JP4886747B2 - Image coding apparatus, image coding method, and program - Google Patents

Description

  The present invention relates to moving image processing, specifically, a moving image encoding technique.

  There are a variable bit rate method and a fixed bit rate method for encoding moving images. The variable bit rate method is a method in which image quality is prioritized and the frame bit rate is not constant (Patent Documents 1 and 2). The fixed bit rate method is a method in which the bit rate of each frame is the same.

  In an apparatus that decodes an encoded image (hereinafter referred to as a compressed image), an upper limit value per frame is set according to conditions such as system layer restrictions and disk drive read performance, and the code amount of the frame is Exceeding the upper limit may interfere with system operation. A variable bit rate compressed image has an indefinite bit rate and may exceed the upper limit depending on the frame. On the other hand, if the bit rate is fixed so that the code amount is equal to or less than the upper limit value, the compressed image of the fixed bit rate method does not exceed the upper limit value of the frame. There is a problem that image quality deterioration such as noise occurs.

Patent Document 2 describes H.264. A technique for suppressing the amount of code when encoding with a variable bit rate using the H.264 / AVC (Advanced Video Coding) method is disclosed. In this method, it is determined whether or not the code amount per frame is likely to exceed the upper limit value. When it is determined that the code amount is likely to exceed the upper limit, a part of the transform coefficient after orthogonal transform is set to 0 (zero) for the frame. The amount of code is suppressed by replacing with.
JP 2006-333444 A JP 2008-22405 A

  H. mentioned above. The compression method such as H.264 / AVC cuts the information of the high frequency component more than the information of the low frequency component by utilizing the characteristic that the human eye is sensitive to the low frequency component but insensitive to the high frequency component. Yes. Specifically, when the transform coefficient after orthogonal transform is quantized, a larger quantization table value is assigned as the frequency is higher.

  In the method of Patent Document 1 in which a part of the transform coefficient after the orthogonal transform is replaced with 0, the code amount is further suppressed by manipulating the transform coefficient before quantization.

  However, since many high-frequency components are cut in the quantization, even if the conversion coefficient of the high-frequency components is replaced with 0 before the quantization, the effect of suppressing the code amount cannot be expected to be so great. On the other hand, if the conversion coefficient close to the low frequency component is replaced with 0 in order to surely suppress the code amount, the image quality deterioration becomes remarkable.

  One aspect of the present invention is an image encoding method. This method uses a frame immediately before a scene change frame, which is a frame whose bit rate exceeds a predetermined threshold when a moving image is compressed by a variable bit rate method among a plurality of frames constituting a moving image as a target frame. And processing for copying a partial region of the scene change frame to the target frame corresponding to the scene change frame. Then, the processed target frame is used, and the original frame is used for frames other than the target frame, and compression is performed by the fixed bit rate method.

  In addition, what expressed the said method as an apparatus, a system, or a program is also effective as an aspect of this invention.

  According to the technology of the present invention, when a moving image is compressed, it is possible to reliably realize the suppression of the code amount and the reduction of the image quality deterioration.

  Before describing a specific embodiment of the present invention, first, the principle of the encoding technique according to the present invention will be described.

As described above, when a moving image is compressed, if a variable bit rate method is used, deterioration in image quality can be suppressed, but the code amount may exceed the upper limit value, while a fixed bit rate method is used. Although it is possible to prevent the code amount from exceeding the upper limit, there is a problem of image quality deterioration such as block noise conspicuous. The inventor of the present application has earnestly searched for a trade-off between image quality and code amount, and as a result, has found the following two points.
1. A frame in which the amount of code increases when encoding by the variable bit rate method and a frame in which block noise is conspicuous when encoding by the constant bit rate method have common characteristics.

Such a frame is one in which frame prediction is not effective, that is, a frame having a large variation from the immediately preceding frame. Such a frame is hereinafter referred to as a scene change frame.
2. When a scene change occurs, the screen changes greatly, so even if the image quality of the frame immediately before the screen change, that is, immediately before the scene change frame is reduced, the human eye does not feel uncomfortable.

  The inventor of the present application has established the technology according to the present invention based on these findings. In this technique, after a part of a scene change frame is copied to a frame (target frame) immediately before the scene change frame, compression is performed by a fixed bit rate method. Since it is a fixed bit rate, the code amount is constant. In addition, since the frame immediately before the scene change is processed, it is possible to avoid deterioration in image quality without giving a sense of incongruity to human eyes.

  Note that the processing is not limited to a single frame immediately before the scene change frame, but a plurality of frames immediately before the scene change frame may be the target frame. In this case, a partial area of the scene change frame is divided into a plurality of areas, and one area is copied to a plurality of target frames corresponding to the scene change frame. By doing this, the amount of variation caused by the scene change is distributed to a plurality of frames before the scene change frame, so that it is possible to more reliably avoid a deterioration in image quality.

  Although not particularly limited, it is preferable to detect the scene change frame by the following method.

  In this method, first, a functional block (compression unit) responsible for compression performs pre-compression by a variable bit rate method using an original frame of a moving image. A frame in which the bit rate obtained by the pre-compression exceeds a predetermined threshold is detected as a scene change frame. By doing so, the compression unit may perform pre-compression and control may be performed so that the compression unit performs main compression by a fixed rate method after processing according to the result of the pre-compression. That is, it is not necessary to provide a dedicated device for detecting a scene change frame, and the size of the device can be suppressed.

Based on the above principle, a specific embodiment according to the technique of the present invention will be described.
FIG. 1 shows an image compression apparatus 100 according to an embodiment of the present invention. The image compression apparatus 100 is an H.264 filer. The moving image is compressed by the H.264 / AVC method, and includes a data supply unit 110, an adder 120, an orthogonal transform unit 122, a quantization unit 124, a variable length coding unit 126, a rate control unit 130, and an inverse quantization unit 140. , An inverse orthogonal transform unit 142, a frame memory 150, a motion compensation processing unit 160, an image processing unit 170, a stream output control unit 180, a stream buffer 182, and an operation mode control unit 190. The data supply unit 110 includes a pre-compression image data memory 112, a main compression image data memory 114, and a switching unit 116. The rate control unit 130 includes a rate adjustment unit 132 and a rate evaluation unit 134. The frame memory 150 includes a pre-compression prediction frame memory 152 and a main compression prediction frame memory 154. The image processing unit 170 includes a processing execution unit 172 and a bit rate value storage memory 174. The adder 120, the orthogonal transform unit 122, the quantization unit 124, the variable length coding unit 126, the rate control unit 130, the inverse quantization unit 140, the inverse orthogonal transform unit 142, and the motion compensation processing unit 160 constitute a compression unit. .

  In FIG. 1, each element described as a functional block for performing various processes can be configured by a processor, a memory, and other circuits in terms of hardware, and recorded in the memory in terms of software. Alternatively, it is realized by a loaded program. Therefore, it is understood by those skilled in the art that these functional blocks can be realized in various forms by hardware only, software only, or a combination thereof, and is not limited to any one. Also, for the sake of clarity, only those necessary for explaining the technique of the present invention are shown in these drawings.

  When compressing a moving image, the image compression apparatus 100 performs the first half processing for performing compression (pre-compression) for grasping the approximate code amount of each frame of the moving image, and the actual compression (using the result of the first half processing). The latter half of the processing is performed sequentially. The pre-compression is performed by a variable bit rate method, and the main compression is performed by a fixed rate method. Image processing is performed between the first half process and the second half process, and switching between the first half process and the second half process is controlled by the operation mode control unit 190.

  FIG. 2 shows only functional blocks related to the first half process in the image compression apparatus 100. The moving image is input to the data supply unit 110 in units of frames. The pre-compression image data memory 112 in the data supply unit 110 stores input data of each frame. Hereinafter, the data of each frame stored in the pre-compression image data memory 112 is referred to as pre-compression image data. H. In the H.264 / AVC format, the image display order and the compression order are different, but it is assumed here that frames are input to the data supply unit 110 in the compression order.

  During the first half processing, the switching unit 116 outputs frame data (pre-compression image data) from the pre-compression image data memory 112 to the adder 120 and the motion compensation processing unit 160.

  The adder 120 combines the pre-compression image data from the data supply unit 110 and the calculation result of the motion compensation processing unit 160 to create input data to the orthogonal transform unit 122.

  The orthogonal transform unit 122 performs orthogonal transform on the input data created by the adder 120 to obtain transform coefficients. This transform coefficient is output from the orthogonal transform unit 122 to the quantization unit 124.

  The quantization unit 124 performs a quantization process on the transform coefficient of the orthogonal transform unit 122. The quantization process is a process that is responsible for reducing the amount of code, and how much it is reduced is in accordance with an instruction from the rate adjustment unit 132 of the rate control unit 130. The quantization unit 124 outputs the result of the quantization process to the variable length coding unit 126 and the inverse quantization unit 140.

  The variable length encoding unit 126 encodes the quantization result output from the quantization unit 124 and the motion vector output from the motion compensation processing unit 160, and outputs the encoded data to the stream output control unit 180. To do. The encoded data is the compressed data of the moving image itself. However, since the purpose is to grasp the code amount in the first half process (pre-compression), the variable length encoding unit 126 further obtains the encoded data here. The received bit rate of the frame is output to the rate evaluation unit 134 of the rate control unit 130 and the image processing unit 170.

  The rate evaluation unit 134 outputs a comparison result between the bit rate from the variable length coding unit 126 and a preset bit rate to the rate adjustment unit 132.

  The rate adjustment unit 132 gives a rate adjustment instruction according to the comparison result of the rate evaluation unit 134 to the quantization unit 124. In the first half of the processing, the rate adjustment unit 132 performs adjustment so that the variable bit rate method is used.

  The stream output control unit 180 does not output an encoded frame during the first half processing.

  The inverse quantization unit 140 outputs the result of the inverse quantization process to the inverse orthogonal transform unit 142, and the inverse orthogonal transform unit 142 outputs the result of the inverse orthogonal transform to the frame memory 150. The processing result of the inverse orthogonal transform unit 142 is frame data for motion prediction, and is stored in the pre-compression prediction frame memory 152 of the frame memory 150. Hereinafter, the frame data stored in the pre-compression predicted frame memory 152 is referred to as pre-compression predicted frame data.

  H. In H.264 / AVC, block noise reduction processing may be performed by a deblocking filter before the processing result of the inverse orthogonal transform unit 142 is stored in the frame memory 150 (here, the pre-compression prediction frame memory 152). This is omitted here.

  The motion compensation processing unit 160 performs motion prediction based on the pre-compression prediction frame data stored in the pre-compression prediction frame memory 152 during the first half processing. The motion prediction includes inter-frame prediction and intra-frame prediction. The motion compensation processing unit 160 performs these two processes.

  The image processing unit 170 stores the bit rate output from the variable length encoding unit 126 during the first half processing in the bit rate value storage memory 174. The bit rate value storage memory 174 holds the bit rate of the current frame and one or more frames immediately before the current frame. In the present embodiment, as an example, the bit rate value storage memory 174 Hold the bit rate of a total of five frames.

  Based on the bit rate of the current frame, the image processing unit 170 determines whether or not the previous four frames are to be processed. If it is determined that processing is to be performed, the data of the previous four frames (pre-compression image data) The frame data stored in the memory 112 is processed and output to the data supply unit 110. On the other hand, if it is determined not to process, the previous four frames are output to the data supply unit 110 as they are. Details of the operation of the processing execution unit 172 will be described later.

  The data supply unit 110 stores the data from the image processing unit 170 in the main compression image data memory 114. The frame data stored in the main compression image data memory 114 will be referred to as main compression image data hereinafter because it is used in the latter half of the process, that is, main compression.

  Next, the operation of the image compression apparatus 100 for the second half process will be described. The latter half process is different from the first half process in the following three points: “the data supplied from the data supply unit 110 is different”, “the operation of the rate adjustment unit 132 is different”, and “the stream output control unit 180 sends a stream Output to buffer 182 is performed. " The latter half of the process will be described with reference to FIG. FIG. 3 shows only functional blocks related to the latter half of the process.

  In the latter half of the process, the switching unit 116 of the data supply unit 110 outputs the main compression image data stored in the main compression image data memory 114 to the adder 120 and the motion compensation processing unit 160.

  The adder 120, the orthogonal transform unit 122, the quantization unit 124, the inverse quantization unit 140, and the inverse orthogonal transform unit 142 operate in the same manner as in the first half processing except that the input data is different.

  The frame memory 150 stores the output of the inverse orthogonal transform unit 142 in the main compression prediction frame memory 154 and supplies it to the motion compensation processing unit 160.

  The motion compensation processing unit 160 performs motion prediction using the frame data stored in the compression prediction frame memory 154.

  The rate adjustment unit 132 instructs the quantization unit 124 to adjust the bit rate of the frame to a preset bit rate. That is, the image compression apparatus 100 according to the present embodiment performs compression by the fixed bit rate method in the second half process.

  That is, the rate adjustment unit 132 differs in the bit rate adjustment method between the first half process and the second half process. Specifically, adjustment is performed so that the variable bit rate method is used during the first half processing, and adjustment is performed so that the fixed bit rate method is used during the second half processing.

  The stream output control unit 180 outputs nothing during the first half process, but outputs the compressed data (the bit stream resulting from the compression of the main compression) obtained by the variable length coding unit 126 to the stream buffer 182 during the second half process.

  The stream buffer 182 temporarily stores the bit stream from the stream output control unit 180 and outputs the bit stream to the outside such as an output device (not shown).

Next, the processing execution unit 172 in the image processing unit 170 will be described in detail with reference to FIGS.
H. In the H.264 / AVC format, a process called cropping is performed when a compressed moving image is decoded and reproduced. In this process, the frame is cut out so as to match the reproduction form such as the resolution of the display device. That is, instead of displaying all the pixels for each frame obtained by decoding the compressed moving image, only a part is displayed. Usually, the center area of the frame is displayed and the edge area of the frame is truncated.

  On the other hand, in inter-frame prediction at the time of encoding, motion vector prediction using a motion vector from a non-display area can be performed.

  As shown in FIG. 4, with respect to the frame 300 obtained by decoding, the display area 304 at the center part is cut out for display, and the edge part indicated by the shadow is cut off as the non-display area 302 area.

  Note that the length from the upper end of the frame to the upper end of the display area 304, the length from the lower end of the frame to the lower end of the display area 304, the length from the left end of the frame to the left end of the display area 304, the right end of the frame to the display area The lengths up to the right end of 304 are Fram_Crop_Top_Offset (FCTO) 312, Fram_Crop_Bottom_Offset (FCBO) 314, Fram_Crop_Left_Offset (FCLO) 316, and Fram_Crop_Roff 18

  The image processing unit 170 operates only during the first half process, and generates frame data (main compression image data) used for the main compression of the second half process from the pre-compression image data stored in the pre-compression image data memory 112. Is. In the present embodiment, the processing execution unit 172 sets a frame whose coding amount is large when compressed by the variable bit rate method, that is, a frame where block noise is conspicuous when compressed by the fixed bit rate method as a scene change frame, The four frames immediately before the scene change frame are set as target frames corresponding to the scene change frame, and a part of the display area of the scene change frame is divided into a plurality of areas, one in the non-display area of the plurality of target frames. make a copy.

  As shown in FIG. 5, during the first half processing, the bit rate value storage memory 174 stores the current frame 401 and the four frames immediately before the frame 401 (the previous frame 402, the previous frame 403, 3 The previous frame 404 and the previous frame 405) are stored.

  First, the processing execution unit 172 compares the bit rate of the current frame 401 with a preset threshold value Th, and determines whether or not to process the frames 402 to 405. Specifically, when the bit rate of the frame 401 is equal to or less than the threshold Th, it is determined that the frames 402 to 405 are not processed, and when the bit rate of the frame 401 exceeds the threshold Th, that is, the frame 401 is a scene change. If it can be determined that it is a frame, it is determined that the frames 402 to 405 are to be processed.

  FIG. 6 shows an example of bit rates of five frames stored in the bit rate value storage memory 174. In this example, the current frame 401 is a scene change frame, and the bit rate M0 exceeds the threshold Th. The bit rates M1 to M4 of the immediately preceding four frames 402 to 405 are below the threshold value. In this case, the processing execution unit 172 determines to process the frames 402 to 405.

  When it determines with processing, the process execution part 172 calculates the process amount with respect to the frames 402-405 further, respectively. In the present embodiment, during processing, the processing execution unit 172 cuts out data for the area exceeding the threshold from the display area of the scene change frame 401, further divides the cut data area into four regions, Copy one by one to the non-display area of four frames. Therefore, the processing amount is the size of each area copied to the frames 402 to 405. A specific method for calculating the machining amount will be described later.

  If it is determined not to be processed, the data of the frames 402 to 405 stored in the pre-compression image data memory 112 is stored in the main compression image data memory 114 as it is as main compression image data.

  FIG. 7 shows frames 502 to 505 obtained by processing the frames 402 to 405 by the processing execution unit 172. The frame 501 is the same as the current frame 401, and is a scene change frame in which the bit rate exceeds the threshold Th. In FIG. 7, only the display area of each frame and the non-display area at the bottom are displayed for easy understanding.

  As shown in FIG. 7, “area 1 + area 2 + area 3 + area 4” included in the display area among all areas of the frame 501 is an area to be copied. Regions 1 to 4 in the frame 501 are denoted as regions 601 to 604, respectively.

  The processing execution unit 172 copies the area 601 to the non-display area of the frame 402 as the area 701. Except for the region 701, the other parts of the frame 502 are the same as the corresponding parts in the frame 402.

  When encoding a motion vector, since the amount of code can be reduced when the amount of motion vector is small, that is, the distance moved is short, in this embodiment, the processing execution unit 172 determines the position in the frame 501 of the area to be copied. The area is copied to the position in the non-display area of the frame 402 corresponding to.

  Therefore, the area 601 located on the lower left side of the frame 501 is copied to the left side of the non-display area at the lower end of the frame 402 so that the motion vector amount is reduced. As shown in FIG. 7, an area 701 obtained by copying the area 601 is located on the left side of the non-display area at the lower end of the frame 502.

  Similarly, the area 602 of the frame 501 is copied to the left side of the non-display area at the lower end of the frame 403. In the frame 503 obtained thereby, a region 702 on the left side of the non-display region at the lower end is the same as the region 602.

  The area 603 of the frame 501 is copied to the right side of the non-display area at the lower end of the frame 404. In the frame 504 obtained thereby, a region 703 on the right side of the non-display region at the lower end is the same as the region 603.

  Further, the area 604 of the frame 501 is copied to the right side of the non-display area at the lower end of the frame 404. In the frame 505 obtained thereby, a region 704 on the right side of the non-display region at the lower end is the same as the region 604.

  The frames 502 to 505 obtained in this way are stored in the main compression image data memory 114 as main compression image data.

  Here, a calculation method of the machining amount by the machining execution unit 172 will be described. In the description, the example shown in FIG. 6 is used.

  The purpose of processing the frame immediately before the scene change frame is to avoid deterioration in image quality even if compression is performed at a fixed bit rate during the latter half of the processing. Specifically, the processing execution unit 172 distributes the amount of the bit rate exceeding Th to the previous four frames for the frame in which the bit rate obtained during the first half processing exceeds the threshold Th.

  As shown in FIG. 8, the amount to be shared by the four frames before the current frame is (M0-Th).

  First, the processing execution unit 172 calculates, for each frame, the product (Mi × θi) of the bit rate Mi of the frame and the sharing coefficient θi. That is, (M1 × θ1), (M2 × θ2), (M3 × θ3), and (M4 × θ4) are obtained for the previous frame to the fourth frame. The sharing coefficients θ1 to θ4 are preset numerical values from 0 to 1. The closer the value is to 1, the more the image quality degradation of the frame is suppressed, and the closer the value is to 0, the larger the image quality degradation.

Then, according to the following equation (1), the share amounts of the previous four frames are calculated.
ΔMi = (M0−Th) × {(Th−Mi × θi) / Σ _i (Th−Mi × θi)} (1)
Where Th: threshold
M0: bit rate of the current frame
Mi: bit rate of i frames before the current frame
θi: sharing coefficient set for i frames before the current frame
ΔMi: Share of i frames before the current frame

  By doing so, as shown in FIG. 9, the amount by which the bit rate of the current frame exceeds the threshold value Th is distributed to the previous four frames.

  Then, the processing execution unit 172 calculates the ratio of the share amount ΔMi of the i-th previous frame to the total of the share amounts of the previous four frames (ΔM1 + ΔM2 + ΔM3 + ΔM4) and the total size of the area copied in the current frame. The size of the area to be copied to the i-th previous frame is calculated by multiplication. This size is the processing amount of the i-th previous frame.

Next, an operation flow of the image compression apparatus 100 will be described with a focus on the operation mode control unit 190 with reference to a flowchart of FIG.
The moving images are sequentially input to the data supply unit 110 in units of frames and stored in the pre-compression image data memory 112. For the first half process, the following control is performed on the other functional blocks of the image processing apparatus 100 by the operation control mode control unit 190.

A. The data supply unit 110 is switched to use the pre-compression image data memory 112. That is, the switching unit 116 outputs frame data (pre-compression image data) from the pre-compression image data memory 112 (S100).
B. The frame memory 150 is switched to use the pre-compression prediction frame memory 152 (S102). That is, the frame data from the inverse orthogonal transform unit 142 is stored in the pre-compression prediction frame memory 152, and is output from the pre-compression prediction frame memory 152 to the motion compensation processing unit 160.
C. The rate adjustment unit 132 switches the bit rate adjustment method so that the variable bit rate method is used (S103).
D. In the stream output control unit 180, switching is performed so as not to output the bit stream (S104).

  Then, the first half processing (pre-compression) of the current frame is performed (S110).

  The bit rate of the current frame is obtained by the first half processing of step 110, and the bit rate is stored in the bit rate value storage memory 174 of the image processing unit 170 (S112).

  If the bit rate is equal to or lower than the threshold Th (S114: No), the data of the frame stored in the pre-compression image data memory 112 is transferred to the main compression image data memory 114 (S116), and the processing is step. The process proceeds to S122. On the other hand, when the bit rate exceeds the threshold Th (S114: Yes), it is determined that the frame immediately before the current frame should be processed, and the processing amount is calculated (S118).

  Then, the four frames immediately before the current frame are processed by the amount calculated in step S118, and are output to the main compression image data memory 114 as the main compression image data of these frames ( S120). When the main compression image data of these frames is already stored in the main compression image data memory 114, the data being stored is updated with the processed data. The current frame data is transferred to the main compression image data memory 114 as it is. Then, the process proceeds to step S122.

  For the latter half of the processing, the operation control mode control unit 190 performs the following control on the other functional blocks of the image processing apparatus 100.

A. The data supply unit 110 is switched to use the main compression image data memory 114. That is, the switching unit 116 outputs frame data (main compression image data) from the main compression image data memory 114 (S122).
B. The frame memory 150 is switched to use the main compression prediction frame memory 154 (S124). That is, the frame data from the inverse orthogonal transform unit 142 is stored in the main compression prediction frame memory 154 and output from the main compression prediction frame memory 154 to the motion compensation processing unit 160.
C. The rate adjustment unit 132 switches the bit rate adjustment method so as to be the fixed bit rate method (S126).
D. In the stream output control unit 180, switching is performed so as to output the bit stream (S128).

  Then, the latter half processing (main compression) is performed for the current frame (S130).

  As described above, the image compression apparatus 100 according to the present embodiment performs compression in the order of pre-compression → main compression image data generation → main compression. When generating the image data for main compression, a part of the display area of the frame whose bit rate obtained by the pre-compression exceeds the threshold Th is divided into a plurality of areas, one for each non-display area of the immediately preceding plurality of frames. make a copy. At the time of the main compression, for example, the data in the area 601 of the current frame 501 shown in FIG. 7 has already been encoded as a part of the area 701 of the previous frame 502, so the code amount of the current frame 501 is reduced by inter-frame prediction. . Further, since the processing is performed on the frame immediately before the scene chain frame, it is difficult for the human eye to perceive it, so that deterioration of image quality can be reduced.

  The present invention has been described above based on the embodiment. The embodiment is an exemplification, and various changes and increases / decreases may be added without departing from the gist of the present invention. It will be understood by those skilled in the art that modifications to which these changes and increases / decreases are also within the scope of the present invention.

It is a figure which shows the image compression apparatus concerning embodiment of this invention. It is a figure which shows the functional block in connection with the first half process in the image compression apparatus shown in FIG. It is a figure which shows the functional block concerning the latter half process in the image compression apparatus shown in FIG. H. 2 is a diagram for explaining cropping processing in the H.264 / AVC format. FIG. It is a figure which shows the flame | frame in which the bit rate was stored in the bit rate value preservation | save memory in the image process part of the image compression apparatus shown in FIG. FIG. 3 is a diagram for explaining a processing process by an image processing unit of the image compression apparatus illustrated in FIG. 1 (part 1); FIG. 8 is a diagram for explaining a processing process by an image processing unit of the image compression apparatus illustrated in FIG. 1 (part 2); FIG. 3 is a diagram for explaining a processing process by an image processing unit of the image compression apparatus shown in FIG. 1 (No. 3). FIG. 4 is a diagram for explaining a processing process by an image processing unit of the image compression apparatus shown in FIG. 1 (No. 4). 3 is a flowchart showing a flow of processing performed in the image compression apparatus shown in FIG. 1.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Image compression apparatus 110 Data supply part 112 Pre compression image data memory 114 Main compression image data memory 116 Switching part 120 Adder 122 Orthogonal transformation part 124 Quantization part 126 Variable length encoding part 130 Rate control part 132 Rate adjustment part 134 Rate evaluation unit 140 Inverse quantization unit 142 Inverse orthogonal transformation unit 150 Frame memory 152 Pre-compression prediction frame memory 154 Main compression prediction frame memory 160 Motion compensation processing unit 170 Image processing unit 172 Processing execution unit 174 Bit rate value storage memory 180 Stream output control unit 182 Stream buffer 190 Operation mode control unit 302 Non-display area 304 Display area

Claims (8)

Of the plurality of frames constituting the moving image, the frame immediately before the scene change frame that is a frame in which the bit rate when the moving image is compressed by the variable bit rate method exceeds a predetermined threshold is set as the target frame, and the scene change An image processing unit for performing processing for copying a partial area of the frame to a target frame corresponding to the scene change frame;
A compression unit that performs compression by a fixed bit rate method using a frame that has been processed by the image processing unit for a target frame, and an original frame for a frame other than the target frame. An image encoding device.   The compression unit is H.264. The image encoding apparatus according to claim 1, wherein the image encoding apparatus conforms to H.264 / AVC.   3. The image encoding according to claim 2, wherein the image processing unit copies the partial area in the display area of the scene change frame to a non-display area of the target frame corresponding to the scene change frame. apparatus.   The image processing unit takes n frames (n ≧ 2) immediately before the scene change frame as a target frame, and divides the partial area of the scene change frame into n to correspond to the scene change frame. 4. The image encoding device according to claim 1, wherein the image encoding device copies one by one to n target frames. 5. A pre-compression of the moving image by a variable bit rate method, further comprising an operation mode control unit that causes the compression unit to perform the pre-compression using the original frame for all frames;
5. The image code according to claim 1, wherein the image processing unit sets, as the scene change frame, a frame in which a bit rate obtained by the pre-compression exceeds the predetermined threshold value. 6. Device.   The said image process part calculates the magnitude | size of the said one part area | region according to the quantity by which the said bit rate of the scene change frame exceeded the said predetermined threshold value. The image encoding device according to item. Of the plurality of frames constituting the moving image, the frame immediately before the scene change frame that is a frame in which the bit rate when the moving image is compressed by the variable bit rate method exceeds a predetermined threshold is set as the target frame, and the scene change Processing to copy a part of the frame to the target frame corresponding to the scene change frame,
An image encoding method, wherein the target frame is processed and the original frame is used for frames other than the target frame, and compression is performed by a fixed bit rate method. Of the plurality of frames constituting the moving image, the frame immediately before the scene change frame that is a frame in which the bit rate when the moving image is compressed by the variable bit rate method exceeds a predetermined threshold is set as the target frame, and the scene change Processing to copy a part of the frame to the target frame corresponding to the scene change frame,
A program that causes a computer to execute a compression process using a fixed bit rate method by using the processed target frame and using the original frame for frames other than the target frame.

Images