JP2005277982A - Image compression processor - Google Patents

Abstract

【Task】
An image compression processing apparatus that processes compression-encoded image data, and reliably controls the bit rate of output data to be equal to or lower than a target bit rate.
[Solution]
When the bit rate determination processing means determines that the effective bit rate exceeds the target bit rate based on the compression-encoded image data, the data deletion timing is supplied. The compressed data deleting unit deletes predetermined compressed data with reference to the data deletion timing supplied from the bit rate determining unit. The quantization setting means refers to the data deletion timing supplied from the bit rate determination means, and sets the quantization setting to be equal to or lower than the target bit rate in the compression encoding process of the next frame. The encoding type designation means refers to the data deletion timing supplied from the bit rate determination means, and switches the encoding type to I-VOP in the compression encoding process for the next frame.
[Selection] Figure 1

Description

  The present invention relates to an image compression processing device that processes compression-encoded image data, and more particularly to an image compression processing device capable of bit rate control.

  With digitalization in recent years, moving image data compression / decompression technology such as MPEG (Moving Picture Experts Group) has begun to be used in surveillance systems, DVD (Digital Versatile Disc), BS (Broadcasting Satellite) broadcasting, and the like. MPEG-4, which is a video compression method with high compression efficiency, is not limited to broadband such as ADSL (Asymmetric Digital Subscriber Line) and CATV (Cable Television), but also in narrowband such as mobile phones and ISDN (Integrated Services Digital Network). TV (Tele-Vision) telephone and video content distribution using the (Moving Picture Experts Group Phase 4) method are becoming widespread.

As the use of image transmission by digitization expands in this way, there is an increasing demand for transmitting high-quality images in various infrastructures. For example, a mobile phone TV phone has a line capacity of 64 kbps, and the bandwidth allocated for image transmission is about 40 kbps. Since audio and control data are transmitted in other bands, image data must be controlled to 40 kbps or less before transmission. On the other hand, moving image compression coding technology such as MPEG-4 is processed by assuming a certain number of data generated by compression coding, but it is a compression algorithm that is difficult to control with high accuracy such as tens of bytes. When compressing a moving image with constantly changing image complexity, it may output 1.5 to 2 times the expected bandwidth. In the example of the mobile phone described above, if the bit rate of image data exceeds 40 kbps, the image data, audio data, and other data are destroyed. If the image data is destroyed, if it is a predictive coding method that refers to the image data of the previous and subsequent frames, such as MPEG, the frame from which the compressed data is destroyed becomes a prediction reference frame (hereinafter referred to as I-VOP). The image is continuously disturbed and collapsed (refer to Patent Document 1).
Henceforth, the frame which performs predictive coding other than I-VOP is called P-VOP. I-VOP is an inter-coded VOP (VOP: Video Object Plane), and P-VOP is a forward-predicted VOP.

JP 2003-309841 A

As described above, the image compression processing device is a compression algorithm that makes it difficult to accurately control the number of compression-encoded data, so that more data than the limited bandwidth is output from the image compression processing device. Data may be destroyed, and it may be difficult to reliably maintain high-quality image transmission.
An object of the present invention is to provide an image compression processing apparatus that solves the above-described problems and does not cause image distortion or collapse.

In order to achieve the above object, an image compression processing apparatus of the present invention processes image data that has been compression-encoded as follows.
That is, when the bit rate determination processing means determines that the effective bit rate exceeds the target bit rate based on the compression-encoded image data, the data deletion timing is supplied. Then, the compressed data deleting unit deletes predetermined compressed data with reference to the data deletion timing supplied from the bit rate determining unit. Further, the quantization setting means refers to the data deletion timing supplied from the bit rate determination means, and sets the quantization setting to be equal to or lower than the target bit rate in the compression encoding process of the next frame. The encoding type designation means refers to the data deletion timing supplied from the bit rate determination means, and switches the encoding type to I-VOP in the compression encoding process for the next frame.
In this way, it has a bit rate determination processing unit that controls so that the effective bit rate of the compressed data by compression encoding does not exceed the given band, and deletes the image data as necessary, while maintaining high quality Reliably maintain image transmission.
Therefore, it is possible to control the bit rate of compression-encoded image data within a bit rate determined by the system, and to realize a bit rate control function without causing image distortion or collapse. Is possible.

Various types of image data to be processed may be used. For example, moving image data composed of a plurality of temporally continuous frames, still image data that is not temporally continuous, etc. Can be used.
Various means may be used as means for supplying the data deletion timing from the image data to be processed. For example, information can be input from an external bit rate measuring device or the like.
Further, as a condition for supplying the data deletion timing, not only when the effective bit rate exceeds the target bit rate, but also a deletion timing supply condition based on the ratio of the effective bit rate to the target bit rate, etc. It is possible to make the conditions satisfy the required specifications.
In addition, not only the target bit rate but also the deletion timing supply condition based on the ratio of the effective bit rate to the transmission channel bit rate is satisfied as the condition for supplying the data deletion timing, and the required specifications of the image compression processing apparatus and system are satisfied. It can be a condition.

  As described above, according to the image compression processing device of the present invention, the effective bit rate is acquired from the compression-encoded image data, and the effective bit rate is set to the target bit rate using the acquired effective bit rate. Since the compression encoded data is deleted when it exceeds, the encoding type of the next frame to be compressed is set to I-VOP, and the quantization setting of that frame is set to be less than the target bit rate. It is possible to realize a rate control function that does not cause image distortion / disintegration while reliably maintaining a bit rate lower than or equal to the compression-coded data.

That is, in the image compression apparatus of the present invention, the bit rate determination processing means includes means for supplying data deletion timing when the effective bit rate of the compressed data output from the image compression processing unit exceeds the target bit rate.
In the image compression apparatus according to the present invention, the compressed data deletion means includes means for deleting the compressed data to be output from the image compression processing apparatus in synchronization with the data deletion timing.
Further, in the image compression apparatus of the present invention, the quantization setting means includes the number of compressed data so that the effective bit rate is equal to or less than the target bit rate in the compression encoding in the next frame of the frame in which the data deletion timing has occurred. Means for adjusting the quantization setting to reduce.
In the image compression apparatus of the present invention, the compression-encoded image data to be processed is image data that has been compression-encoded using predictive encoding.

An embodiment of the present invention will be described with reference to FIG. FIG. 1 is a block diagram showing a configuration of an embodiment of an image compression apparatus of the present invention.
In the embodiment shown in FIG. 1, for example, a compression encoding method using predictive encoding such as MPEG-4 (ISO / IEC 14496) is used. 1-1 is the first adder, 1-2 is the DCT (Discrete Cosine Transform) / quantization processing unit, 1-3 is the variable length coding processing unit, 1-4 is the bit rate determination processing unit, 1-5 Is a data output switch, 1-6 is a coding type changeover switch, 1-7 is an inverse quantization / inverse DCT processing unit, 1-8 is a second adder, 1-9 is a frame buffer, 1-10 is a motion A search processing unit, 1-11 is a motion compensation prediction processing unit, 100 is an image data input terminal, and 200 is a compressed data output terminal. 1- (A) is image data, 1- (B) is difference image data, 1- (C) is quantized data, 1- (D) is compressed data, 1- (E) is the number of compressed data, 1- (F) is output compressed data, 1- (G) is inverse DCT data, 1- (H) is added image data, 1- (I) is delayed image data 1- (J) is P-VOP encoded Motion prediction data, 1- (K) is I-VOP encoded motion prediction data, 1- (L) is motion prediction data, 1- (M) is motion information, 1- (N) is a data reduction timing signal It is.

An example of operations performed on moving image data will be described with reference to FIG.
First, the processing of the first frame of moving image data will be described. In compression encoding methods using predictive encoding such as MPEG-4, compression encoding to I-VOP that does not perform prediction from the previous and subsequent frames is performed periodically, and the frame that serves as a reference for prediction (I-VOP) is generated periodically. By inserting it periodically, the image quality can be maintained to some extent, and the image can be restored from image distortion and collapse due to an error. For this reason, since the first frame needs to be a prediction reference, the encoding type is I-VOP.
The image data 1- (A) is input to the input terminal 100 in FIG. 1 successively one after another in units of frames.

When the first image data 1- (A) is input to the image compression processing device via the input terminal 100, the image data 1- (A) is first converted into the first adder 1-1 and the motion search processing unit 1-10. Is input.
Since the first (that is, the first frame) image data 1- (A) has the encoding type I-VOP, the encoding type changeover switch 1-6 uses I as the motion prediction data 1- (L). -Output motion prediction data 1- (K) for VOP coding and prediction data 1- (L).
The encoding type changeover switch 1-6 is initially set so that input a and output c are connected, and after I-VOP encoding, input b and output c are connected during the specified period P-VOP process. Can be switched automatically. The motion prediction data 1- (L) that is the output of the coding type changeover switch 1-6 is output to the first adder 1-1 and the second adder 1-8.

The value of I-VOP coded motion prediction data 1- (K) is always “0”. Therefore, the first adder 1-1 subtracts the motion prediction data 1- (L) having the value “0” from the input image data 1- (A), and outputs the difference image data 1- (B). Will do. That is, the same data as the image data 1- (A) input to the input terminal 100 is output to the DCT / quantization processing unit 1-2 as the difference image data 1- (B).
The DCT / quantization processing unit 1-2 performs DCT processing and quantization processing on the input differential image data 1- (B), and converts the quantized data 1- (C) into a variable length coding processing unit 1-3. And the inverse quantization / inverse DCT processing section 1-7.

The inverse quantization / inverse DCT processing unit 1-7 performs inverse quantization processing and inverse DCT processing on the input quantized data 1 (C), and adds the inverse DCT data 1- (G) to the second addition. Output to machine 1-8.
Also, motion prediction data 1- (L) is input to the other input of the second adder 1-8, but since the encoding type is I-VOP, motion prediction data 1- (L) is , I-VOP encoded motion prediction data 1- (K) (that is, the value “0”).

The second adder 1-8 adds the inverse DCT data 1- (G) and the motion prediction data 1- (L) and outputs the added image data 1- (H). That is, since the value of the input motion prediction data 1- (L) is “0”, the second adder 1-8 adds the same data as the input inverse DCT data 1- (G) to the addition image. Output as data 1- (H) to frame buffer 1-9.
The frame buffer 1-9 accumulates the input added image data 1- (H) until the next frame processing, and then outputs it to the motion compensation prediction processing unit 1-11.

On the other hand, the motion search processing unit 1-10 does not perform motion detection processing at the time of I-VOP encoding. Therefore, the motion information 1- (M) is information indicating “no motion”, and this information is output to the motion compensation prediction processing unit 1-11 and the variable length coding processing unit 1-3.
The variable length coding processing unit 1-3 is given the quantized data 1- (C) and the motion information 1- (M), and the input quantized data 1- (C) and the motion information 1- (M) Variable-length coding processing is performed, compressed data 1- (D) is output to the input terminal of the data output switch 1-5, and the number of compressed data 1- (E) is sent to the bit rate determination processing unit 1-4. Output.

The bit rate determination processing unit 1-4 calculates the effective bit rate of the compressed data 1- (D) from the result of adding the number of compressed data 1- (E) over the past n frames, and the effective bit rate is the target bit rate. (N is a natural number). Here, assuming that the target bit rate is not exceeded, the bit rate determination processing unit 1-4 holds 1- (E) over the past n frames including the current frame, and does not send any signal. Do not output.
The data output switch 1-5 that does not receive a signal from the bit rate determination processing unit 1-4 outputs the compressed data 1- (D) as output compressed data 1- (F) from the image compression processing device via the output terminal 200. To do.

Next, processing after the second frame of moving image data will be described. In the second and subsequent frames, encoding is performed using the P-VOP encoding type for predictive encoding, and P-VOP encoding is repeated until the I-VOP encoding period is reached.
In FIG. 1, when image data 1- (A) of the first second frame is input to the image compression processing apparatus, it is first input to the first adder 1-1 and the motion search processing unit 1-10. .
The motion search processing unit 1-10 searches the first frame of the image data 1- (A) for a location that has moved, and uses the information as motion information 1- (M), which is variable with the motion compensation prediction processing unit 1-11. Output to long encoding processing unit 1-3. Here, the added image data 1- (H) accumulated in the frame buffer 1-9 in the first frame is output to the motion compensation prediction processing unit 1-11 as delayed image data 1- (I).

The motion compensated prediction processing unit 1-11 uses the delayed image data 1- (I) and the motion information 1- (M) to generate an image of coordinates that would have generated motion from the delayed image data 1- (I). Is output to the input terminal b of the encoding type changeover switch 1-6 as motion prediction data 1- (J) for P-VOP encoding.
Since the encoding type is P-VOP, the encoding type selector switch 1-6 is in a state where input terminal b and output terminal c are connected, and P- is used as motion prediction data 1- (L). VOP encoded motion prediction data 1- (J) is output from output terminal c.

The first adder 1-1 subtracts the motion prediction data 1- (L) from the input image data 1- (A) and converts the difference image data 1- (B) into the DCT / quantization processing unit 1 Output to -2.
The DCT / quantization processing unit 1-2 performs DCT processing and quantization processing on the difference image data 1- (B), and performs inverse quantization with the variable-length coding processing unit 1-3 as quantized data 1- (C). / Outputs to reverse DCT processing section 1-7.
The inverse quantization / inverse DCT processing unit 1-7 performs inverse quantization processing and inverse DCT processing on the quantized data 1 (C), and outputs the result to the second adder 1-8. Here, the coding type changeover switch 1-6 gives the motion prediction data 1- (L) to the second adder 1-8. Since the coding type is P-VOP, the motion prediction data 1- ( L) is P-VOP coded motion prediction data 1- (J).

  The second adder 1-8 adds the inverse DCT data 1- (G) and the motion prediction data 1- (L), and outputs the added image data 1- (H) to the frame buffer 1-9. The frame buffer 1-9 accumulates the added image data 1- (H) until the next frame processing, and then outputs it to the motion compensation prediction processing unit 1-11.

On the other hand, the variable length coding processing unit 1-3 that receives the quantized data 1- (C) and motion information 1- (M) receives the quantized data 1- (C) and motion information 1- (M). The variable length encoding process is performed on the data, compressed data 1- (D) is input to the data output switch 1-5, and the number of compressed data 1- (E) is output to the bit rate determination processing unit 1-4.
The bit rate determination processing unit 1-4 calculates the effective bit rate of the compressed data 1- (D) from the result of adding the number of compressed data 1- (E) over the past n frames, and the effective bit rate is the target bit. Determine if the rate is exceeded. Here, assuming that the target bit rate is not exceeded, the bit rate determination processing unit 1-4 holds the number of compressed data 1- (E) over the past n frames including the current frame. Does not output any signal. The data output switch 1-5 that does not receive a signal from the bit rate determination processing unit 1-4 uses the compressed data 1- (D) as the output compressed data 1- (F) from the image compression processing device via the output terminal 200. Output.
The same processing as described above is repeated for the third and subsequent frames.

  Here, a case where the effective bit rate exceeds the target bit rate as a result of the processing in the bit rate determination processing unit 1-4 will be described. This is the same process regardless of I-VOP encoding and P-VOP encoding. When the effective bit rate exceeds the target bit rate, the bit rate determination processing unit 1-4 sends the data deletion timing signal 1- (N) to the data output switch 1-5, the DCT / quantization processing unit 1-2, and Output to encoding type selector switch 1-6.

  Data output switch 1-5 that receives data deletion timing signal 1- (N) opens the switch circuit and does not output compressed data 1- (D) of the current frame. That is, no data is output from the compression encoding processing unit in the compression encoding processing of the current frame.

  In addition, the encoding type changeover switch 1-6 to which the data deletion timing signal 1- (N) is input forcibly sets the encoding type of the next frame to I-VOP encoding. That is, the switch is switched so that the motion prediction data 1- (L) of the output of the next frame becomes the motion prediction data 1- (K) of I-VOP coding.

  In addition, the DCT / quantization processing unit 1-2 receiving the data deletion timing signal 1- (N) receives an extremely small amount of data of the quantized data 1- (C) to be output in the next frame. Adjust the quantization settings.

Next, an example of the flow of compressed data output processing performed by the image compression processing apparatus of this example will be described with reference to FIGS. FIG. 2 is a diagram showing an example of image compression processing when the present invention is implemented, and FIG. 4 is a diagram showing an example of image compression processing when the present invention is not implemented. 2 and 4, the horizontal axis indicates time, and the vertical axis indicates (a) output data 1- (F), (b) compressed data 1- (E) (unit: byte / frame), ( c) indicates the execution bit rate (unit: bit / sec).
FIG. 3 is a block diagram showing the configuration of the image compression processing apparatus that reaches the result of FIG. The configuration of FIG. 3 is the same as the configuration of FIG. 1 except for the bit rate determination processing unit 1-4, the data output switch 1-5, and the connection lines between them.

  FIG. 2 shows an example of a specific output frame for the output compressed data 1- (F) shown in FIG. 1, and as an example, the number of compressed data 1- (E) and the bit rate determination processing unit The transition of the effective bit rate calculated in 1-4 is shown. The horizontal direction in FIG. 2 represents the first to thirteenth frames as a temporal processing flow (flow in frame units).

  The first frame 2-1 to the fifth frame 2-5 output from the image compression processing apparatus are cases where the effective bit rate does not exceed the target bit rate. Next, assuming that the sixth frame 2-6 is output from the image compression processing unit, the bit rate determination processing unit 1-4 determines that the effective bit rate exceeds the target bit rate. This corresponds to the broken line in the sixth frame in FIG. 2, but is the state when the phenomenon of the sixth frame 4-6 in FIG. 4 occurs. At this time, the sixth frame 2-6 is not output from the image compression processing unit and is deleted in the image compression processing unit. For this reason, the number of compressed data 1- (E) is 0 [byte / frame], and the effective bit rate rapidly decreases. This corresponds to the solid line in the sixth frame in FIG.

Next, in the seventh frame, since it is the next frame from which the output compressed data 1- (F) has been deleted, it is forcibly set to the I-VOP encoding type. In addition, since the quantization is forcibly set so that the amount of generated codes is low, the size of the compressed data 1- (E) is also small for the I-VOP. For this reason, it is avoided that the effective bit rate rapidly increases and exceeds the target bit rate. For example, the quantization setting is 1.3, assuming that the compression rate of I-VOP in normal time is 1.6. This compression average rate is the result of modeling, for example, on condition that the target bit rate is 800 kbps, the frame rate is 30 fps, the resolution is VGA, and the intra rate is 30 frames.
From the 8th frame onward, the effective bit rate does not exceed the target bit rate, so the I-VOP encoding type is not forced.

  As described above, the image compression processing apparatus of this example performs compression coding processing on an input image, and outputs compressed data generated thereby to transmission means having a limited transmission capacity. If it is determined that the execution bit rate exceeds the transmission capacity by continuously monitoring the amount of generated data and outputting the compressed data to the transmission means, for example, automatically deleting the compressed data without outputting it . Further, in the image compression processing apparatus of this example, in the compression processing of the next frame from which the compressed data is deleted, a function for setting the quantization type so that the number of compressed data is reduced by using a coding type that is a reference for predictive coding It is equipped with.

  As described above, the image compression processing apparatus of the present invention realizes a rate control function for reliably controlling the bit rate below the target bit rate by deleting compressed data exceeding the target bit rate. In addition, by forcing the compression encoding process of the next frame from which the compressed data is deleted to be the encoding type that is a reference for predictive encoding, it is possible to reliably avoid data disturbance and corruption due to data reduction. . Furthermore, it is possible to enhance the avoidance of the target bit rate over by adjusting the quantization setting so that the generated code amount is reduced in the frame that is forcibly set as the reference encoding type.

Here, in this example, the case where moving image data composed of a plurality of continuous frames is processed has been shown. However, there is no particular limitation on image data to be processed, It is also possible to process a frame in which still image frame data is continuously arranged.
In the above embodiment, the present invention is applied to the image compression processing apparatus.

  Further, in the image compression processing apparatus of this example, the bit rate determination processing means is configured by the function of the bit rate determination processing unit 1-4, and the compressed data deletion means is configured by the data output switch 1-5. The DCT / quantization processing unit 1-2 constitutes quantization setting means, and the coding type changeover switch 1-6 constitutes coding type designation means.

The background of the technology related to the present invention will be described below. Note that the matters described here are not necessarily limited to the conventional technology.
An example of a conventional image compression apparatus will be described with reference to FIG. FIG. 3 is a block diagram showing the configuration of an embodiment of a conventional image compression apparatus.
FIG. 3 shows a case where a compression encoding method using predictive encoding such as MPEG-4 (ISO / IEC 14496) is used.
3 includes a first adder 1-1, a DCT / quantization processing unit 1-2, a variable length coding processing unit 1-3, a bit rate determination processing unit 1-4, , Data output switch 1-5, encoding type changeover switch 1-6, inverse quantization / inverse DCT processing unit 1-7, second adder 1-8, frame buffer 1-9, motion A search processing unit 1-10 and a motion compensation prediction processing unit 1-11 are provided.

First, the processing of the first frame of moving image data will be described. As described above, since the first frame needs to be a prediction reference, the encoding type is I-VOP.
When the first image data 1- (A) is input to the image compression processing device via the input terminal 100, the image data 1- (A) is first converted into the first adder 1-1 and the motion search processing unit 1-10. Is input.
Since the first image data 1- (A) has the encoding type I-VOP, the encoding type changeover switch 1-6 uses I-VOP encoding motion prediction data as motion prediction data 1- (L). Output 1- (K) as prediction data 1- (L).
The encoding type changeover switch 1-6 is initially set so that input a and output c are connected, and after I-VOP encoding, input b and output c are connected during the specified period P-VOP process. Can be switched automatically. The motion prediction data 1- (L) that is the output of the coding type changeover switch 1-6 is output to the first adder 1-1 and the second adder 1-8.

The value of I-VOP coded motion prediction data 1- (K) is always “0”. Therefore, the first adder 1-1 subtracts the motion prediction data 1- (L) having the value “0” from the input image data 1- (A), and outputs the difference image data 1- (B). Will do. That is, the same data as the image data 1- (A) input to the input terminal 100 is output to the DCT / quantization processing unit 1-2 as the difference image data 1- (B).
The DCT / quantization processing unit 1-2 performs DCT processing and quantization processing on the input differential image data 1- (B), and converts the quantized data 1- (C) into a variable length coding processing unit 1-3. And the inverse quantization / inverse DCT processing section 1-7.

The value of I-VOP coded motion prediction data 1- (K) is always “0”. Therefore, the first adder 1-1 subtracts the motion prediction data 1- (L) having the value “0” from the input image data 1- (A), and outputs the difference image data 1- (B). Will do. That is, the same data as the image data 1- (A) input to the input terminal 100 is output to the DCT / quantization processing unit 1-2 as the difference image data 1- (B).
The DCT / quantization processing unit 1-2 performs DCT processing and quantization processing on the input differential image data 1- (B), and converts the quantized data 1- (C) into a variable length coding processing unit 1-3. And the inverse quantization / inverse DCT processing section 1-7.

The inverse quantization / inverse DCT processing unit 1-7 performs inverse quantization processing and inverse DCT processing on the input quantized data 1 (C), and adds the inverse DCT data 1- (G) to the second addition. Output to machine 1-8.
Also, motion prediction data 1- (L) is input to the other input of the second adder 1-8, but since the encoding type is I-VOP, motion prediction data 1- (L) is , I-VOP encoded motion prediction data 1- (K) (that is, the value “0”).

The second adder 1-8 adds the inverse DCT data 1- (G) and the motion prediction data 1- (L) and outputs the added image data 1- (H). That is, since the value of the input motion prediction data 1- (L) is “0”, the second adder 1-8 adds the same data as the input inverse DCT data 1- (G) to the addition image. Output as data 1- (H) to frame buffer 1-9.
The frame buffer 1-9 accumulates the input added image data 1- (H) until the next frame processing, and then outputs it to the motion compensation prediction processing unit 1-11.

On the other hand, the motion search processing unit 1-10 does not perform motion detection processing at the time of I-VOP encoding. Therefore, the motion information 1- (M) is information indicating “no motion”, and this information is output to the motion compensation prediction processing unit 1-11 and the variable length coding processing unit 1-3.
The variable length coding processing unit 1-3 is given the quantized data 1- (C) and the motion information 1- (M), and the input quantized data 1- (C) and the motion information 1- (M) Variable-length coding processing is performed, compressed data 1- (D) is output to the input terminal of the data output switch 1-5, and the number of compressed data 1- (E) is sent to the bit rate determination processing unit 1-4. Output.

Next, processing after the second frame of moving image data will be described. In the second and subsequent frames, encoding is performed using the P-VOP encoding type for predictive encoding, and P-VOP encoding is repeated until the I-VOP encoding period is reached.
In FIG. 3, when image data 1- (A) of the first second frame is input to the image compression processing apparatus, it is first input to the first adder 1-1 and the motion search processing unit 1-10. .
The motion search processing unit 1-10 searches the first frame of the image data 1- (A) for the moved part, and uses the information as motion information 1- (M), which is variable with the motion compensation prediction processing unit 1-11. Output to long encoding processing unit 1-3. Here, the added image data 1- (H) accumulated in the frame buffer 1-9 in the first frame is output to the motion compensation prediction processing unit 1-11 as delayed image data 1- (I).

The motion compensated prediction processing unit 1-11 uses the delayed image data 1- (I) and the motion information 1- (M) to generate an image of coordinates that would have generated motion from the delayed image data 1- (I). Is output to the input terminal b of the encoding type changeover switch 1-6 as motion prediction data 1- (J) for P-VOP encoding.
Since the encoding type is P-VOP, the encoding type selector switch 1-6 is in a state where input terminal b and output terminal c are connected, and P- is used as motion prediction data 1- (L). VOP encoded motion prediction data 1- (J) is output from output terminal c.

The first adder 1-1 subtracts the motion prediction data 1- (L) from the input image data 1- (A) and converts the difference image data 1- (B) into the DCT / quantization processing unit 1 Output to -2.
The DCT / quantization processing unit 1-2 performs DCT processing and quantization processing on the difference image data 1- (B), and performs inverse quantization with the variable-length coding processing unit 1-3 as the quantized data 1- (C). / Outputs to reverse DCT processing section 1-7.
The inverse quantization / inverse DCT processing unit 1-7 performs inverse quantization processing and inverse DCT processing on the quantized data 1 (C), and outputs the result to the second adder 1-8. Here, the coding type changeover switch 1-6 gives the motion prediction data 1- (L) to the second adder 1-8. Since the coding type is P-VOP, the motion prediction data 1- ( L) is P-VOP coded motion prediction data 1- (J).

  The second adder 1-8 adds the inverse DCT data 1- (G) and the motion prediction data 1- (L), and outputs the added image data 1- (H) to the frame buffer 1-9. The frame buffer 1-9 accumulates the added image data 1- (H) until the next frame processing, and then outputs it to the motion compensation prediction processing unit 1-11.

On the other hand, the variable length coding processing unit 1-3 that receives the quantized data 1- (C) and motion information 1- (M) receives the quantized data 1- (C) and motion information 1- (M). On the other hand, variable length coding processing is performed, and output compressed data 1- (F) is output from the image compression processing apparatus via the output terminal 200.
The same processing as described above is repeated for the third and subsequent frames.

Next, an example of the flow of processing performed by the image compression processing apparatus shown in FIG. 3 will be described with reference to FIG.
FIG. 4 shows an example of a specific output frame for the output compressed data 1- (F) shown in FIG. 3, and the transition of the number of compressed data and the effective bit rate is shown as an example. The horizontal direction in FIG. 4 represents the first to thirteenth frames as a temporal processing flow (flow in frame units). The first frame 4-1 to the fifth frame 4-5 output from the image compression processing apparatus are cases where the effective bit rate does not exceed the target bit rate. Here, as an example, assume that the sixth frame 4-6 is output from the image compression processing unit, and the effective bit rate exceeds the target bit rate. The 6th frame 4-6 may be damaged because it tries to output more than the bit rate of the line. If the sixth frame 4-6 is damaged, it will be distorted and collapsed, and even after the seventh frame 4-7, the disturbance and collapse of the sixth frame 4-6 will be inherited.

  This disturbance and collapse are resolved when the 11th frame 4-11, which is the next I-VOP, is output. In a low bit rate line such as a narrow band, since the I-VOP code amount is large, the I-VOP insertion period tends to be as long as possible. At this time, if the effective bit rate exceeds the target bit rate, a long time is required until the image distortion / disintegration is resolved.

Here, the configuration of the image compression processing apparatus according to the present invention is not necessarily limited to the above-described configuration, and various configurations may be used. The invention can also be provided, for example, as a method or method for executing the processing according to the present invention, a program for realizing such a method or method, a recording medium for recording the program, and the like. Also, it can be provided as various devices and systems.
In addition, the application field of the present invention is not necessarily limited to the above-described fields, and the present invention can be applied to various fields.
In addition, as various processes performed in the image compression processing apparatus according to the present invention, for example, the processor executes a control program stored in a ROM (Read Only Memory) in a hardware resource including a processor and a memory. For example, various functional means for executing the processing may be configured as an independent hardware circuit.
The present invention can also be understood as a computer-readable recording medium such as a flexible disk or a CD (Compact Disk) -ROM storing the above control program, or the program (itself). The processing according to the present invention can be performed by inputting a medium into a computer and causing the processor to execute it.

1 is a block diagram showing a configuration of an image compression processing apparatus according to an embodiment of the present invention. The figure which shows an example of the process performed by the image compression processing apparatus of one Example of this invention. The figure which shows the structural example of the conventional image compression processing apparatus. The figure which shows an example of the process performed by the conventional image compression processing apparatus.

Explanation of symbols

1-1: First adder 1-2: DCT / quantization processing unit 1-3: Variable length coding processing unit 1-4: Bit rate determination processing unit 1-5: Data output switch 1-6: Encoding type selector switch, 1-7: Inverse quantization / inverse DCT processing unit, 1-8: Second adder, 1-9: Frame buffer, 1-10: Motion search processing unit, 1 -11: Motion compensation prediction processing unit, 100: Image data input terminal, 200: Compressed data output terminal, 1- (A): Image data, 1- (B): Difference image data, 1- (C): Quantization Data, 1- (D): Compressed data, 1- (E): Number of compressed data, 1- (F): Output compressed data, 1- (G): Inverse DCT data, 1- (H): Addition image data 1- (I): Delayed image data, 1- (J): P-VOP encoded motion prediction data, 1- (K): I-VOP encoded motion prediction data, 1- (L): Motion Predicted data, 1- (M): motion information, 1- (N): data reduction timing signal.

Claims (2)

In an image compression processing apparatus that processes data of a compression-encoded image,
Compressed data deleting means for deleting predetermined compressed data when the effective bit rate is greater than a predetermined value;
Deleting with the compressed data deleting means, and coding type designation means for switching the coding type to intra-frame coding in the compression coding processing of the next frame;
An image compression processing apparatus comprising: 2. The image compression processing apparatus according to claim 1, wherein the compressed data deleting means is a compressed data deleting means for deleting predetermined compressed data when an effective bit rate becomes larger than a transmission channel bit rate. Image compression processing device.