outputs a signal obtained by interpolating the class value of the block that has been thinned out, and a unit that determines the quantization step width from the output signal of the interpolating unit.