JP4253276B2 - Image coding method - Google Patents

Description

The present invention relates to an image coding how for moving image or a still image.

  In video coding standard systems such as ITU-T H.261, H.263, H.264, ISO / IEC MPEG-2, MPEG-4 Part 2, the prediction mode is switched in units of coding called macroblocks. Encoding is performed. It is said that how to perform this switching operation greatly affects the coding performance.

  One of the prediction mode switching methods is TMN9 (ITU-T, Study Group 16: “Video Codec Test Model, Near-Term, Version 9 (TMN9),” which was a test model of ITU-T H.263, for example. Document: Q15-C-15) (Non-Patent Document 1). In Non-Patent Document 1, threshold value processing is performed using an absolute value difference sum between an input image and a predicted image obtained by motion compensation, and an absolute value difference sum between an input image and a macroblock average of the input image. Switching between three prediction modes, that is, an inter-frame prediction mode of × 8 block, an inter-frame prediction mode of 16 × 16 block, and an intra-frame prediction mode is performed.

  On the other hand, Gary J. Sullivan and Thomas Wiegand, “Rate-Distortion Optimization for Video Compression”, IEEE Signal Processing Magazine, Vol.15, No.6, pp. 74-90, Nov. 1998. (Non-Patent Document 2) Introduces a method of selecting a prediction mode of each macroblock based on Lagrange's undetermined multiplier method. More specifically, after encoding is actually performed in each prediction mode to obtain a code amount (rate) and encoding distortion, a Lagrange's undetermined multiplier is assumed to be a function of the quantization parameter, and the same quantization parameter is used. The cost of each prediction mode is calculated and the prediction mode with the lowest cost is selected. It has been introduced that this technique significantly improves the coding efficiency compared to TMN9.

Also, TM5 (International Organization for Standardization, Test Model Editing Committee, 1993. Test Model 5. April. ISO-IEC / JTC1 / SC29 / WG11 / N0400), which was a test model of ISO / IEC MPEG-2. In 3), a method for correcting a quantization parameter considering visual characteristics is introduced. This is a technique for calculating the activity of the input image and correcting the quantization parameter value of the flat part to a small value in consideration of the property that human visual characteristics are sensitive to distortion of the flat part.
ITU-T, Study Group 16: “Video Codec Test Model, Near-Term, Version 9 (TMN9),” Document: Q15-C-15 Gary J. Sullivan and Thomas Wiegand, “Rate-Distortion Optimization for Video Compression”, IEEE Signal Processing Magazine, Vol.15, No.6, pp.74-90, Nov. 1998. International Organization for Standardization, Test Model Editing Committee, 1993. Test Model 5. April. ISO-IEC / JTC1 / SC29 / WG11 / N0400.

  The prediction mode switching method disclosed in Non-Patent Document 1 is a simple method, and the prediction mode can be switched with a small amount of calculation. However, since the actual code amount (rate) and distortion are not taken into consideration, it cannot be said that the optimum prediction mode is switched based on the relationship between the code amount and distortion. Therefore, a significant improvement in coding efficiency cannot be obtained.

  In the method of Non-Patent Document 2, since the code amount-encoding distortion function having the same gradient with respect to the same quantization parameter is assumed, the case where the quantization parameter is different or the form of the code amount-encoding distortion function is used. Is different from the assumption, there is a problem that correct comparison and evaluation of coding costs cannot be performed and an appropriate prediction mode is not selected. Also, it is not possible to compare the coding costs when the quantization parameters when combined with the code amount control are different. Therefore, a significant improvement in coding efficiency cannot be expected even with this method.

  Although the technique of Non-Patent Document 3 considers visual characteristics, the quantization parameter is directly controlled, so the relationship between visual characteristics and distortion is unclear. In addition, since mode selection is not taken into account, there is a drawback in that encoding efficiency is reduced.

An object of the present invention is to further improve the coding efficiency, it is to provide an image coding how visual characteristics also allows for selection of the prediction mode and the quantization parameter in consideration.

In order to solve the above-described problem, in one aspect of the present invention, an image is encoded while controlling a quantization parameter using one prediction mode selected from a plurality of prediction modes for each macroblock unit. in the encoding method, corresponding to that sign-amount into a plurality of combinations of a plurality of quantization parameters before and after the plurality of prediction modes and the reference parameter for each said macro block - generating a coding distortion pair set A step of calculating an undetermined multiplier from the reference parameter, a step of calculating each cost of the code amount-encoding distortion pair set using the undetermined multiplier, and a code amount-encoding distortion pair set. Selecting a code amount-encoding distortion pair with the minimum cost from the code amount, a prediction mode used for the encoding from the code amount-encoding distortion pair with the minimum cost. To provide an image coding method and a step of determining a quantization parameter.

The undetermined multiplier is calculated by, for example, λ = f × exp (gQ ₀ ) (where λ is an undetermined multiplier, Q ₀ is a reference parameter, and f and g are constants).

  When calculating the cost of each code amount-coding distortion pair set, it is desirable to weight the coding distortion in consideration of visual characteristics. More specifically, the cost of each code amount-coding distortion pair set is, for example, a value obtained by weighting the encoding distortion in consideration of visual characteristics and a value obtained by multiplying the code amount by the undetermined multiplier. Calculated as the sum.

  According to the present invention, by determining an appropriate prediction mode and quantization parameter, it is possible to perform image coding with high efficiency and considering visual characteristics and the like. In addition, since the prediction mode and the quantization parameter can be simultaneously optimized in consideration of the visual characteristics, it is possible to expect an improvement in coding efficiency in consideration of the visual characteristics.

Embodiments of the present invention will be described below with reference to the drawings.
FIG. 1 shows an image encoding apparatus for video encoding according to an embodiment of the present invention. For example, a moving image signal is input to the image encoding device in units of frames as the image signal 100. The subtracter 101 calculates the difference between the input image signal 100 and the predicted image signal 102 and generates a prediction error signal 103. The generated prediction error signal 103 is subjected to orthogonal transform, for example, discrete cosine transform (DCT) by the orthogonal transformer 104. The orthogonal transformer 104 obtains orthogonal transform coefficient information 105, for example, DCT coefficient information. The orthogonal transform coefficient information 105 is quantized by the quantizer 106 and then branched into two. One of the bifurcated quantized orthogonal transform coefficient information 107 is guided to the entropy encoder 108.

  The other of the bifurcated quantized orthogonal transform coefficient information 107 is subjected to sequential processing reverse to the processing of the quantizer 106 and the orthogonal transformer 104 by the inverse quantizer 109 and the inverse orthogonal transformer 110, and the prediction error signal. Then, the adder 110 adds the predicted image signal 102 to the local decoded image signal 112. The locally decoded image signal 112 is input to the frame memory / predicted image generator 108.

  The frame memory / predicted image generator 113 generates a predicted image signal from the input image signal 100 and the locally decoded image signal 112. Details of the frame memory / predictive image generator 113 will be described later. In the entropy encoder 108, the quantized orthogonal transform coefficient information 107 and the motion vector information / prediction mode information 114 are entropy-encoded, and each code generated thereby is multiplexed by the multiplexer 116, and then output buffer. It is smoothed by 117. Thus, the encoded data 118 output from the output buffer 117 is sent to a transmission system or storage system (not shown).

  The mode selector 121 directly selects a prediction mode by controlling a switch in the frame memory / predictive image creator 113 and controls a quantization parameter. Details of the mode selector 121 will be described later. The encoding controller 115 controls the encoding unit 115 including elements from the subtracter 101 to the frame memory / predictive image creator 113. In the present embodiment, code amount allocation for each encoding unit is performed while considering the buffer amount of the buffer 117. In this embodiment, the encoding unit is, for example, a macroblock unit, but is not limited to this.

(About the frame memory / predicted image generator 113)
FIG. 2 shows a specific example of the frame memory / predicted image generator 113 that generates the predicted image signal 102 from the input image signal 100 and the locally decoded image signal 112. The locally decoded image signal 112 from the adder 111 in FIG. 1 is temporarily stored in the frame memory 200. The motion vector detector 201 performs matching (block matching) between the input image signal 100 and the locally decoded image signal stored in the frame memory 200 for each block in the frame, and detects a motion vector.

  The inter-frame predictor 202 performs motion compensation on the locally decoded image signal in the frame memory 200 based on the motion vector detected by the motion vector detector 201, and creates a predicted image signal based on the inter-frame prediction. On the other hand, the intra-frame predictor 203 creates a predicted image signal based on intra-frame prediction from the input image signal 100 and a locally decoded image signal in a region already encoded in a frame in the frame memory 200.

  The inter-frame predictor 202 has K (K is plural) inter-frame prediction modes, and the intra-frame predictor 203 has L (L is plural) intra-frame prediction modes. A switcher 204 is connected to the outputs of the inter-frame predictor 202 and the intra-frame predictor 203. The switch 204 is controlled by the mode selector 121 to switch to a prediction image signal based on one prediction mode selected from K inter-frame prediction modes, or one prediction mode selected from L intra-frame prediction modes. Based on this, a predicted image signal is output.

  Motion vector information / prediction mode 114, that is, motion vector information output from motion vector detector 201 and prediction mode information indicating the prediction mode selected by switch 204 by mode selector 121 are sent to entropy encoder 108. It is done. The motion vector information is output from the motion vector detector 201 only when the inter-frame prediction mode is selected.

(About the encoding controller 119 and the mode selector 121)
FIG. 3 shows a flow of operations of the encoding controller 119 and the mode selector 121 in FIG. First, the encoding controller 119 controls the encoding unit 115 to cause the encoding unit 115 to sequentially perform encoding in a plurality of prediction modes (K inter-frame prediction modes and L intra-frame prediction modes). In addition, the quantization parameter is switched within each prediction mode.

  Accordingly, the mode selector 121 provides the code amount R and the coding distortion for each prediction mode and each quantization parameter based on the reference parameter which is provided from the coding control unit 119 for each macroblock and serves as a reference for the quantization parameter. A pair with D (hereinafter referred to as RD point) is obtained, and a code amount-coding distortion pair set (hereinafter referred to as RD point set) is generated (step S30).

  Next, the undetermined multiplier λ is calculated from the reference parameter (step S302). The undetermined multiplier λ will be described later.

  Next, based on the undetermined multiplier λ calculated in step S302, the cost of the combination of the prediction mode and the quantization parameter at each RD point generated in step S301 is calculated (step S304). Finally, a combination of a prediction mode and a quantization parameter used for encoding is selected from the RD points with the lowest cost calculated in step S304 in the RD point set generated in step S302 (step S302). 305).

Next, a specific example of the processing of steps S301 to S305 in FIG. 3 will be shown.
[Step S301]
First, the reference parameter Q ₀ given by the encoding controller 109 and the N quantization parameters Q ₀ -2N, Q ₀ -2N + 2,..., Q ₀ -2, Q ₀ , Q ₀ +2, every other point. .., Q ₀ + 2N−2, Q ₀ + 2N in total, 2N + 1 points of quantization parameters are encoded for each prediction mode, and a code amount R and encoding distortion D are obtained.

Here, for example, as shown in the following formula (1), the coding distortion D is obtained by the square error of each of the luminance signal Y and the color difference signals Cb and Cr for each macroblock.

Next, since the values of the quantization parameters that are actually encoded are every other point, the code amount R of the preceding and subsequent quantization parameters and the encoding distortion are determined for the points of the quantization parameter of the skipped values. Estimate from D points. It is assumed that the relationship represented by the following equations (2) and (3) is established between the code amount R and the encoding distortion D, where Q is the quantization parameter. a, b, c, and d are coefficients.

Coefficients a, b, c, and d assumed in equations (2) and (3) using code amounts R ₁ and R ₂ and encoding distortions D ₁ and D ₂ at two points before and after the actual encoding. Is estimated by the following equations (4) to (7).

From these four coefficients a, b, c, and d, the code amount R and coding distortion D for each quantization parameter skipped are estimated according to the following equations (7) and (8).

  FIG. 4 shows an example of the estimated values estimated by the measured values of the code amount R and the coding distortion D and the equations (8) and (9).

[Step S302]
In step S302, the reference parameter Q ₀ given from the encoding controller 109, for example, estimates the undetermined multiplier λ using the following equation (10).

  Here, f and g are constants determined by a prediction structure of a picture or a slice, such as a picture type or a slice type.

[Step S303]
In step S303, from the RD point set obtained in step S301, an RD point that minimizes the value of cost J calculated by the following equation (11) is selected.

Here, α is a weighting factor considering visual characteristics. As an example, the weighting factor α is given by the following equation (12) as a scale similar to step 3 of MPEG-2.

Where V _Y is the variance value of the input image of the current macroblock (currently the macroblock whose prediction mode and quantization parameter are to be determined), and ave_act is an image or image area unit such as a frame, field, or slice The average value of activity. A is a constant for scaling the value of the weight coefficient α to a value of 1 / A to A.

  When the variance of the input image of the current macroblock is smaller than the average, that is, when the current macroblock is a relatively flat region as a picture, the value of α is larger than 1, so that the coding distortion (for example, the square error) ) The value of αD is larger than other regions where D is the same. On the contrary, when the variance of the input image of the current macroblock is larger than the average, that is, when the current block is a relatively complex part as a picture, the value of α is smaller than 1, so the coding distortion D is the same. The value of αD is smaller than in other regions. That is, even if the value of the coding distortion D is the same, the weighting factor α weights D for the property that the distortion is more visually perceptible in a relatively flat area.

It is also possible to give the weighting coefficient α as a value in consideration of the visual characteristics regarding color as in the following formula (13).

Here, ΔE is the color difference in the CIE1976L * a * b * color space, assuming that an error of a constant k has occurred between the luminance and the color difference after encoding the current macroblock. Is obtained as the square of. Now, when the average value of luminance and color difference of the input image of the current macroblock is C _yuv = (Y _av , U _av , V _av ), the average value of luminance and color difference after encoding the current macroblock is C ′ _{Assume yuv} = (Y _av + k, U _av + k, V _av + k). At this time, if C _yuv and C ' _yuv are converted into colors in L * a * b * space, C _Lab = (L, a, b), C' _Lab = (L ', a', b ') and In this case, the square color difference ΔE is calculated according to the following equation (14).

  Here, ave_act is an average value of activities (act) in units of images or image areas such as frames, fields, or slices. A is a constant for scaling the value of the weight coefficient α to a value of 1 / A to A.

  The CIE1976L * a * b * space is a color space that is relatively close to the color difference felt by humans, and the perceived color difference increases as the distance in this space increases. Therefore, when the square color difference ΔE of the input image of the current macroblock is larger than the average, that is, when the current macroblock has a pattern in which distortion has a relatively large effect on color perception, the weight of Expression (13) Since the value of the coefficient α is larger than 1, the value of αD is larger than other regions having the same coding distortion D.

  On the contrary, when the square color difference ΔE of the input image of the current macroblock is smaller than the average, that is, when the current macroblock has a pattern in which distortion has a relatively small effect on color perception, the current macroblock is When the influence of distortion on color perception is relatively small, the value of the weighting coefficient α is smaller than 1, so that the value of αD is smaller than that in other regions where the coding distortion D is the same. In other words, even with the same square error (MSE) value, the weight of the encoded distortion D is weighted by the weighting coefficient α in such a way that the way the distortion is visually detected depends on the color.

  As shown in FIG. 6, the point of the same quantization parameter value Q in a certain prediction mode moves in the horizontal direction due to the variation of the weighting coefficient α in consideration of the visual characteristics. Therefore, if the weighting factor α is larger than 1 for the same λ, the prediction mode or the quantization parameter Q having a higher rate is selected. Also, if the weighting factor α is smaller than 1, the prediction mode or the quantization parameter Q having a lower rate is selected.

  In the above-described embodiment, two methods, ie, a method of weighting the coding distortion D by an activity representing the pattern of the input image and a method of weighting by the color are shown. Other methods using visual characteristics Needless to say, the coding distortion D can also be weighted using.

  In the above embodiment, an example in which a coding unit is a macroblock and a pair of a prediction mode and a quantization parameter is determined for each macroblock has been described. However, a coding unit may be a plurality of macroblock units, slices, fields, or the like. Other units such as a frame, a picture, or a GOP may be used.

  Furthermore, although the above embodiment has been described by taking moving image coding as an example, the present invention can also be applied to still image coding.

The block diagram which shows the structure of the image coding apparatus according to one Embodiment of this invention. The block diagram which shows the structure of the frame memory / predictive image creation device in FIG. The flowchart which shows the process sequence of the encoding controller and mode selector in FIG. The figure explaining the process of estimating code amount and coding distortion from actual measurement The figure which shows the process of selecting the code amount-encoding distortion pair with the minimum cost from the code amount-encoding distortion pair set. The figure which shows the change of the optimal point when weighting value alpha which considered the visual characteristic is changed

Explanation of symbols

100: Input video signal;
101 ... subtractor;
102 ... predicted image signal;
103 ... Prediction residual signal;
104 ... Orthogonal transformer;
105: orthogonal transform coefficient information;
106: Quantizer;
107: Quantized orthogonal transform coefficient information;
108 ... entropy encoder;
109 ... inverse quantizer;
110: inverse orthogonal transformer;
111 ... adder;
112 ... Locally decoded image signal;
113 ... Frame memory / predictive image creator;
114 ... motion vector information, prediction mode information;
115 ... encoding unit;
116: multiplexer;
117 ... output buffer;
118 ... encoded data;
119 ... Coding controller;
120 ... target code amount information;
121 ... mode selector;
200 ... frame memory;
201 ... motion vector detector;
202 ... inter-frame predictor;
203 ... intra-frame predictor;
204 ... Switcher

Claims (2)

In an image encoding method for encoding an image while controlling a quantization parameter using one prediction mode selected from a plurality of prediction modes for each macroblock unit,
Generating a code amount-coding distortion pair set respectively corresponding to a plurality of combinations of the plurality of prediction modes and a plurality of quantization parameters before and after the reference parameter for each macroblock;
Calculating an undetermined multiplier from the reference parameter;
Calculating the cost of each of the code amount-encoding distortion pair set using the undetermined multiplier;
Selecting the code amount-encoding distortion pair with the minimum cost from the code amount-encoding distortion pair set;
Determining a prediction mode and a quantization parameter used for the encoding from the code amount-coding distortion pair with the minimum cost ,
The step of calculating the cost includes a step of weighting the encoding distortion in consideration of visual characteristics . In an image encoding method for encoding an image while controlling a quantization parameter using one prediction mode selected from a plurality of prediction modes for each macroblock unit,
Generating a code amount-coding distortion pair set respectively corresponding to a plurality of combinations of the plurality of prediction modes and a plurality of quantization parameters before and after the reference parameter for each macroblock;
Calculating an undetermined multiplier from the reference parameter;
Calculating the cost of each of the code amount-encoding distortion pair set using the undetermined multiplier;
Selecting the code amount-encoding distortion pair with the minimum cost from the code amount-encoding distortion pair set;
Determining a prediction mode and a quantization parameter used for the encoding from the code amount-coding distortion pair with the minimum cost ,
The step of calculating the cost comprises: calculating a sum of a value obtained by weighting the coding distortion in consideration of visual characteristics and a value obtained by multiplying the code amount by the undetermined multiplier as the cost. .

Images