JP2003284077A - Apparatus and method for coding image, and image coding program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image coding apparatus, an image coding method, and an image coding program which suppresses the quantity of code generation to a minimum, and perform high-image-quality coding, when the stationary part of an image is coded by compression coding. <P>SOLUTION: A stationary frame coding judgment means 15 which judges whether or not a frame to be intended to code is a frame which is stationary (a stationary frame) from the difference value between the frame and a reference frame just before it, and a quantization parameter value control means 16 which controls quantization parameter values to be used for coding only for one frame at an initial time when its image begins to be stationary, are provided. Coding is performed whose coding error is smaller than that to be generated when other frames are coded. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image coding device,
The present invention relates to an image encoding method and an image encoding program, and in particular, specifies a still frame among encoded frames,
The present invention relates to a high-quality encoding of the frame.

[0002]

2. Description of the Related Art In order to efficiently store or transmit digital images, compression coding is required. As a method for compressing and coding a digital image, there are waveform coding methods such as subband, wavelet, and fractal in addition to the discrete cosine transform (DCT). Also, the image is
At the time of encoding, for example, in the case of an image having a similar part in one screen, redundant information in the screen or continuous display screens, by removing redundant information between screens such as a still part It is compressed and this is encoded.

As a method for removing the redundant information in the screen, there is a method called intra-screen prediction. In this method, the portion to be encoded is obtained as a difference signal by obtaining the difference from the already encoded adjacent portion, and this difference signal is waveform-encoded.

As a method of removing redundant image information between screens of adjacent frames, a method of inter-screen prediction using motion compensation, that is, the pixel value of a pixel of the current screen is compared with that of the previous screen. There is a method in which the difference from the pixel value of a pixel is obtained and expressed as a difference signal, and this difference signal is waveform-coded.

An example of a conventional image predictive coding apparatus using the above inter-picture prediction method will be described with reference to FIG.
FIG. 16 is a block diagram showing the configuration of a general image encoding device 1500.

In FIG. 16, an image encoding device 1500
Is a blocker 1 that divides an input image into macroblocks each having a size of 16 × 16 pixels, and subtraction means that obtains a difference value s1 between a target macroblock to be encoded and a prediction macroblock corresponding to the macroblock. 2, the switch 3 that turns ON / OFF based on the determination result of the skip macroblock determiner 4, the absolute value of the difference value s1 and the quantization parameter given from the outside, and the macroblock is displayed on the immediately preceding screen. Of the same position, the skip macroblock determiner 4 that determines whether to be the skip macroblock to be replaced or the difference value s1 is encoded, and switches the switch 3, and the compression code that compresses and encodes the difference value s1. Localizing means 5, local decoding means 9 for performing local decoding based on the quantized coefficient output from the compression coding means 5, and local decoding Is obtained by a local predictive image generation means 12 for generating a predictive macroblock from the output of stage 9.

The compression coding means 5 comprises a frequency converter 6 for converting the difference value s1 into a frequency domain signal, and a frequency converter 6
Quantizer 7 which quantizes the output of the above using an externally input quantization parameter, and variable length coding of the output (quantization coefficient) of the quantizer 7, or the skip macroblock determiner 4 A skip length macroblock signal s2 which is an output is provided with a variable length encoder 8 which multiplexes the differential signals of other macroblocks with variable length encoded data and outputs the multiplexed signal as a code string.

The local decoding means 9 dequantizes the output (quantization coefficient) of the quantizer 7 and the dequantizer 10 and the output (frequency domain signal) of the dequantizer 10 into a spatial domain signal. The frequency inverse converter 11 for converting is provided.

The local predicted image generating means 12 stores a frame memory 14 for storing data obtained by adding the output (locally decoded data) of the local decoding means 9 and the output (predicted macroblock) of the predicted image generating means 13, And a prediction image generator 13 for generating a prediction macroblock.

The image coding apparatus 1 configured as described above
The operation of 500 will be described. The input image is divided by the blocker 1 into macroblocks. Then, the difference value s between the target macroblock to be the target of the encoding process and the prediction macroblock corresponding to the target macroblock
1 is obtained by the subtracting means 3, and the skip macroblock determiner 4 determines whether or not the macroblock is a skip macroblock using the absolute value of the difference value s1 and the quantization parameter given from the outside. Will be held in.

Here, the skip macroblock determiner 4
The processing method of will be described with reference to FIG. First, the difference value s1 of the macroblock output from the subtraction unit 2
The absolute value sum X is calculated (step S1600), and the calculated absolute value sum X is compared with the threshold value THR_L obtained by multiplying the quantization parameter by a constant value (step S1601). If the absolute value sum X is larger than the threshold value THR_L, step S1
Proceeding to 602, it is determined that the macro block is an inter macro block, and the difference value s1 is encoded. If the sum of absolute values X is smaller than the threshold value THR_L, step S1
In step 603, it is determined that the macro block is a skip macro block. That is, the macroblock determined to be the skip macroblock is a case where there is little motion or no motion (stationary) in the immediately preceding frame, the macroblock to be encoded in the frame, and the vicinity thereof.

When the skip macroblock determiner 4 determines that the macroblock is not a skip macroblock, the difference value s1 is output to the compression coding means 5 by the switch 3, and the difference value s1 is compressed compression means 5.
Is converted into a frequency domain signal by the frequency converter 6 in the
The frequency domain signal is quantized by a quantizer 7 using a quantization parameter input from the outside. Further, the output (quantization coefficient) from the quantizer 7 is variable-length coded in the variable-length encoder 8. The output of the quantizer 7 is locally decoded by the local decoding means 9, and the locally decoded data is the locally predicted image generating means 1
2 is used to generate a predicted image (predicted macroblock).

If the skip macroblock determiner 4 determines that the macroblock is a skip macroblock, the skip macroblock determiner 4 sends the skip macroblock signal s to the variable length encoder 8.
2 is output, and the variable length encoder 8 multiplexes the differential signals of the other macro blocks with the variable length encoded data,
It is output as a code string. In this way, a still image can be determined while encoding the image, and the determined still image can be encoded with high image quality.

The image coding method described with reference to FIG. 16 is a one-pass coding (one-pass coding) method in which the image signal to be coded is input only once to the coding apparatus. According to this, the coded signal can be coded immediately after input, and real-time coding can be performed.

Also, unlike the above-mentioned one-pass coding, there is also a two-pass coding (two-pass coding) method in which an image signal to be coded is once coded and then coded again using the coded information. According to the 2-pass coding, it is possible to perform the coding using the first information, but the time required for the coding is longer than that in the 1-pass coding.

[0016]

[Patent Document 1] Japanese Patent Laid-Open No. 2000-188735

[Patent Document 2] Japanese Patent Laid-Open No. 2000-125301

[0017]

However, in the above-mentioned conventional method, since the entire image is coded, there arises a problem that the generated code amount increases. Therefore, if the difference value is encoded by using the inter-frame prediction used from the conventional method, or if the encoding is performed by replacing the macro block that is the skip macro block with the macro block of the immediately preceding frame, The amount of generated bits in encoding is smaller than that in encoding that does not perform inter-frame prediction, and in the case of a still image or an image with many still parts, particularly when encoding using inter-frame prediction, a large compression rate is achieved. The effect is obtained.

Further, according to the above-mentioned conventional method, in a still image or an image having many still portions, an inter-screen residual signal is hardly generated in a frame other than the first frame of the still image,
There is a problem that the number of skip macroblocks that refer to the immediately preceding frame increases. At this time, if the image quality deterioration (encoding error) due to the encoding of the first still frame is sufficiently small, the image quality of the subsequent still frame referring to this can be expected to improve, but if the encoding error of the first still frame is large, The coding error affects the entire still part that refers to this, and the image quality deterioration is not improved. In order to prevent such image quality deterioration, it is necessary to reduce the number of skip macroblocks or to reduce the quantization width for quantization.

The present invention has been made in order to solve the above problems, and in the coding of a still image or a still part, an image coding apparatus for preventing an increase in the generated code amount and coding with high image quality. , An image encoding method, and an image encoding program.

[0020]

In order to solve the above-mentioned problems, the image coding method according to claim 1 of the present invention divides a coded frame to be coded into macroblocks which are processing units. A still frame determination method for determining whether or not the encoded frame is a still frame, which is an image encoding method for obtaining a difference value between the frame and a reference frame and encoding the difference value And a high image quality code for performing still frame encoding in which at least one frame among the frames determined to be still frames in the still frame determination step has higher image quality than the encoding of other frames. And a converting step. This allows
By specifying the encoded frame that is a still frame and performing high image quality encoding on at least one frame, it is possible to increase the amount of generated code by encoding all still frames with high image quality as in the conventional case. The image quality can be suppressed and the quality of the subsequent frames can be improved.

An image coding method according to claim 2 of the present invention is the image coding method according to claim 1, wherein
The still frame determination step is characterized in that the determination is performed based on a difference value between the encoded frame and the reference frame. As a result, the next frame of the frame that has started to be stationary can be identified, and high quality coding can be performed on the next frame, and as a result, the subsequent frames also have high quality.

An image coding method according to claim 3 of the present invention is the image coding method according to claim 1, wherein
The high image quality encoding step includes a still frame encoding determination step of determining whether to perform still frame encoding on the encoded frame determined to be the still frame, or to perform normal frame encoding, In the quantization parameter value control step for controlling the quantization width of the quantization parameter input from the outside based on the encoding method determined in the still frame encoding determination step, and in the quantization parameter value control step A skip macroblock determining step of determining whether each macroblock of the encoded frame is a skip macroblock using the controlled quantization parameter and the difference value, and the skip macroblock determining step Based on the judgment result of each macro block obtained in A coding step of coding arm, in which characterized in that it comprises a. This allows
The number of macroblocks that are determined to be skip macroblocks can be suppressed, and as a result, it is possible to prevent an increase in the amount of generated code and prevent encoded frames that have been determined to be still frames without degrading image quality. It can be encoded with high image quality.

An image coding method according to claim 4 of the present invention is the image coding method according to claim 1, wherein
The high image quality encoding step includes a still frame encoding determination step of determining whether to perform still frame encoding on the encoded frame determined to be the still frame, or to perform normal frame encoding, When it is determined in the still frame coding determination step that the coded frame is a frame in which still frame coding is not performed, a difference value between the coded frame and a reference frame and a predetermined threshold value are used. It is determined whether or not each macroblock of the encoded frame is a skip macroblock, and it is determined in the still frame encoding determination step that the encoded frame is a frame for still frame encoding. In this case, the predetermined threshold is controlled to be small, and the difference between the controlled threshold and the encoded frame and the reference frame is A skip macro-block determination step of determining whether each macroblock of 該被 coded frame is a skip macro-block by using a value,
An encoding step of encoding the frame to be encoded based on the determination result of each macroblock obtained in the skip macroblock determination step. As a result, the number of macroblocks that are determined to be skip macroblocks can be suppressed, and as a result, it is possible to prevent an increase in the amount of generated code and prevent a coded code that is determined to be a still frame without degrading image quality. The encoded frame can be encoded with high image quality.

The image coding method according to claim 5 of the present invention is the image coding method according to claim 1, wherein
Whether the next frame of the coded frame based on the coding information of the coded frame determined in the coding of the preceding frame of the coded frame is a frame for still frame coding, or Still frame coding determination step of determining whether or not the frame is a frame to be frame-encoded, and control of quantization width of quantization parameter input from the outside based on the encoding method determined in the still frame determination step Whether each macroblock of the coded frame is a skip macroblock by using the quantization parameter value control step for performing the quantization parameter value control step, the quantization parameter controlled in the quantization parameter value control step, and the difference value. The skip macroblock determination step for determining whether or not there is the still frame coding determination step Determines the encoding method of the next frame of the encoded frame based on the determination result of, and performs the encoding of the encoded frame based on the encoding information determined in the encoding of the previous frame And an encoding step. As a result, it is possible to perform high image quality encoding with less quantization error.

An image coding method according to claim 6 of the present invention is the image coding method according to claim 1, wherein
The still frame determining step is performed based on a coding type that represents a coding method of the macroblock of each macroblock of the reference frame. As a result, it is possible to reduce the calculation amount in the still frame determination of the encoded frame,
In addition, the frame two frames after the frame that has started to be stationary is specified, and only that frame is encoded with high image quality. As a result, the generated code amount can be suppressed more than before, and the subsequent frames also have high image quality. Becomes

An image coding method according to claim 7 of the present invention is the image coding method according to claim 6, wherein
The coding type of each macroblock includes a first flag indicating whether the coded macroblock is a skip macroblock, and a coded frame when the coded macroblock is a skip macroblock. Is a ternary flag indicating whether or not is the still frame coding, and the coded macroblock is coded according to the flag value of the macroblock referred to by the coded macroblock. Flag is to be updated. As a result, once the macroblock that has been encoded with high image quality is determined to be a skip macroblock without being encoded with high image quality in the subsequent processing,
As a result, only the first frame can be encoded with high image quality, and the amount of calculation in the still frame determination of the encoded frame can be suppressed.

An image coding method according to claim 8 of the present invention is the image coding method according to claim 1, wherein
The high image quality encoding step is characterized in that the high image quality encoding is not performed on the I frame for which only intra-picture predictive encoding is performed. As a result, still frame coding is not performed for a frame for which intra-frame predictive coding is performed, and as a result, it is not necessary to reduce the quantization width,
It is possible to further suppress an increase in the generated code amount.

An image coding method according to claim 9 of the present invention is the image coding method according to claim 1, wherein
The high image quality encoding step is characterized in that the high image quality encoding is not performed on the macroblock for which the intra-picture predictive encoding is performed. As a result, still frame coding is not performed for macroblocks for which intra-picture predictive coding is performed, and as a result, there is no need to reduce the quantization width, and the increase in the amount of generated code can be further suppressed.

According to a tenth aspect of the image encoding method of the present invention, the encoded frame to be encoded is divided into macroblocks, which are processing units, and the difference value between the frame and the reference frame is obtained. An image encoding method for encoding the difference value, the still macroblock determining step of predicting whether or not each macroblock of the encoded frame is still, and the still macroblock determining step The macroblock coding step of coding a macroblock determined to be still in 1. with higher image quality than the coding of other macroblocks is included. This allows
It is possible to suppress the generation of the code amount and perform encoding with high image quality.

An image coding method according to claim 11 of the present invention is the image coding method according to claim 1, wherein the high image quality coding step is a frame to be coded judged to be the still frame. In the still macroblock determination step of predicting whether or not each macroblock is still in the screen, and for the macroblock determined to be still in the still macroblock determination step, other macroblocks And a macroblock coding step for coding with high image quality as compared with coding. As a result, it is possible to perform high-quality coding with few coding errors due to fluctuations in the quantization parameter even in a frame in which still frame coding is performed.

The image coding apparatus according to claim 12 of the present invention blocks the coded frame to be coded into macroblocks as a processing unit, and calculates the difference between the coded frame and the reference frame. An image coding apparatus that obtains a value and encodes the encoded frame based on the difference value, and determines whether the encoded frame is a still frame based on the difference value. If the frame to be coded, which is determined to be the still frame, is a frame to be subjected to still frame coding which is coded with higher image quality than the coding to other frames, or is subjected to normal frame coding. Still frame coding determining means for determining whether the frame is a frame, and the amount of quantization parameter input from the outside based on the output signal of the still frame coding determining means. By using the quantization parameter value control means for controlling the quantization width, the quantization parameter controlled by the quantization parameter value control means and the difference value, each macroblock of the encoded frame is a skip macroblock. It is characterized by further comprising: skip macroblock determining means for determining whether or not the encoding target frame is encoded based on an output of the skip macroblock determining means. .
As a result, the number of macroblocks determined to be skip macroblocks can be suppressed, and the frame next to the frame that has started to be stationary can be encoded with high image quality.
It is possible to suppress the amount of generated codes as compared with the conventional technique and improve the image quality of subsequent frames.

An image coding apparatus according to a thirteenth aspect of the present invention is the image coding apparatus according to the twelfth aspect, wherein the coding means is a quantization controlled by the quantization parameter value control means. The above-mentioned encoding is performed using a parameter. As a result, it is possible to perform high image quality encoding with less quantization error.

According to a fourteenth aspect of the present invention, the image coding apparatus blocks the coded frame to be coded into macroblocks, which are processing units, and refers to the coded frame and the coded reference. An image coding apparatus that obtains a difference value from a frame and encodes the coded frame based on the difference value, and determines whether the coded frame is a still frame based on the difference value. Whether the frame to be coded, which is determined to be a still frame, is a frame for performing still frame encoding for encoding with higher image quality than other frames, or a frame for performing normal frame encoding Is determined using the still frame coding determination means, the output of the still frame coding determination means, and the difference value. A skip macroblock determining unit that determines whether or not the block is a skip macroblock, and an encoding unit that encodes the encoded frame based on the output of the skip macroblock determining unit. It is characterized by that. As a result, the number of macroblocks that are determined to be skip macroblocks can be suppressed, and the frame next to the frame that has started to be stationary can be encoded with high image quality. As a result, the generated code amount can be suppressed more than before. And the subsequent frames also have high image quality.

An image coding apparatus according to a fifteenth aspect of the present invention is the image coding apparatus according to the fourteenth aspect, wherein the quantum input from the outside based on the output signal of the still frame coding determination means. A quantization parameter value control means for controlling the quantization width of the quantization parameter, wherein the encoding means encodes the encoded frame using the quantization parameter controlled by the quantization parameter value control means. It is characterized by performing. As a result, it is possible to perform high image quality encoding with less quantization error.

According to a sixteenth aspect of the present invention, the image coding apparatus blocks the coded frame to be coded into macroblocks which are processing units, and the difference between the coded frame and the reference frame. An image coding apparatus that obtains a value and encodes the encoded frame based on the difference value, and performs intra-screen prediction as to whether or not each macroblock of the encoded frame is stationary. , A skip macroblock determining unit that determines the macroblock as a still macroblock when a macroblock in the vicinity of the macroblock is a skip macroblock, and a macro that is determined as a still macroblock by the skip macroblock determining unit. Make the quantization parameter value of the block smaller than the quantization parameter value input from the outside. The quantization parameter value control means to be controlled and the macroblock determined to be the still macroblock using the quantization parameter controlled by the quantization parameter value control means are higher than those of other macroblocks. And a coding means for coding in image quality. As a result, it is possible to suppress the generation of the code amount and perform encoding with high image quality.

An image coding apparatus according to claim 17 of the present invention divides a coded frame to be coded into macroblocks as a processing unit and obtains a difference value between the frame and the reference frame. An image encoding device for encoding the difference value, wherein the next frame of the encoded frame is based on the encoding information of the encoded frame determined in the encoding of the previous frame of the encoded frame. Still frame coding determining means for determining whether the frame is a frame for performing still frame coding for coding with higher image quality than that for coding for other frames, or a frame for performing normal frame coding. A quantization parameter value control means for controlling the quantization width of a quantization parameter input from the outside based on the output of the still frame determination means, Using the difference value and the quantization parameter controlled by the quantization parameter value control means, skip macroblock determination means for determining whether each macroblock of the encoded frame is a skip macroblock, The encoding method of the next frame of the encoded frame is determined based on the determination result of the still frame encoding determination means, and the encoded frame is determined based on the encoding information determined in the encoding of the previous frame. And encoding means for performing encoding of. As a result, the number of macroblocks that are determined to be skip macroblocks can be suppressed, and the frame next to the frame that has started to be stationary can be encoded with high image quality. As a result, the generated code amount can be suppressed more than before. And the subsequent frames also have high image quality.

The image coding apparatus according to claim 18 of the present invention is the image coding apparatus according to claim 17, wherein the coding means is a quantization controlled by the quantization parameter value control means. The above-mentioned encoding is performed using a parameter. As a result, it is possible to perform high image quality encoding with less quantization error.

An image coding apparatus according to a nineteenth aspect of the present invention is the image coding apparatus according to the seventeenth aspect, wherein the skip macroblock determining means encodes the encoded frame from the outside. Starting the determination processing of the skip macroblock by inputting a frame encoded signal indicating a method, the encoding means, when the frame encoded signal is a signal indicating an intra-frame prediction encoding method,
Each of the macroblocks of the frame to be coded is coded, and when the frame coded signal is not a signal indicating the intra-frame predictive coding method, the difference value is coded. As a result, still frame coding is not performed for frames and macroblocks for which intra-frame predictive coding is performed, and as a result, there is no need to reduce the quantization width, and the increase in the amount of generated code can be further suppressed. it can.

Further, an image coding apparatus according to claim 20 of the present invention divides a coded frame to be coded into macroblocks which are processing units, and obtains a difference value between the frame and the reference frame. An image encoding apparatus for encoding the difference value, wherein whether each macroblock of the encoded frame is a skip macroblock using the difference value and a quantization parameter input from the outside A skip macroblock determining means for determining whether or not the next frame of the encoded frame is based on the encoding information of the encoded frame determined in the encoding of the preceding frame of the encoded frame, It is a frame that performs still frame encoding that encodes with higher image quality than frame encoding, or a frame that performs normal frame encoding. And the coding method of the next frame of the coded frame based on the judgment result of the still frame coding judging means The encoding is characterized by comprising: an encoding unit that performs the encoding based on the encoding information determined in the encoding of the previous frame. As a result, it is possible to easily realize the determination process of whether or not to encode the next frame code as a still frame, suppress the number of macroblocks determined to be skip macroblocks, and reduce the number of Two
The frame after the frame can be encoded with high image quality, and as a result, the generated code amount can be suppressed more than before, and the subsequent frame also has high image quality.

An image coding apparatus according to a twenty-first aspect of the present invention is the image coding apparatus according to the twentieth aspect, wherein the quantization input from the outside based on the output of the still frame coding determination means. It is characterized in that it comprises a quantization parameter value control means for controlling the quantization width of the parameter, wherein the encoding means performs the encoding using the quantization parameter controlled by the quantization parameter value control means. It is a thing. As a result, it is possible to perform high image quality encoding with less quantization error.

An image coding apparatus according to a twenty-second aspect of the present invention is the image coding apparatus according to the twentieth aspect, wherein the skip macroblock determining means encodes the encoded frame from the outside. By inputting a frame encoded signal indicating a method, the skip macroblock determination process is started, and the encoding means, when the frame encoded signal is a signal indicating an intra-frame predictive encoding method, the encoded Each of the macroblocks of the encoded frame is encoded, and when the frame encoded signal is not a signal indicating the intra-frame predictive encoding method, the difference value is encoded. As a result, still frame coding is not performed for frames and macroblocks for which intra-frame predictive coding is performed, and as a result, there is no need to reduce the quantization width, and the increase in the amount of generated code can be further suppressed. it can.

Further, the image coding program according to claim 23 of the present invention divides the coded frame to be coded into macroblocks as a processing unit, and obtains a difference value between the frame and the reference frame. Is a program for performing the process of encoding the difference value by a computer, and it is determined whether or not the encoded frame is a still frame. As a result of the determination, it is determined that the frame is a still frame. The program is characterized by being a program for causing a computer to perform a high image quality encoding process of encoding at least one of the frames to have a higher image quality than the encoding of other frames. This allows
It is possible to specify a frame to be coded that is a still frame and perform high-quality encoding on at least one frame, and increase the amount of generated code by encoding all still frames with high image quality as in the past. It is possible to suppress the image quality and to make the subsequent frames have high image quality.

[0043]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings. The embodiment shown here is merely an example, and the present invention is not necessarily limited to this embodiment.

(Embodiment 1) An image coding apparatus and an image coding method according to Embodiment 1 of the present invention will be described below. Image coding apparatus 10 according to the first embodiment
0 is one-pass coding that encodes the image signal to be encoded with one input, and determines whether the frame to be encoded is a still frame and starts the still frame. In a frame (one frame later) subsequent to (a frame immediately before a frame determined to be a still frame), high-quality coding (still frame coding) is performed as compared with other frames.

FIG. 1 is a block diagram showing the configuration of the image coding apparatus 100 according to the first embodiment. In the figure, the same or corresponding components as those in FIG. 16 are designated by the same reference numerals, and the description thereof will be omitted.

In FIG. 1, reference numeral 15 denotes a still frame coding determiner, which determines whether the frame is a frame for still frame coding or a frame for normal (similar to the conventional method) coding. To do. A quantization parameter value controller 16 controls the quantization width of the quantization parameter input from the outside based on the output of the still frame coding determiner 15.

The image coding apparatus 1 configured as described above
The operation of 00 will be described. First, the luminance / color difference signal of the frame to be encoded is divided into macroblocks as a processing unit by the blocker 1, the difference value s1 between the frame and the reference frame is obtained by the subtraction unit 2, and the code The still frame coding determiner 15 determines whether the encoded frame is a still frame (a frame that is completely still as compared to the immediately preceding encoded frame, or a frame that has many still portions). For a frame determined to be present, it is determined whether to perform still frame encoding or normal frame encoding. The quantization width of the quantization parameter input from the outside based on this determination result (output of the still frame coding determination unit 15) is the quantization parameter value controller 16
And the controlled quantization parameter s3
(The output of the quantization parameter value controller 16) is input to the skip macroblock determiner 4.

The skip macroblock determiner 4 uses the difference value s1 and the quantization parameter s3 to determine whether the macroblock is a skip macroblock.

When the skip macro block determiner 4 determines that the macro block is not the skip macro block, the switch 3 is turned on and the difference value s1
To the compression coding means 5 and the difference value s1 is converted into a frequency domain signal by the frequency converter 6 in the compression coding means 5. The frequency domain signal is quantized by the quantizer 7 using the quantization parameter s3 output from the quantization parameter value controller 16, and the output (quantization coefficient) of the quantizer 7 has a variable length. Variable-length coding is performed by the encoder 8. The output of the quantizer 7 is locally decoded by the local decoding unit 9, and the locally decoded data is used by the local prediction image generation unit 12 to generate a prediction image (prediction macroblock).

On the other hand, when the skip macroblock determiner 4 determines that the macroblock is a skip macroblock, the skip macroblock signal s2 is sent from the skip macroblock determiner 4 to the variable length encoder 8. The data is output, and the variable length encoder 8 multiplexes the differential signals of the other macroblocks with the variable length encoded data, and outputs the multiplexed signal as a code string.

Here, the operation of the still frame coding decision unit 15 will be described in detail. First, a still frame determination method by the still frame coding determination unit 15 will be described with reference to FIG.
Will be described in detail.

FIG. 2 shows each of frames 201 to 208 of an image to be encoded, which are encoded in order from the frame 201. It should be noted that the frames 203 to 208 are still frames for which inter-frame predictive coding is performed, and the frame 2
03-205 and frame 208 are forward prediction frames (prediction is performed from past frames) frames (P frames),
The frames 206 and 207 are bidirectional prediction frames (prediction is made from past or future or any one of the frames) (B frames). Each frame is a frame indicated by an arrow, for example, a frame 204 is a frame 20.
3 is a reference frame.

The still frame coding determiner 15 uses the frame 203 as a reference frame when coding each macroblock of the frame 204 which is a P frame, and codes the difference value from the corresponding macroblock. At this time, if the sum of the absolute values of the difference values of the frame is small, it is determined that the frame is stationary. Similarly, frame 205
The frame 204 is referred to in the encoding of B, the frame 205 is referred to in the encoding of frame 208, and B is referred to.
When encoding the frames 206 and 207 which are frames, the frame 208 is referred to and the difference value from the corresponding macroblock is encoded.

In this way, the still frame coding decision unit 1
5 determines that the frame is a still frame when the absolute value sum of the difference values between the frame and the reference frame is smaller than a predetermined threshold. Therefore, a still frame is determined to be one frame at the shortest after the image has stopped. In FIG. 2, the first still frame is the frame 204, and the frames 205 to 208 in which the still image continues are followed. It becomes the still frame of.

Next, with reference to FIG. 3, a detailed description will be given of an encoding determination method for determining whether to perform still frame encoding or normal frame encoding on a frame determined to be a still frame in the still frame determination. Explained. Here, the value of FLAG used for the coding determination will be described using 0 and 1, but the value is not limited to this.

The still frame coding determiner 15 first
The value THR determined by a predetermined method with respect to the sum Y of the absolute values of the differences between the screens of the frame and the reference frame
It is determined whether it is smaller than a (step S303).

In step S303, if the sum Y of absolute values of inter-screen differences is equal to or greater than the threshold value THRa, the frame is normally subjected to inter-screen predictive coding (step S3).
04), FLAG for determining still frame coding is updated to 0 (step S305).

On the other hand, in step S303, when the absolute value sum Y of the inter-screen differences is smaller than the threshold value THRa, it is determined whether FLAG for determining still frame coding is 0 (step S306). If FLAG is not 0, the frame is subjected to normal inter-picture predictive coding (step S307). If FLAG is 0, the frame is assumed to be still frame encoded (step S308), and FLAG for determining still frame encoding is updated to 1 (step S309). FLAG for determining still frame coding has an initial value of 0, and FLAG is set at the time of still frame coding.
Is updated to 1 and this value is not updated while the image is still (during a still frame). FLAG is updated to 0 when it is determined that the frame is not a still frame, and this value is not updated until the next still frame. In this way, it is determined whether or not the frame is coded as a still frame based on the value of FLAG, the frame one frame after the start of stillness is specified, and still frame coding is performed only for that frame.

Next, the control of the quantization parameter value by the quantization parameter value controller 16 according to the output of the still frame coding determiner 15 will be described in detail with reference to FIG. The quantization parameter value controller 16 receives the output of the still frame coding determiner 15 (a signal indicating the determination result of whether or not the frame is still frame encoded) and the quantization parameter given from the outside. Then, the control process of the quantization parameter value is started, and it is determined whether or not the frame is still frame encoded, based on the output signal of the still frame encoding determiner 15 (step S401).

If it is determined in step S401 that the frame is not still frame encoded,
The quantization parameter value QPinput given from the outside is set as the quantization parameter value QP of the frame (step S402).

On the other hand, when it is determined in step S401 that the frame is still frame encoded, the quantization parameter value QP of the frame is set to the quantization parameter value QPinput given from the outside by a predetermined method. The determined value β (0 <β <1) is multiplied (step S403).

In this way, the still frame coding decision unit 1
When it is determined by 5 that the frame is a frame for still frame coding (frame one frame after the start of stillness), the quantization parameter value is set to a value smaller than the normal quantization parameter value (quantization By reducing the width), in the skip macroblock determiner 4, the threshold value THR_L in step 1601 of FIG. 17 shown in the conventional image encoding method becomes a small value as a result, and the frame is not still frame encoded. The ratio of skip macroblocks is smaller than that. This is equivalent to that the reference macroblock does not replace the macroblock and increases the portion (inter-macroblock) that encodes the difference, and the frame one frame after the frame at which the image starts to be stationary is encoded with high image quality. It is effective to do. Further, in the still frame to be encoded in the still frame, the quantizer 7 performs quantization using the quantization parameter set to a value smaller than that in the normal frame encoding, so that the quantization error is small, that is, high. Image quality coding can be performed.

The quantization parameter used in the quantizer 7 may be a quantization parameter input from outside instead of the output s3 of the quantization parameter value controller 16.

Further, in the above-mentioned image coding apparatus 100, the input to the skip macroblock determiner 4 is the quantization parameter s3 which is the output of the quantization parameter value controller 16, but the still frame coding determiner is used. The output of 15 may be used.

Here, the skip macroblock determiner 4
An image coding apparatus in which the input to the still frame coding determination unit 15 is the output will be described with reference to FIG. Note that the skip macroblock determiner 4 in the image encoding device 100 of FIG. 1 and the conventional image encoding device 1500 determines from the quantization parameter value to determine whether or not the encoding macroblock is a skip macroblock. Although the threshold THR_L to be used is used, in the image encoding device 110 shown in FIG. 5, the skip macroblock determiner 4 uses the output of the still frame encoding determiner 15 to make the above determination.

First, the processing method of the skip macroblock determiner 4 in the image coding apparatus 110 shown in FIG. 5 will be described with reference to FIG. The skip macroblock determiner 4 outputs the still frame coding determiner 15 (a signal indicating whether or not the frame is still frame encoded).
Is input, the skip macroblock determination process is started, and it is determined based on the input signal whether or not the frame is a frame for still frame coding (step S601).

As a result of the above determination, when it is determined that the frame is a frame not subjected to still frame coding,
The threshold value THRskip is a threshold value THRsk determined by a predetermined method.
It is set to ip_base (step S602). On the other hand, when it is determined that the frame is a frame for still frame encoding, the threshold value THRskip is set to a value obtained by multiplying the threshold value THRskip_base by a value γ (0 <γ <1) determined by a predetermined method (step S603). ). The threshold value THRsk
The ip is used to determine whether or not the macro block is a skip macro block.

Next, by using the threshold value THRskip determined in step S602 or step S603, it is determined whether the sum X of absolute values of the inter-screen difference values s1 of the macro block input from the subtraction means 2 is smaller than the threshold value THRskip. (Step 604). As a result of the determination, if the absolute value sum X of the difference values of the macroblock is equal to or more than THRskip, it is determined that the macroblock encodes the inter-screen difference value (step S605), and if not, it is determined to be a skip macroblock. The determination is made (step S606).

As described above, when the frame is still frame-encoded, the threshold value THRskip used for the determination by the skip macroblock determiner 4 is set to a value smaller than that when the still frame is not encoded, so that a skip macroblock is obtained. Is suppressed. This is equivalent to increasing the portion for encoding the difference without replacing the macro block with the reference macro block, and is effective for encoding still frame encoded frames with high image quality.

In the image coding apparatus 110 shown in FIG. 5, except for the quantization parameter value controller 16, the quantization parameter used in the quantization by the quantizer 7 is used as the quantization parameter input from the outside. Good.

In the image coding apparatus according to the first embodiment, the image to be coded in the 1-pass coding is higher in the frame one frame after the frame in which the image starts to be stationary than in the other frames. Since the image quality is encoded, the subsequent still frame can refer to the image encoded with high image quality, and compared with the two-pass encoding in which the encoded data is referred to and encoded. The image quality can be improved for a frame while the image is still in a short coding time.

In the first embodiment, the quantization parameter value controller 16 is used for the encoding of the frame determined by the still frame encoding determiner 15 as the subsequent still frame. If the quantization parameter value is controlled to a value larger than the quantization parameter value input from the outside, the number of skip macroblocks in the frame will be larger than in other frames, but the reference macro to be replaced as a skip macroblock Since the blocks have already been encoded with high image quality, the image quality does not deteriorate, and the generated code amount can be suppressed as compared with other frames.

In the first embodiment, immediately before the coding of the coded frame is started, the difference between the frames of the coded frame and the reference frame is calculated, and the coding method of the coded frame (still Whether or not to encode the frame) is determined, but the difference value calculated at the time of encoding each macroblock of the encoded frame is held, and after the encoding of the encoded frame is completed, The difference value of the entire frame may be obtained from the held difference value, and the encoding method for the next encoded frame may be determined from the obtained difference value.

In the first embodiment, the coding method is one-pass coding, but the coding method may be two-pass coding in which coding is performed with reference to data that has been coded once. In this case, it is determined whether or not the frame is a still frame based on the generated code amount of each frame once encoded, a flag indicating that the frame is a still frame given in the first pass encoding, a frame difference value, or the like. If high-quality coding (still frame coding) is performed on the still start frame, the coding process takes longer than the one-pass coding method. It is possible to specify the frame accurately, and it is possible to encode the frame to be encoded in the still frame with high image quality one frame earlier than the image encoding device shown in FIGS. 1 and 5 of the first embodiment.

Here, the image coded data (bit stream) coded by the image coding apparatuses 100 and 110 according to the first embodiment has the following features. Since the image encoding devices 100 and 110 always encode the next frame of the frame that has started to be stationary with high image quality, the encoded bit stream is a bit encoded by a conventional image encoding device. Compared to streams
The frame immediately after the start of stillness has a smaller quantization parameter value than the other P frames, and the frame immediately after the frame having the smaller quantization parameter value has a higher proportion of skip macroblocks. It has the feature. This is because the higher the image quality of the high-quality coded frame, the smaller the difference value in the subsequent still frame, and the larger the proportion of skip macroblocks, and the higher the quality of the subsequent frame. You can

Further, the image coding devices 100 and 110
Thus, if the frame immediately after the start of stillness is encoded as an I frame, the encoded bitstream has a feature that the frame immediately after the start of stillness is always encoded as an I frame. Therefore, since the I frame has a better image quality than the P frame, the image quality can be better than that of the still frame coding.

Further, the image coding devices 100 and 110
Thus, if the quantization parameter value is controlled to be small for the frame immediately after the start of stillness and then the quantization parameter value is controlled to be higher than the normal value in the subsequent still frames, the encoded bitstream becomes a still image. In the following still frame, the quantization parameter value has a large value. This means that when the quantization parameter value of the subsequent still frame is increased, the number of skip macroblocks is increased, and a still frame-coded image is displayed, so that a high quality image can be displayed. Also, when the frame immediately after the start of stillness is set as a still frame and the quantization parameter value is reduced and encoded, the generated code amount of one frame increases, but if the quantization parameter value of the subsequent still frame is increased, it becomes 1 Since the generated code amount per frame is small, the generated code amount between the still images is averaged, and as a result, the average generated code amount (bit rate) becomes substantially constant.

A bitstream having such characteristics can be decoded by a commonly used image decoding device, and can reproduce an image having a higher image quality than a bitstream by a conventional image coding device. You can Further, as the image decoding device, the number of skip macroblocks is not less than the number obtained by a predetermined method with respect to the total number of macroblocks in the decoding of the subsequent still frame, instead of performing the above-described normal decoding processing. If it is determined that the frame is a skip frame not to be displayed and the decoding process of the next frame is started, not only high-quality image reproduction can be obtained,
It is also possible to reduce the calculation amount in the decoding process.

The image coding method according to the first embodiment is based on MPEG-2, MPEG-4, H.264. 261 and H.264. It can be incorporated into standardized image coding methods such as H.263. However, the MPEG-4 standard stipulates that the variation of the quantization parameter value of the macroblock is within ± 2.

Further, although the still frame is determined between the frames in the above, as a method of determining the stillness of the image, intra-frame prediction, that is, prediction from the already encoded neighboring macroblock of the same frame is also performed. good. For example, in the image stillness determination method of the apparatus shown in FIGS. 1 and 5, the determination is made only by the difference value from the reference image, but the already encoded neighboring macroblock of the same frame (the upper part of the macroblock, the left Part, upper left portion, and the like) may be determined as stillness of each macroblock based on information regarding whether the macroblock is a skip macroblock. Further, the fluctuation of the quantization parameter value for each macroblock may be determined not only from the difference value and the flag but also by referring to the neighboring macroblocks that have already been coded and by their states. For example, when the macroblock immediately before the macroblock is a skip macroblock, the macroblock can be encoded with high image quality by controlling the quantization parameter value of the macroblock to be small.

An image coding apparatus using the image stillness determination method based on such intra-frame prediction is configured, for example, as shown in FIG. In FIG. 7, the same or corresponding components as those in FIG. 1 are designated by the same reference numerals, and the description thereof will be omitted.

Below, the image coding device 12 shown in FIG.
A method for determining a 0 skip macroblock will be described. The skip macroblock determiner 4 first determines whether or not the left macroblock of the macroblock is a skip macroblock (step S70).
1). Although the macro block to be referred to is described as a left macro block here, it is not necessary to limit the macro block to the left macro block, that is, the immediately preceding macro block, and the neighboring macro blocks (the upper part, the left part of the macro block, Any macro block located on the upper left or the like may be used. Further, the greater the number of neighboring macroblocks to be referred to, the more detailed skip macroblock determination can be performed.

When the left macroblock is not the skip macroblock, the value of the quantization parameter QP of the macroblock is the quantization parameter QP given from the outside.
The same value as input is set (step S702). On the other hand, if the left macroblock is a skip macroblock,
The value of the quantization parameter QP of the macroblock is set smaller than the value of the quantization parameter QPinput given from the outside (step S703).

In this way, even in the case of only the intra-picture prediction that does not perform the differential coding with the reference frame, the still part of the image can be predicted and the quantization parameter value can be controlled small. As a result, encoding can be performed with high image quality while suppressing the generation of code amount. When this method is used in combination with the image coding method of the apparatus shown in FIG. 1 or FIG. 5, high-quality coding with few coding errors due to fluctuations in quantization parameters even in a frame to be coded in still frame coding. It can be performed.

(Second Embodiment) An image coding apparatus and an image coding method according to the second embodiment of the present invention will be described below. Image coding apparatus 20 according to the second embodiment
0 has a configuration as shown in FIG. 9, and determines a still frame for an image to be encoded, and compares a frame that is two frames after the frame that started to be stationary with another frame. High-quality coding (still frame coding) is performed.

The image coding apparatus 10 according to the first embodiment
The difference from the configuration of 0 is that in the configuration of the image encoding device 100 for 1-pass coding shown in FIG. 1, the processing of the still frame encoding determiner 15 that determines whether or not to perform still frame encoding, and Quantization parameter value controller 16
Is the difference in processing. The difference will be described below with reference to FIGS. 10 and 12.

First, a still frame coding determination processing method by the still frame coding determination unit 15 in the second embodiment will be described with reference to FIG. A ternary flag indicating the encoding method (encoding type) of the macroblock is used for the determination of the still frame encoding. It should be noted that this flag is given for each macroblock, which is a unit of encoding processing, and is referred to when encoding the next frame. In the figure, a flag 0 represents a case where the macro block is not a skip macro block, and a flag 1 is a macro block which is not still frame coded in a past coded frame, and the macro block is a skip macro block. Represents the case
The flag 2 represents a case where the macro block has already been coded in a still frame in the past coded frame and the macro block is a skip macro block.

When the coding of the frame is started and the processing of the still frame coding determiner 15 is started, the still frame coding determiner 15 firstly based on the output s4 of the skip macroblock determiner 4. It is determined whether the macro block is a skip macro block (step S803).

If the macroblock is not a skip macroblock, the flag value of the macroblock is set to 0 (step S804). If the macro block is a skip macro block, the flag value of the reference macro block is referenced (step S805), and the flag value is 0.
If so, the flag value of the macroblock is set to 1 (step S806), and if not, it is determined whether the frame is still frame coding (step S806).
807).

If the result of determination in step S807 is that still frame coding has not been performed, the flag value of the macroblock is set as the flag value of the reference macroblock (step S808). If the frame is still frame encoded, set the flag value of the macroblock to 2
(Step S809). When this processing is performed in the coding of all the macroblocks of the frame and the coding of all the macroblocks of the frame is completed, it is determined whether the number F of macroblocks having a flag value of 1 is larger than the threshold value THRc ( Step S810). The threshold T
HRc is a value obtained by multiplying the total number of macroblocks in the frame by a value δ (0 <δ <1) determined by a predetermined method.

When the number F of macroblocks having the flag value of 1 is larger than the threshold value THRc in step S810, the next coded frame is still frame coded (step S812). If not,
The next encoded frame is encoded normally (step S8).
11).

Here, regarding the above-mentioned three-valued flag, the state in which the flag value changes from the flag value of the reference macroblock depending on the state of the macroblock is shown in the state transition diagram of FIG.

The flag value of the reference macroblock is 0 (N9
01), the flag value remains 0 if the macro block is not a skip macro block, and if the macro block is a skip macro block (M
904), the flag value of the macroblock is 1 (N9
02).

The flag value of the reference macroblock is 1 (N9
02), if the macroblock is a skip macroblock and the frame is not still frame coded, the flag value of the macroblock remains 1, and the macroblock is a skip macroblock and the frame is Is still frame encoded (M905), the flag value of the macroblock is 2
(N903), and if the macroblock is not a skip macroblock, the flag value of the macroblock is 0 (N901).

The flag value of the reference macroblock is 2 (N9
03), if the macro block is a skip macro block, the flag value of the macro block remains 2. If the macro block is not a skip macro block, the flag value of the macro block is 0 ( N901).

As described above, the still frame coding determiner 15 outputs the signal s indicating the coding method of the macroblock.
4 (signal indicating whether or not it is a skip macro block)
From the skip macroblock determiner 4 and further using the flag value representing the coding method of the frame determined in the immediately preceding frame coding and the coding method of the macroblock given to the reference macroblock, It is determined whether or not the encoded frame is still frame encoded, and the result determined at the time of encoding the previous frame is used in encoding the frame. In addition, in FIG.
The three-valued flag described in FIG. 11 does not limit its value to 0, 1, 2.

Next, a method of controlling the quantization parameter value by the quantization parameter value controller 16 in the second embodiment will be described with reference to FIG. Here, information on whether or not the reference frame is still frame coding, which is held by the still frame coding determiner 15,
The quantization parameter value controlled by the flag value of the reference macroblock will be described.

The quantization parameter value controller 16 gives a flag value of the reference macroblock determined by the still frame coding determiner 15 and a signal s5 indicating whether or not the reference frame is still frame coding from the outside. And the quantization parameter to be input are input, and control processing of the quantization parameter value is started.

First, based on the output signal s5 of the still frame coding determiner 15, it is judged whether or not the frame is still frame coded (step S1001). When it is determined that the frame is not still frame encoding, the quantization parameter value QPinput given from the outside
Is set as the quantization parameter value QP of the macroblock (step S1002). On the other hand, when it is determined that the frame is still frame encoding, it is determined whether the flag value of the reference macroblock is 0 (non-skip macroblock) (step S1003).

When the flab value of the reference macroblock is not 0 in step S1003, the quantization parameter value of the macroblock is a value ε determined by a predetermined method for the quantization parameter value QPinput given from the outside. The value is multiplied by (0 <ε <1) (step S1004). When the flab value of the reference macroblock is 0, the quantization parameter value of the macroblock is a quantization parameter value QPinpu given from the outside.
t is multiplied by a value ε (0 <ε <1) determined by a predetermined method, and further, a value ζ (0 <0 <determined by a predetermined method.
ζ) is added (step S1005).

As described above, the quantization parameter value controller 1
6 controls the quantization parameter value input from the outside based on the output s5 of the still frame coding determiner 15 to perform still frame coding before the immediately preceding frame coding in the coding of the macroblock. When a macro block that is not present is used as a reference macro block, it is determined that the macro block is a macro block of the second frame after the macro block starts to be stationary, and the quantization parameter value QP is input from the outside in the encoding of the normal frame. It can be set to a value smaller than the quantization parameter value QPinput. When referring to a macroblock that has already undergone still frame coding, the reference macroblock has already been coded with a small coding error, so the quantization parameter value is larger than the value of QPinput × ε. Even if it is set to, the coding error does not increase.

According to the MPEG-4 standard, the fluctuation of the quantization parameter value of the macroblock is regulated within the range of ± 2. That is, since the difference between the quantization parameter values of adjacent macroblocks is only ± 2, the quantization parameter value QP of 1 which varies depending on ε and ζ is shown in FIG.
The difference in value for each macroblock is +2, +1, 0, -1,-.
It must be controlled to be either of the two.

Here, considering that the generated code amount increases by controlling the quantization parameter value to a small value,
For a macroblock (step S1005) for which still frame coding is not performed, the quantization parameter value Q
Quantization parameter value QPinpu input P from the outside
It should be t or a value close to it.

In the image coding apparatus 200, the quantization parameter value used for the differential coding in the quantizer 7 is not the output s3 of the quantization parameter value controller 16 but the quantizer input from the outside. It may be a parameter value.

Further, in the judgment by the skip macroblock coding judgment unit 4 as to whether or not the macroblock should be a skip macroblock, as shown in the image coding apparatus 110 in FIG. 5, still frame coding judgment means. You may use the output of.

Further, the flag is used to determine whether or not the unit is still in macroblock units, but the present invention is not limited to this, and another unit may be used. For example, the determination may be made in each of the four divided parts of the frame, or in any area.

In the image coding apparatus 200 according to the second embodiment, a flag used to determine whether or not still frame coding is performed is given to each macroblock, the coding method of the reference frame, and the flag value of the reference macroblock. , And the state of the macroblock (whether it is a skip macroblock, whether still frame coding or not), it is possible to easily realize a determination process as to whether or not the next coded frame is still frame coded. By controlling the quantization parameter value in units of macroblocks, it is possible to determine whether or not to make the macroblock a skip macroblock and to perform quantization at the time of differential encoding, as shown in the first embodiment. It is possible to code with as little coding error as possible.

Here, the image coded data (bit stream) coded by the image coding apparatus 200 according to the second embodiment has the following features. Since the image encoding device 200 always encodes a frame next to a frame that has started to be stationary with high image quality, the encoded bit stream becomes a bit stream encoded by a conventional image encoding device. In comparison, the frame immediately after the start of stillness has a smaller quantization parameter value than the other P frames, and the frame immediately after the frame having the smaller quantization parameter value has a higher proportion of skip macroblocks. It has the characteristic of being This is because the higher the image quality of the high-quality coded frame, the smaller the difference value in the subsequent still frame, and the larger the proportion of skip macroblocks, and the higher the quality of the subsequent frame. You can

Further, by the image coding device 200,
If the frame immediately after the start of stillness is encoded as an I frame, the encoded bit stream has a feature that the frame immediately after the start of stillness is always encoded as an I frame. Therefore, the I frame is P
Since the image quality is better than that of frames, the image quality can be better than that of still frame coding.

Further, by the image coding device 200,
If the quantization parameter value is controlled to be small for the frame immediately after the start of stillness, and the quantization parameter value is controlled to be larger than the normal value in the subsequent still frames,
The encoded bit stream has a feature that the quantization parameter value is large in the subsequent still frame of the still image. This means that when the quantization parameter value of the subsequent still frame is increased, the number of skip macroblocks is increased, and a still frame-coded image is displayed, so that a high quality image can be displayed. Also, when the frame immediately after the start of stillness is set as a still frame and the quantization parameter value is reduced and encoded, the generated code amount of one frame increases, but if the quantization parameter value of the subsequent still frame is increased, it becomes 1 Since the generated code amount per frame is small, the generated code amount between the still images is averaged, and as a result, the average generated code amount (bit rate) becomes substantially constant.

The image coded data (bitstream) having such characteristics can be decoded by a commonly used image decoding device, and the image quality is higher than that of the bitstream by the conventional image coding device. It is possible to reproduce various images. Also, as an image decoding device,
Instead of performing the above normal decoding process, when decoding the subsequent still frame, if the number of skip macroblocks is equal to or more than the number obtained by a predetermined method with respect to the total number of macroblocks, the frame concerned is skipped. If it is determined that the skip frame is not displayed and the decoding process of the next frame is started, not only a high-quality image can be reproduced, but also the amount of calculation in the decoding process can be reduced. .

In addition, bidirectional predictive coding is performed by the method of determining the stillness of an image by the flag shown in the second embodiment and coding with high image quality using the flag (temporally past frame or temporally. It does not apply to a frame (referring to a future frame) to a frame (B frame). That is,
If normal encoding is performed without updating the flag or controlling the quantization parameter accordingly, a frame (B frame) that is not used as a prediction image (that does not serve as a reference frame of a subsequent frame) is generated. It is possible to suppress an increase in code amount. And it started to stop 2
Since the P frame after the frame (that is, the frame referred to by the subsequent frame) is encoded with a high image quality as a still frame, the image quality of the subsequent frames that refer to the still frame encoded frame is improved.

If the frame is determined to be a still frame and the frame is a B frame, high quality image is encoded with a quantization parameter smaller than usual by controlling the quantization parameter, but the flag is updated. If not performed, the B frame that refers to the frame that is determined to be a still frame is still frame encoded (encoded with high image quality) and becomes the subsequent reference frame.
Since the frame is also still frame encoded, the image quality can be improved in the subsequent still frame.

The image coding method according to the second embodiment is based on MPEG-2, MPEG-4, H.264. 261 and H.H. It can be incorporated into standardized image coding methods such as H.263. However, the MPEG-4 standard stipulates that the variation of the quantization parameter value of the macroblock is within ± 2.

(Embodiment 3) An image coding apparatus and an image coding method according to Embodiment 3 of the present invention will be described below. Image coding apparatus 30 according to the third embodiment
0 is a configuration as shown in FIG. 13, which determines a still frame for an image to be encoded, and a frame two frames after the frame that has started to be stationary is compared with other frames. High-quality coding (still frame coding) is performed.

The image coding apparatus according to the third embodiment and the image coding apparatus 200 according to the second embodiment include the still frame coding determiner 15 that determines whether or not to perform still frame coding. Since the processing of 1 and the processing of the skip macroblock determiner 4 are different, the difference will be described below.

First, the operation of the skip macroblock determiner 4 in the third embodiment will be described. The skip macroblock determiner 4 is a frame coded signal s1 that is input from the outside and indicates the coding method of the frame.
When 1 is input, the skip macroblock determination process is started.

When the input signal s11 is a signal indicating an intra-frame predictive coding frame (I frame), the image signal s10 output from the blocker 1 is compressed and coded by the compression coding means 5.
The switch 17 is controlled so as to output the data, and the data is compressed and encoded by the compression encoder 5.

When the input signal s11 is a signal that does not indicate an I frame, the sum of absolute values X of the difference values s1 of the macro block output from the subtracting means 2 and the absolute value sum X are determined by a predetermined method. The threshold value THR_H is compared (step S1201). The threshold value THR_H is obtained by multiplying the quantization parameter value s3 by the value α1 (0 <α1 <1).

The sum X of absolute values of the difference value s1 is the threshold value T
If it is greater than or equal to HR_H, the process proceeds to step 1202, and the macro block is determined to be an intra macro block. As a result, the output signal s10 of the blocker 1 is compression encoded by the compression encoding means 5.

On the other hand, if the absolute value sum X of the difference values s1 is less than or equal to the threshold value THR_H, the process proceeds to step 1203,
The sum X of absolute values and the threshold T determined by a predetermined method
It is compared with HR_L (S1203). Where the threshold T
HR_L has a value α2 (0 <α
2 <α1).

When the absolute value sum X of the difference values s1 is equal to or more than the threshold value THR_L in step S1203, the process proceeds to step 1204, and the macro block is set as an inter macro block. Thereby, the subtraction means 2
The output s1 of the above is compressed and encoded by the compression encoding means 5.

If the sum X of the absolute values of the difference value s1 is smaller than the threshold value THR_L in step S1203, the flow advances to step 1205 to set the macroblock as a skip macroblock. As a result, the skip macroblock signal s2 is output to the variable length encoder 8.

As described above, the skip macroblock determiner 4 outputs the signal s11 indicating the frame coding method of intraframe predictive coding or interframe predictive coding of the frame input from the outside, The quantization parameter value s3, which is the output from the quantization parameter value controller 16, is input and the macroblock is intra-picture predictive coding (intra macroblock) or inter-picture predictive coding (inter macroblock). Determine whether to do.

Next, the processing of the still frame coding decision unit 15 of the image coding apparatus 300 according to the third embodiment will be described with reference to FIG. The flag used in FIG. 15 is a ternary flag indicating the encoding method (encoding type) of the macroblock, and is the same as that described in the second embodiment. The three-valued flag used in FIG. 15 is not limited to 0, 1, and 2.

When the coding of the frame is started and the processing of the still frame coding determiner 15 is started, the still frame coding determiner 15 first uses the output s4 of the skip macroblock determiner 4. Then, it is determined whether the macro block is a skip macro block (step S1301). If the macro block is not a skip macro block, set the flag value of the macro block to 0.
(Step S1302).

If the macro block is a skip macro block, the flag value of the reference macro block is referenced (step S1303). If the flag value of the reference macroblock is 0, the flag value of the macroblock is set to 1 (step S1304). If not, the process proceeds to step S1305, and the frame is still frame encoded or not. To determine. If the frame is not still frame coded, the flag value of the macroblock is set to be the same as the flag value of the reference macroblock (step S1306). If the frame is still frame encoded, the flag value of the macroblock is set to 2 (step S1307).

The above processing is performed in the coding of all the macroblocks of the frame, and is repeated until the coding of all the macroblocks of the frame is completed. When the encoding process of all macroblocks is completed, it is determined whether the number F of macroblocks having flag 1 is larger than the threshold value THRc (step S1308). Here, the threshold value THRc is a value obtained by multiplying the total number of macroblocks in the frame by a value δ (0 <δ <1).

At step S1308, if the number F of macroblocks having flag 1 is larger than the threshold value THRc, the next coded frame is assumed to be still frame coded (step S1310), and if not,
The next encoded frame is encoded normally (step S1).
309).

As described above, the still frame coding decision unit 15 obtains the coding method (whether or not it is a skip macro block) of the macro block from the output of the skip macro block decision unit 4, and further The encoding method of the frame (whether or not it is a still frame) determined in the immediately preceding frame encoding held in the frame encoding determiner 15, and the encoding method of the macroblock given to the reference macroblock. (Encoding type)
It is determined whether or not the next coded frame is to be still frame coded by using a flag value that represents, and the result determined at the time of coding the previous frame is used in the coding of the frame.

In the image coding device 300,
In the differential encoding in the quantizer 7, the quantization parameter value may be the quantization parameter value input from the outside.

In the coding of the macroblock, the skip macroblock determiner 4 determines whether or not the macroblock is a skip macroblock, as shown in the image coding apparatus 110 of FIG. The output of the still frame coding determination means may be used.

In the image coding apparatus according to the third embodiment as described above, a flag used to determine whether or not still frame coding is performed is given to each macroblock, the reference frame coding method, and the reference macroblock flag value. , And the state of the macroblock (whether it is a skip macroblock,
Whether or not still frame coding is performed), it is possible to easily realize whether or not the next coded frame is still frame coded, and by controlling the quantization parameter value in units of macroblocks, As shown in the first embodiment, it is possible to perform coding with the least possible coding error in the skip macroblock determination of the macroblock and in the quantization at the time of differential coding, and the intraframe prediction coding is possible. By not applying the still frame coding to the frames and macroblocks for which the coding is performed, it is possible to suppress the generated code amount in the coding without reducing the quantization width.

In the image coding apparatus according to the third embodiment, when the signal indicating the coding method of the frame input from the outside is the signal indicating the intra-frame predictive coding (I frame), the coding is performed. If it is determined that the frame is a still frame, inter-frame predictive coding is performed instead of encoding as an I frame, so that the reference image is high-quality data, and image quality deterioration is small, and The generated code amount can be reduced. Here, the image coded data (bit stream) coded by the image coding apparatus 300 according to the third embodiment has the following features.

Since the image encoding device 300 always encodes the frame next to the frame that has started to be stationary with high image quality, the encoded bit stream was encoded by the conventional image encoding device. Compared to bitstream
The frame immediately after the start of stillness has a smaller quantization parameter value than the other P frames, and the frame immediately after the frame having the smaller quantization parameter value has a higher proportion of skip macroblocks. It has the feature. This is because the higher the image quality of the high-quality coded frame, the smaller the difference value in the subsequent still frame, and the larger the proportion of skip macroblocks, and the higher the quality of the subsequent frame. You can

Further, by the image coding device 300,
If the frame immediately after the start of stillness is encoded as an I frame, the encoded bitstream has a feature that the frame immediately after the start of stillness is always encoded as an I frame. Therefore, since the I frame has a better image quality than the P frame, the image quality can be better than that of the still frame coding.

Further, by the image coding device 300,
For the frame immediately after the start of stillness, the quantization parameter value is controlled to be small, and the quantization parameter value is controlled to be larger than the normal value in the subsequent still frames. It is characterized in that the quantization parameter value is large in the subsequent still frame. This means that when the quantization parameter value of the subsequent still frame is increased, the number of skip macroblocks is increased, and a still frame-coded image is displayed, so that a high quality image can be displayed. Further, if the frame immediately after the start of stillness is set as a still frame and the quantization parameter value is reduced and encoding is performed, 1
Although the generated code amount of a frame increases, the generated code amount per frame decreases on the contrary when the quantization parameter value for the subsequent still frame is increased. Therefore, the generated code amount between still images is averaged. As a result, the average generated code amount (bit rate) becomes almost constant.

Further, by the image coding device 300,
When it is determined that the frame is a still frame, even if an I-frame coding instruction is input from the outside, if the inter-frame predictive coding is performed, the I frame is always inserted in the subsequent still frame in which stillness continues. It has the characteristic that it does not exist. Since the reference image has high image quality, it is possible to prevent image quality deterioration and reduce the generated code amount.

The coded image data (bitstream) having such characteristics can be decoded by a commonly used image decoding device, and further has a higher image quality than a bitstream by a conventional image coding device. It is possible to reproduce various images. Also, as an image decoding device,
Instead of performing the above normal decoding process, when decoding the subsequent still frame, if the number of skip macroblocks is equal to or more than the number obtained by a predetermined method with respect to the total number of macroblocks, the frame concerned is skipped. If it is determined that the skip frame is not displayed and the decoding process of the next frame is started, not only a high-quality image can be reproduced, but also the amount of calculation in the decoding process can be reduced. .

The image coding method according to the third embodiment is based on MPEG-2, MPEG-4, H.264. 261 and H.H. It can be incorporated into standardized image coding methods such as H.263. However, the MPEG-4 standard stipulates that the variation of the quantization parameter value of the macroblock is within ± 2.

In each of the first to third embodiments described above, the input image is described in frame units, but the present invention is not limited to this, and similar effects can be obtained in field units. You can Furthermore, in each of the above-described first to third embodiments, the image coding method of the present invention is realized by hardware, but it may be realized by software.

[0142]

As described above, according to the image coding method of the first aspect of the present invention, the coded frame to be coded is divided into macroblocks which are processing units, and An image encoding method for obtaining a difference value from a reference frame and encoding the difference value, the still frame determining step of determining whether or not the encoded frame is a still frame, In the frame determination step, a high image quality encoding step of performing still frame encoding in which at least one frame of the frames determined to be still frames is encoded with a higher image quality than the encoding of other frames, Than including
By specifying the encoded frame that is a still frame and performing high image quality encoding on at least one frame, it is possible to increase the amount of generated code by encoding all still frames with high image quality as in the conventional case. The image quality can be suppressed and the quality of the subsequent frames can be improved.

According to the image coding method of the second aspect of the present invention, in the image coding method of the first aspect, the still frame determining step includes the coded frame and the reference frame. Since the above determination is made based on the difference value between the frame and the frame, the next frame of the frame that has started to be stationary can be specified, and high-quality encoding can be performed on the next frame. Also becomes high quality.

According to the image coding method of the third aspect of the present invention, in the image coding method of the first aspect, the high image quality coding step is performed on the object determined to be the still frame. The still frame coding determination step for determining whether to perform still frame coding or normal frame coding for the coded frame, and the coding method determined in the still frame coding determination step A quantization parameter value control step of controlling the quantization width of the quantization parameter input from the outside based on
Skip macroblock determination for determining whether or not each macroblock of the frame to be encoded is a skip macroblock using the quantization parameter and the difference value controlled in the quantization parameter value control step It is determined to be a skip macroblock by including a step and an encoding step that encodes the encoded frame based on the determination result of each macroblock obtained in the skip macroblock determination step. As a result, the number of generated macroblocks can be suppressed, and as a result, it is possible to prevent an increase in the amount of generated code and to encode a coded frame determined to be a still frame with high image quality without degrading image quality. can do.

Further, according to the image coding method of the fourth aspect of the present invention, in the image coding method of the first aspect, the high image quality coding step is performed on the object determined to be the still frame. In the still frame coding determination step of determining whether to perform still frame coding or normal frame coding for the coded frame, and in the still frame coding determination step, the coded frame remains stationary. When it is determined that the frame is not frame-encoded, whether each macroblock of the encoded frame is a skip macroblock using the difference value between the encoded frame and the reference frame and a predetermined threshold value. It is determined whether or not the frame to be encoded is a frame for still frame encoding in the still frame encoding determination step. If the predetermined threshold is controlled to be small, each macroblock of the encoded frame is a skip macroblock using the controlled threshold and the difference value between the encoded frame and the reference frame. A skip macroblock determining step for determining whether or not there is, and an encoding step for encoding the encoded frame based on the determination result of each macroblock obtained in the skip macroblock determining step, ,
By including, it is possible to suppress the number of macroblocks determined to be skip macroblocks, as a result,
It is possible to prevent an increase in the amount of generated code, and to encode a coded frame determined to be a still frame with high image quality without causing deterioration in image quality.

Further, according to the image coding method of the fifth aspect of the present invention, in the image coding method of the first aspect, it is determined that the preceding frame of the frame to be coded is determined. Still frame coding for determining whether the next frame of the coded frame is a frame for still frame coding or a frame for normal frame coding based on the coding information of the frame to be coded A determination step, a quantization parameter value control step of controlling a quantization width of a quantization parameter input from the outside based on the encoding method determined in the still frame determination step, and the quantization parameter value control step Each macroblock of the coded frame is skipped by using the quantization parameter and the difference value controlled by A skip macroblock determining step for determining whether or not the block is a black block, and a coding method for the next frame of the coded frame based on the determination result of the still frame coding determining step, and By including the encoding step of encoding the encoded frame based on the encoding information determined in the encoding of the previous frame, there is less quantization error,
High quality coding can be performed.

According to the image coding method of the sixth aspect of the present invention, in the image coding method of the first aspect, the still frame determination step has each macroblock of the reference frame. Since the coding method of the macroblock is performed based on the coding type that represents the coding method, it is possible to reduce the amount of calculation in the still frame determination of the frame to be coded. Is specified and only that frame is encoded with high image quality. As a result, the generated code amount can be suppressed more than before, and the subsequent frames also have high image quality.

According to the image coding method of the seventh aspect of the present invention, in the image coding method of the sixth aspect, the coding type of each macroblock is the coded macroblock. Is a skip macroblock, and a second flag indicating whether the coded frame is still frame coded when the coded macroblock is a skip macroblock. Three
Is a three-valued flag, and the flag of the encoded macroblock is updated according to the flag value of the macroblock referred to by the encoded macroblock. The determined macroblock is judged as a skip macroblock without being subjected to high image quality encoding in the subsequent processing, and as a result, only the first frame can be encoded in high image quality and the encoded frame It is possible to suppress the calculation amount in the still frame determination.

According to the image coding method of the eighth aspect of the present invention, in the image coding method of the first aspect, the high image quality coding step only performs intra-frame predictive coding. Since the high quality coding is not performed for the I frame, the still frame coding is not performed for the frame for which the intra prediction coding is performed, and as a result, it is not necessary to reduce the quantization width. , It is possible to further suppress an increase in the generated code amount.

According to the image coding method of the ninth aspect of the present invention, in the image coding method of the first aspect, the high image quality coding step is a macro for performing intra-picture predictive coding. Since high-quality coding is not performed for blocks, still frame coding is not performed for macroblocks that perform intra-picture predictive coding, and as a result,
It is not necessary to reduce the quantization width, and it is possible to further suppress the increase in the generated code amount.

According to the image coding method of the tenth aspect of the present invention, the coded frame to be coded is divided into macroblocks as processing units, and the difference between the frame and the reference frame is calculated. An image encoding method for obtaining a value and encoding the difference value, the method comprising: a still macroblock determining step of predicting in-screen whether or not each macroblock of the encoded frame is still; A macroblock coding step for coding a macroblock that is determined to be still in the block determination step with higher image quality than the coding of other macroblocks is included. It is possible to suppress the occurrence of the error and to encode with high image quality.

Further, according to the image coding method of the eleventh aspect of the present invention, in the image coding method of the first aspect, the high image quality coding step is performed on the object determined to be the still frame. In the encoded frame, a still macroblock determination step of predicting whether or not each macroblock is still in the screen, and other macroblocks determined to be still in the still macroblock determination step By including a macroblock coding step for coding with higher image quality than that for macroblock coding, a high-quality code with less coding error due to fluctuations in quantization parameters even in a frame for still frame coding. Can be converted.

According to the twelfth aspect of the image coding apparatus of the present invention, the coded frame to be coded is divided into macroblocks which are processing units, and the coded frame and the reference frame are coded. An image encoding apparatus that obtains a difference value between the encoded frame and the encoded frame based on the difference value, and determines whether the encoded frame is a still frame based on the difference value. Whether the frame to be coded, which has been determined to be the still frame, is a frame for performing still frame encoding that encodes with higher image quality than the encoding of other frames, or a normal frame code Still frame coding determining means for determining whether the frame is a frame to be encoded, and a quantization parameter input from the outside based on the output signal of the still frame encoding determining means. Using the quantization parameter value control means for controlling the quantization width of, the quantization parameter controlled by the quantization parameter value control means and the difference value, each macroblock of the encoded frame is skipped macroblock Skip macro block determining means for determining whether or not, and the encoding means for encoding the frame to be encoded based on the output of the skip macro block determining means, the skip macro block Suppress the number of macroblocks that are determined to be
It is possible to encode the frame next to the frame that has started to be stationary with high image quality, and as a result, it is possible to suppress the amount of generated code as compared with the conventional technique and to improve the image quality of the subsequent frames.

According to the image coding apparatus of the thirteenth aspect of the present invention, in the image coding apparatus of the twelfth aspect, the coding means is controlled by the quantization parameter value control means. Since the above encoding is performed using the quantization parameter, the quantization error is small,
High quality coding can be performed.

According to the image coding apparatus of the fourteenth aspect of the present invention, the coded frame to be coded is divided into macroblocks which are processing units, and the coded frame and the coded frame are coded. An image encoding apparatus that obtains a difference value from an already-referenced frame and encodes the encoded frame based on the difference value, wherein the encoded frame is a still frame based on the difference value. Whether the frame to be coded, which is determined to be a still frame by determining whether there is a still frame, is a frame for performing still frame encoding that encodes with higher image quality than other frames, or a normal frame encoding The still frame coding determining means for determining whether the frame is to be processed, the output of the still frame coding determining means, and the difference value are used to determine each marker of the encoded frame. A skip macroblock determining unit that determines whether or not the block is a skip macroblock, and an encoding unit that encodes the encoded frame based on the output of the skip macroblock determining unit. Therefore, it is possible to suppress the number of macroblocks that are determined to be skip macroblocks, and to encode the frame next to the frame that has started to be still with high image quality, and as a result, reduce the amount of generated code more than before. And the subsequent frames also have high image quality.

According to the fifteenth aspect of the image coding apparatus of the present invention, in the image coding apparatus of the fourteenth aspect, the image is input from the outside based on the output signal of the still frame coding determination means. A quantization parameter value control means for controlling the quantization width of the quantization parameter thus obtained, wherein the encoding means uses the quantization parameter controlled by the quantization parameter value control means to Since the encoding is performed, it is possible to perform high quality encoding with less quantization error.

According to the image coding apparatus of the sixteenth aspect of the present invention, the coded frame to be coded is divided into macroblocks which are processing units, and the coded frame and the reference frame are An image coding apparatus that obtains a difference value between the coded frame and the coded frame based on the difference value, and determines whether or not each macroblock of the coded frame is stationary. Prediction is performed, and when the macroblock in the vicinity of the macroblock is a skip macroblock, the macroblock is determined as a still macroblock by a skip macroblock determining unit that determines the macroblock as a still macroblock and the skip macroblock determining unit. The quantization parameter value of the relevant macroblock is smaller than the quantization parameter value input from the outside. The quantization parameter value control means for controlling so that the macroblock determined to be the still macroblock by using the quantization parameter controlled by the quantization parameter value control means is used for encoding of another macroblock. By providing the encoding means for encoding with high image quality, the generation of the code amount is suppressed, and
It can be encoded with high image quality.

Further, according to the image coding apparatus of the seventeenth aspect of the present invention, the coded frame to be coded is divided into macroblocks as processing units, and the difference between the frame and the reference frame is calculated. An image encoding apparatus for obtaining a value and encoding the difference value, wherein the encoding target is based on the encoding information of the encoding target frame determined in the encoding of the previous frame of the encoding target frame. Still frame coding that determines whether the next frame of the frame is a frame that performs still frame coding that codes with higher image quality than the coding of other frames or a frame that performs normal frame coding Determination means and quantization parameter value control means for controlling the quantization width of a quantization parameter input from the outside based on the output of the still frame determination means , A skip macroblock determining means for determining whether or not each macroblock of the encoded frame is a skip macroblock using the difference value and the quantization parameter controlled by the quantization parameter value control means And a coding method for the next frame of the coded frame based on the determination result of the still frame coding determination means, and the coded information is determined based on the coding information determined in the coding of the preceding frame. The number of macroblocks that are determined to be skip macroblocks is suppressed, and the next frame of the frame that has started to be stationary is encoded with high image quality by including an encoding unit that encodes the encoded frame. As a result, the amount of generated codes can be suppressed more than before, and the subsequent frames also have high image quality.

Further, according to the image coding apparatus of the eighteenth aspect of the present invention, in the image coding apparatus of the seventeenth aspect, the coding means is controlled by the quantization parameter value control means. Since the above encoding is performed using the quantization parameter, the quantization error is small,
High quality coding can be performed.

According to a nineteenth aspect of the present invention, there is provided the image encoding device according to the seventeenth aspect, wherein the skip macroblock determining means is an externally encoded frame. The skip macroblock determination process is started by the input of the frame coded signal indicating the coding method, and the coding means, when the frame coded signal is a signal indicating the intra-frame prediction coding method, Each macroblock of the frame to be coded is coded, and if the frame coded signal is not a signal indicating the intraframe prediction coding method, the difference value is coded. For frames and macroblocks for which the still frame coding is not performed, as a result, there is no need to reduce the quantization width,
It is possible to further suppress an increase in the generated code amount.

Further, according to the image coding apparatus of the twentieth aspect of the present invention, the coded frame to be coded is divided into macroblocks which are processing units, and the difference between the frame and the reference frame. An image encoding apparatus for obtaining a value and encoding the difference value, wherein each macroblock of the encoded frame is a skip macroblock using the difference value and a quantization parameter input from the outside. A skip macroblock determining means for determining whether or not there is a next frame of the encoded frame based on the encoding information of the encoded frame determined in the encoding of the previous frame of the encoded frame. , A frame for which still frame coding is performed to achieve higher image quality than other frames, or normal frame coding is performed Still frame coding determining means for determining whether the frame is a frame, and determining the encoding method of the next frame of the encoded frame based on the determination result of the still frame encoding determining means. The encoding is provided based on the encoding information determined in the encoding of the preceding frame, and the determination processing as to whether or not the next frame code is the still frame encoding is easily realized. In addition, it is possible to suppress the number of macroblocks that are determined to be skip macroblocks, and to encode a frame that is two frames after the frame that has started to be stationary with higher image quality. The generated code amount can be suppressed, and the subsequent frames also have high image quality.

According to a twenty-first aspect of the present invention, there is provided the image encoding device according to the twenty-first aspect, in which the image data is input from the outside based on the output of the still frame encoding determination means. And a quantization parameter value control means for controlling the quantization width of the quantization parameter, wherein the encoding means performs the encoding using the quantization parameter controlled by the quantization parameter value control means. As a result, it is possible to perform high image quality encoding with less quantization error.

According to a twenty-second aspect of the present invention, there is provided the image encoding apparatus according to the twenty-first aspect, wherein the skip macroblock determining means is an externally encoded frame. By the input of a frame coded signal indicating the coding method, the skip macroblock determination process is started, and the coding means, when the frame coded signal is a signal indicating an intra-frame prediction coding method, When each macroblock of the frame to be coded is coded, and when the frame coded signal is not a signal indicating the intra-frame prediction coding method, the difference value is coded. Still frame coding is not performed on the frames and macroblocks to be coded, and as a result, it is not necessary to reduce the quantization width, and the generated code amount is increased. It can be more suppressed.

According to the image coding program of the twenty-third aspect of the present invention, the coded frame to be coded is divided into macroblocks as processing units, and the difference between the frame and the reference frame is calculated. A program for performing a process of obtaining a value and encoding the difference value by a computer, determining whether or not the encoded frame is a still frame, and as a result of the determination, determines that the frame is a still frame. At least one of the determined frames
Since it is a program for a computer to perform high image quality encoding processing that encodes a frame with higher image quality than that of other frames, the encoded frame that is a still frame is specified, and at least one frame is set. On the other hand, it is possible to perform high image quality encoding, suppress an increase in the amount of generated code due to encoding all still frames to high image quality as in the conventional art, and also make the subsequent frames high image quality. .

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an image encoding device according to a first embodiment of the present invention.

FIG. 2 is a conceptual diagram of a frame determined to be a still frame according to the first embodiment and the second embodiment of the present invention.

[Fig. 3] Fig. 3 is a flowchart showing a process of determining whether or not still frame coding is to be performed by the image coding device according to Embodiment 1 of the present invention.

FIG. 4 is a flow chart diagram showing a control process of a quantization parameter value of the image coding apparatus according to the first embodiment of the present invention.

FIG. 5 is a first embodiment in which an input to a skip macroblock determiner is an output of a still frame coding determiner.
FIG. 11 is a diagram showing a modified example of the image coding apparatus according to the present invention.

FIG. 6 is a flowchart showing skip macroblock determination processing of the image encoding device when the input to the skip macroblock determination unit is the output of the still frame coding determination unit.

FIG. 7 is a diagram showing a modified example of the image coding device according to the first embodiment when using intra prediction.

FIG. 8 is a flowchart showing a control process of a quantization parameter value of the image coding apparatus according to the first embodiment when using the intra prediction.

FIG. 9 is a block diagram for explaining an image coding apparatus according to Embodiment 2 of the present invention.

FIG. 10 is a diagram showing a flow of determination processing as to whether or not still frame encoding is performed by the image encoding device according to the second embodiment of the present invention.

FIG. 11 is a state transition diagram showing state transitions of flags used in a process of determining whether or not still frame encoding is performed by the image encoding device according to the second embodiment of the present invention.

FIG. 12 is a diagram showing a flow of a quantization parameter value control process of the image coding apparatus according to the second embodiment of the present invention.

FIG. 13 is a block diagram for explaining an image coding apparatus according to Embodiment 3 of the present invention.

FIG. 14 is a diagram showing a flow of skip macroblock determination processing of the image encoding device according to the third embodiment of the present invention.

FIG. 15 is a diagram showing a flow of determination processing as to whether or not still frame encoding is to be performed by the image encoding device according to the third embodiment of the present invention.

FIG. 16 is a block diagram for explaining a conventional generally used image encoding device.

FIG. 17 is a diagram showing a flow of processing for determining skip macroblocks in the conventional image coding apparatus that is generally used.

[Explanation of symbols]

100, 110, 120, 200, 300, 1500
Image encoder 1 Blocker 2 Subtractor 3 Switch 4 Skip macroblock determiner 5 Compression encoder 6 Frequency converter 7 Quantizer 8 Variable length encoder 9 Local decoder 10 Inverse quantizer 11 Frequency Inverse changer 12 Local predicted image generation means 13 Predicted image generator 14 Frame memory 15 Still frame coding determination device 16 Quantization parameter value controller 201 to 208 frames

Continued front page    F-term (reference) 5C059 MA00 MA05 MA21 MC11 MC15                       MC38 ME01 NN23 PP05 PP06                       PP16 TA48 TB07 TC03 TC13                       TC27 TD06 TD12 UA02 UA33                 5J064 AA02 BB01 BB03 BC16 BC26                       BD01

Claims (23)

[Claims] 1. An image coding method in which a coded frame to be coded is divided into macroblocks as processing units, a difference value between the frame and a reference frame is obtained, and the difference value is coded. Then, at least one frame of at least one of the frames determined to be still frames in the still frame determination step of determining whether the coded frame is a still frame A high image quality encoding step for performing still frame encoding that encodes with higher image quality compared to the encoding of other frames,
An image coding method comprising: 2. The image encoding method according to claim 1, wherein the still frame determining step makes the determination based on a difference value between the coded frame and the reference frame. Image coding method. 3. The image coding method according to claim 1, wherein in the high quality coding step, still frame coding is performed on a coded frame determined to be the still frame, or Still frame coding determination step of determining whether to perform frame coding, and controlling the quantization width of the quantization parameter input from the outside based on the coding method determined in the still frame coding determination step Whether each macroblock of the encoded frame is a skip macroblock using the quantization parameter value control step, the quantization parameter controlled in the quantization parameter value control step, and the difference value The skip macroblock determination step for determining Based on the determination result of each macroblock including a coding step of coding the target encoded frame, image coding method, characterized in that. 4. The image coding method according to claim 1, wherein in the high quality coding step, still frame coding is performed on a coded frame determined to be the still frame, or In the still frame coding determination step of determining whether to perform frame coding, and in the still frame coding determination step, when it is determined that the coded frame is a frame in which still frame coding is not performed, the coded It is determined whether or not each macroblock of the encoded frame is a skip macroblock using a difference value between the encoded frame and the reference frame and a predetermined threshold value, and the still frame encoding determination step is performed to detect the encoded frame. When it is determined that the coded frame is a frame for still frame coding, the predetermined threshold value is controlled to be smaller, and the control is performed. A skip macroblock determining step of determining whether or not each macroblock of the encoded frame is a skip macroblock by using a controlled threshold value and a difference value between the encoded frame and a reference frame, An image encoding method, comprising: an encoding step of encoding the frame to be encoded based on the determination result of each macroblock obtained in the skip macroblock determining step. 5. The image coding method according to claim 1, wherein the encoded frame is encoded based on encoding information of the encoded frame determined in encoding of a previous frame of the encoded frame. A still frame coding determination step of determining whether the next frame is a frame for still frame coding or a frame for normal frame coding, and the coding method determined in the still frame determination step Based on the quantization parameter value control step for controlling the quantization width of the quantization parameter input from the outside, using the quantization parameter and the difference value controlled in the quantization parameter value control step , A skip macroblock for determining whether or not each macroblock of the encoded frame is a skip macroblock. A lock determination step and a coding method for the next frame of the encoded frame are determined based on the determination result of the still frame encoding determination step, and the encoding of the encoded frame is performed by encoding the previous frame. And a coding step performed based on the coding information determined in the coding. 6. The image coding method according to claim 1, wherein the still frame determination step is performed based on a coding type which is included in each macroblock of the reference frame and which represents a coding method of the macroblock. An image coding method characterized by the following. 7. The image coding method according to claim 6, wherein the coding type of each macroblock is a first flag indicating whether or not the coded macroblock is a skip macroblock, When the coded macroblock is a skip macroblock, it is a ternary flag of a second flag and a third flag indicating whether the coded frame is still frame coded. The image coding method is characterized in that the flag of the encoded macroblock is updated according to the flag value of the macroblock referred to by. 8. The image encoding method according to claim 1, wherein the high image quality encoding step does not perform high image quality encoding on an I frame for which only intra-picture predictive encoding is performed. Image coding method. 9. The image encoding method according to claim 1, wherein the high image quality encoding step does not perform high image quality encoding on a macroblock for which intra-picture predictive encoding is performed. Encoding method. 10. An image coding method for blocking a coded frame to be coded into macroblocks as a processing unit, obtaining a difference value between the frame and a reference frame, and coding the difference value. Then, for the still macroblock determination step of predicting whether or not each macroblock of the encoded frame is still, and for the macroblock determined to be still in the still macroblock determination step, And a macroblock coding step for coding with a higher image quality than the coding of other macroblocks. 11. The image coding method according to claim 1, wherein the high image quality coding step displays whether or not each macroblock is still in a coded frame determined to be the still frame. In-prediction still macroblock determination step, and macroblock code that encodes a macroblock that is determined to be still in the still macroblock determination step with higher image quality than the encoding of other macroblocks An image encoding method comprising: an encoding step; 12. An encoding target frame to be encoded is blocked into macroblocks which are processing units, a difference value between the encoding frame and a reference frame is obtained, and the encoding target is based on the difference value. An image encoding device for encoding a frame, wherein the encoded frame is determined based on the difference value to determine whether the encoded frame is a still frame, and the encoded frame is determined to be the still frame. Still frame coding determining means for determining whether the frame is a frame for performing still frame coding for coding with higher image quality than that for coding for other frames, or a frame for performing normal frame coding. A quantization parameter value control means for controlling the quantization width of a quantization parameter input from the outside based on the output signal of the still frame coding determination means, Using the quantization parameter and the difference value controlled by the quantization parameter value control means, a skip macroblock determination means for determining whether each macroblock of the encoded frame is a skip macroblock, An image encoding apparatus comprising: an encoding unit that encodes the frame to be encoded based on an output of the skip macroblock determining unit. 13. The image coding apparatus according to claim 12, wherein the coding means performs the coding by using a quantization parameter controlled by the quantization parameter value control means. Image encoding device. 14. A coding target frame to be coded is divided into macroblocks as a processing unit, a difference value between the coding target frame and a coded reference frame is calculated, and based on the difference value. An image coding apparatus for coding the coded frame, wherein the coded frame is judged as a still frame by judging whether the coded frame is a still frame based on the difference value. Still frame coding determining means for determining whether the coded frame is a frame for performing still frame coding for coding with higher image quality than other frames, or a frame for performing normal frame coding, It is determined whether each macroblock of the coded frame is a skip macroblock using the output of the still frame coding determination means and the difference value. A skip macro-block determination means for, encoding means for encoding the target encoding frame based on the output of the skip macro-block determination unit, comprising a picture coding apparatus, characterized in that. 15. The image coding apparatus according to claim 14, wherein a quantization parameter value control for controlling a quantization width of a quantization parameter input from the outside based on an output signal of the still frame coding determination means. An image coding apparatus, comprising: means, wherein the coding means codes the frame to be coded using the quantization parameter controlled by the quantization parameter value control means. 16. An encoding target frame which is an encoding target is divided into macroblocks which are processing units, a difference value between the encoding target frame and a reference frame is obtained, and the encoding target is based on the difference value. An image coding apparatus for coding a frame, wherein intra-screen prediction is performed as to whether or not each macroblock of the coded frame is stationary, and a macroblock in the vicinity of the macroblock is a skip macroblock. At this time, a skip macroblock determining unit that determines the macroblock as a still macroblock, and a quantization parameter value of the macroblock that is determined as a still macroblock by the skip macroblock determining unit is input from an externally input quantum. Quantization parameter value control means for controlling to be smaller than the quantization parameter value; Coding means for coding the macroblock determined to be the still macroblock using the quantization parameter controlled by the child parameter value control means with higher image quality than the coding of other macroblocks, An image coding apparatus, comprising: 17. An image coding apparatus which blocks a coded frame to be coded into macroblocks which are processing units, obtains a difference value between the frame and a reference frame, and encodes the difference value. Then, based on the encoding information of the encoded frame determined in the encoding of the previous frame of the encoded frame, the next frame of the encoded frame has a higher image quality than the encoding of other frames. Still frame coding determining means for determining whether the frame is a still frame encoding frame or a normal frame encoding frame, and an external input based on the output of the still frame determining means. The quantization parameter value control means for controlling the quantization width of the quantization parameter, and the difference value and the quantization parameter value control means. Using a controlled quantization parameter, the skip macroblock determination means for determining whether or not each macroblock of the encoded frame is a skip macroblock, and the determination result of the still frame coding determination means An encoding means for determining the encoding method of the next frame of the encoded frame based on the encoding means for encoding the encoded frame based on the encoding information determined in the encoding of the previous frame; An image coding apparatus, comprising: 18. The image coding apparatus according to claim 17, wherein the coding means performs the coding using a quantization parameter controlled by the quantization parameter value control means. Image encoding device. 19. The image coding apparatus according to claim 17, wherein the skip macroblock determining means receives the frame coded signal indicating a coding method of the coded frame from the outside, and thereby the skip macroblock. When the frame coded signal is a signal indicating an intra-frame predictive coding method, the coding means performs coding of each macroblock of the coded frame to determine the frame code. An image coding apparatus, characterized in that when the coded signal is not a signal indicating an intra-frame predictive coding method, the difference value is coded. 20. An image coding apparatus for blocking a frame to be coded to be coded into macroblocks as a processing unit, obtaining a difference value between the frame and a reference frame, and coding the difference value. Then, using the difference value and a quantization parameter input from the outside, a skip macroblock determining unit that determines whether or not each macroblock of the encoded frame is a skip macroblock, and the encoded target A still frame in which the next frame of the encoded frame is encoded with higher image quality than the encoding of other frames based on the encoding information of the encoded frame determined in the encoding of the preceding frame of the encoded frame. Still frame coding judging means for judging whether it is a frame for frame coding or a frame for normal frame coding The coding method of the next frame of the coded frame is determined based on the determination result of the still frame coding determination means, and the coding of the coded frame is the code determined in the coding of the previous frame. An image coding apparatus comprising: a coding unit that performs coding based on coding information. 21. The image coding apparatus according to claim 20, wherein the quantization parameter value control means controls the quantization width of the quantization parameter input from the outside based on the output of the still frame coding determination means. The image coding apparatus, wherein the coding means performs the coding using the quantization parameter controlled by the quantization parameter value control means. 22. The image coding apparatus according to claim 20, wherein the skip macroblock determination means receives the frame coded signal indicating an encoding method of the coded frame from the outside, and thus the skip macroblock. The block determining process is started, and when the frame coded signal is a signal indicating an intra-frame predictive coding method, the coding means codes each macroblock of the coded frame, An image coding apparatus, characterized in that, when the coded signal is not a signal indicating an intra-frame predictive coding method, the difference value is coded. 23. A computer performs a process of dividing an encoded frame to be encoded into macroblocks as a processing unit, obtaining a difference value between the frame and a reference frame, and encoding the difference value. Is a program for determining whether or not the encoded frame is a still frame, and as a result of the determination, at least one of the frames determined to be a still frame is a code of another frame. An image coding program, which is a program for causing a computer to perform a high-quality coding process for coding with a higher image quality than encoding.