JP2003092762A - Image encoding data converter - Google Patents

Abstract

PROBLEM TO BE SOLVED: To solve problems of a conventional image encoding data converter that deteriorates image quality due to conversion and increases a delay of the conversion processing resulting in extending the scale of the converter because encoding processing is again applied to image encoding data independently of contents of a decoded image signal after decoding the image encoding data. SOLUTION: An image encoding data analysis section 310 applies first digital signal processing to first image encoding data 220 to generate encoded data 221 after signal processing, an image encoded data composite section 320 receives coded data after the signal processing and a plurality of information items 222 with respect to the fit image encoded data, and applies second digital signal processing to the encoded data after signal processing on the basis of a plurality of information items associated with the first image encoded data to produce second image encoded data 240.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention receives, as an input, first image coded data obtained by performing a coding process on a digitized input image signal, and inputs the first image coded data into a digital signal. The present invention relates to an image coded data conversion device that performs processing and outputs second image coded data.

[0002]

2. Description of the Related Art FIG. 16 shows, for example, Japanese Patent Laid-Open No. 2-17918.
It is a block diagram which shows the conventional image coding data conversion apparatus shown by the 6th publication, and has illustrated the multipoint video conference system. In the figure, 1 is a transmitting station, 2 is a relay station, 3 is a receiving station, 10 is an image encoder of each station, and 20 is an image decoder. Further, 101 is an input image, 102 is image encoded data, 103 is a decoded image signal, 104 is image re-encoded data, and 105 is a decoded image signal.

Next, the operation will be described. The input image 101 input by the image encoder 10 of the transmitting station 1 is encoded, and the obtained image encoded data 102 is sent to the relay station 2. In the relay station 2, the image coded data 10
2 is received by the image decoder 20 and is subjected to decoding processing, and the decoded image signal 103 thus obtained is subjected to re-coding processing by the image encoder 10 to generate image re-encoded data 104. In this way, the image re-encoded data 104 generated by the re-encoding process is sent to the destination station 3, and the destination station 3 transmits an image to the image re-encoded data 104 relayed by the relay station 2. The decoder 20 performs a decoding process and uses it as the decoded image signal 105.

When a video conference is conducted using the relay station 2 having such a decoding relay function, the transmitting station 1 and the receiving station 3
If the encoding methods are different between the two, the relay station 2 temporarily decodes the received image encoded data 102 in order to change the amount of encoded data generated and various parameters (image size, frame rate, etc.). After the decoded image signal 103 is obtained in this way, it is re-encoded into the image re-encoded data 104 to achieve matching.

As described above, in the above-mentioned conventional image coded data conversion apparatus, in order to relay or copy the image coded data, the image coded data is once decoded and then re-encoded. .

[0006]

Since the conventional image coded data conversion apparatus is configured as described above, it is temporarily converted into the decoded image signal 103 in order to relay or convert the image coded data 102. After decoding, the decoded image signal 1
Since the method of performing the encoding process again regardless of the contents of 03, the image quality of the decoded image signal 105 after conversion is deteriorated, or the delay in the relay or conversion process is increased, or the device scale is expanded. There was such a problem.

The present invention has been made in order to solve the above problems, and reduces image quality deterioration, processing delay, device scale reduction, and conversion efficiency while improving image coding data. It is an object of the present invention to obtain an image coded data conversion device capable of performing the conversion.

[0008]

An image coded data conversion apparatus according to the present invention is provided with a motion search section, and the motion vector extracted by the image coded data analysis section according to the arrangement information of the images to be converted. It is designed to estimate the size.

The image coded data conversion apparatus according to the present invention is provided with a quantizer estimator, and decodes the quantization parameter when the first image decoded data is generated, which is output by the image coded data analyzer. It is estimated from the image.

[0010]

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of the present invention will be described below. The fourteenth and sixteenth embodiments described below
Respectively correspond to the inventions of claims 1 and 2 in the claims.

Embodiment 1. 1 is a block diagram showing an image coding data conversion apparatus according to a first embodiment of the present invention, an encoding processing apparatus for coding an input image signal digitally processed by the image coding data conversion apparatus, and the image coding apparatus. It is the block diagram shown with the decoding processing apparatus which performs the decoding processing of the image coding data processed by the data conversion apparatus. In the figure, 30 is an image coded data conversion device, 4
0 is an encoding processing device, 50 is a decoding processing device, and 20
Reference numeral 0 is a digitized input image signal input to the encoding processing device 40, and 220 is a first image encoding obtained by performing an encoding process on the input image signal 200 in the encoding processing device 40. Data 240 is the first image encoded data 22 in the image encoded data conversion device 30.
0 is the second image coded data obtained by performing digital signal processing, and 250 is the decoded image signal obtained by decoding the second image coded data 240 by the decoding processing device 50. .

Further, in the image coded data conversion apparatus 30, reference numeral 310 is an image coded data analysis section for performing the first digital signal processing on the first image coded data 220 from the coding processing apparatus 40. Yes, 320 receives the signal-processed coded data output from the image-coded data analysis unit 310 and a plurality of pieces of information regarding the first image-coded data, and with respect to the signal-processed coded data. , An image coded data synthesizing unit that generates second image coded data 240 by performing second digital signal processing based on a plurality of pieces of information regarding the first image coded data. 221 is coded data after signal processing which is sent from the image coded data analysis section 310 to the image coded data synthesizing section 320, and 222 is the first data.
Of a plurality of pieces of information regarding the image coded data of
Reference numeral 3 is information input to the image coded data synthesizing unit 320 to instruct conversion such as the amount of data to be converted and the image size to be converted.

Further, in the encoding processing device 40,
Reference numeral 401 denotes a converter that performs a calculation such as discrete cosine transform (hereinafter referred to as DCT) to generate a transform coefficient from the input image signal 200, and 402 performs scalar quantization processing on the transform coefficient generated by the converter 401. A quantizer 403 generates a quantization index, and 403 is the quantizer 40.
2 is a variable-length encoder that performs variable-length coding processing such as Huffman coding on the generated quantized index 2 to generate first image coded data 220. Further, in the decoding processing device 50, 501 is a variable length decoder that performs variable length decoding processing on the second image coded data 240, and 502 is the quantization index decoded by the variable length decoder 501. And 503 is an inverse quantizer that performs inverse scalar quantization processing, and an inverse transform 503 that performs an inverse DCT operation to obtain a decoded image signal 250 on the transform coefficient inversely quantized by the inverse quantizer 502. It is a vessel. As the encoding processing device 40 and the decoding processing device 50, only devices for realizing basic functions are shown.

Next, the operation will be described. The digitized input image signal 200 is input to an encoding processing device 40 including a converter 401, a quantizer 402, and a variable length encoder 403, and is subjected to encoding processing to obtain first image encoded data. 220. Hereinafter, the operation of the encoding processing device 40 will be described in detail.

FIG. 2 is a block diagram showing an example of a concrete configuration for explaining the operation of the encoding processing device 40.
In addition, 401 is a converter, 402 is a quantizer, and 403 is a variable length encoder, which realizes the basic function of encoding the input image signal 200 into the first image encoded data 220. The other part is added to realize the function of motion compensation prediction. In the figure, 404 is a current frame memory to be encoded in which the input image signal 200 is stored, and 405 is a current frame image 201 read from the current frame memory 404 to be encoded, from a motion compensation predictor described later. Prediction frame image 2
It is a subtracter that subtracts 10 to generate a prediction error frame image 202. Reference numeral 203 is a transform coefficient to which the result of the DCT operation of the prediction error frame image 202 by the transformer 401 is output, and 204 is generated by the scalar quantization of this transform coefficient 203 by the quantizer 402, and variable length coding is performed. Quantization index output to the device 403, 2
05 is also a quantization parameter. Reference numeral 406 denotes an inverse quantizer that inversely quantizes the quantization index 204 output from the quantizer 402, and 407 denotes the inverse quantizer 40.
Inverse DC to the transform coefficient 206 obtained by inverse quantization by
408 is an inverse converter that performs inverse conversion processing by T calculation.
Is an adder for adding the local prediction error frame image 207 generated by the inverse transform processing by the inverse transformer 407 and the predicted frame image 210 from the motion compensation predictor described later. Reference numeral 409 is a coded preceding frame memory in which the locally decoded frame image 208 which is the addition result of the adder 408 is stored, and 410 is the preceding frame image 20 read from this coded preceding frame memory 409.
9 and the current frame image 201 read from the current memory frame 404 to be encoded, the subtracter 4
05 and the predicted frame image 210 to the adder 408
And a motion compensation predictor that generates a motion vector 211 to the variable length encoder 403. The variable length encoder 4
03 is the first image coded data 2 from the motion vector 211, the quantization parameter 205 from the quantizer 402, and the quantization index 204 from the quantizer 402.
20 is generated.

In the encoding processing device 40 configured as shown in FIG. 2, for example, the international standardization system MPEG1 (Moving Picture) by a joint conference of ISO (International Organization for Standardization) and IEC (International Electrotechnical Commission).
es Expert Group 1), the digitized input image signal 200 is first accumulated in the current frame memory 404 to be encoded, and the current frame image to be encoded read from the current frame memory 404 to be encoded is read. 201 is input to the subtractor 405, and the prediction frame image 21 from the motion compensation predictor 410 is input.
The difference from 0 is calculated to generate the prediction error frame image 202. Next, the transform error processing is performed on the prediction error frame image 202 by the transformer 401 to generate the transform coefficient 203. Next, the quantizer 402 monitors the generated code amount for the transform coefficient 203,
Scalar quantization processing is performed according to a quantization step by feedback control that makes it constant, or by feedforward control by measurement of dispersion of the input image signal 200. With respect to the quantization index 204 generated by the quantizer 402 in this manner, the variable length encoder 403 next moves the quantization parameter 205 generated by the quantizer 402 and the motion generated by the motion compensation predictor 410. Along with the vector 211, variable length coding processing such as Huffman coding is performed to generate the first image coded data 22.
0 is generated and the first image coded data 220 is transmitted via a communication line, or a CD-ROM (Compac
t Disk-Read Only Memory) or a recording medium such as a video tape.

Further, the quantization index 204 generated by the quantizer 402 is inversely quantized by the inverse quantizer 406, and the obtained transform coefficient 206 is inversely transformed by the inverse transformer 4.
In 07, an inverse transform process for performing an inverse DCT operation is performed to generate a local prediction error frame image 207. Next, the local prediction error frame image 207 thus obtained is input to the adder 408, and is added to the prediction frame image 210 from the motion compensation predictor 410 to locally decode the frame image 208.
Is generated and stored in the encoded preceding frame memory 409. Next, in the motion compensation predictor 410, the current frame image 201 read from the current frame memory 404 to be encoded is compared with the coded preceding frame image 209 read from the coded preceding frame memory 409. Motion-compensated prediction frame image 210 that gives a minimum error by performing pattern matching calculation using
Is generated and is output to the subtractor 405 and the adder 408.

Here, in the MPEG1 motion-compensated prediction method, as shown in FIG. 3, there are three types of intra-frame coding mode, forward motion-compensated inter-frame predictive coding mode, and bi-directional motion-compensated inter-frame predictive coding mode. There is an encoding mode. Motion compensation prediction is not performed on I pictures (intra-frame coding mode frame images) F (0) and F (9) (the predicted frame image 210 is not generated). Since the I picture is used as a reference image for motion compensation prediction, it is necessary to improve the image quality of the decoded image. On the other hand, the amount of code is considerably large because motion compensation prediction is not performed. P picture (forward motion compensation inter-frame predictive coding mode frame image) F
(3) and F (6) are only images that are temporally previous (eg, F (0) for F (3) and F (6) for F (3)).
(3)) is used to perform motion compensation prediction. Since the P picture may be used as a reference image for motion compensation prediction,
It is necessary to improve the image quality of the decoded image to some extent. B picture (bidirectional motion compensation interframe predictive coding mode frame image) F (1), F (2), F (4), F (5), F (7), and F (8) are temporally before and after. Motion-compensated prediction is performed using the two images in. Since the B picture is not used as a reference image for motion compensation prediction, it is possible to perform coarse quantization. In this way, since the motion compensation prediction is performed from both directions, if the image quality of the preceding and following I picture and P picture is high in a sequence having a constant motion, for example, a decoded image can be obtained only with the motion vector and the code amount can be small.

The current frame memory 40 to be encoded
4 is F (0) → F when taking the coding mode shown in FIG.
(3) → F (1) → F (2) → F (6) → F (4) → F
The frame images are output in the order of (5) → F (9) → F (7) → F (8).

The encoded image data conversion device 30 uses the "encoded data size" as the first encoded image data 220 generated by the encoding processing device 40 by performing the encoding process on the input image signal 200. , "Size of encoded image", "different encoding method" or "arrangement of images" is converted to newly generate the second encoded image data 240. That is, the first image encoded data 220 from the encoding processing device 40 is received by the image encoded data analysis unit 310, and the first image encoded data 22 is received by the image encoded data analysis unit 310.
0 is analyzed to generate coded data 221 after signal processing, and in the process of generating coded data 221 after signal processing, a plurality of pieces of information 2 relating to the first image coded data 2
22 is extracted and output. Next, the image coded data synthesizing unit 320 newly generates the second image coded data 240 from the coded data 221 after the signal processing and a plurality of information 222 regarding the first image coded data. And outputs it to the decoding processing device 50. In this way, the image coded data synthesizing section 320 is used by the image coded data analysis section 310.
And the image coded data analysis unit 310.
, Which partially decodes the first encoded image data 220, is considered as a part of the image decoder, the encoded image data combining unit 320 is partially decoded by the encoded image data analyzing unit 310. Coded data after signal processing 2
It can be considered as part of an image encoder that performs 21 re-encodings.

A detailed description will be given below with reference to the drawings. In the encoding processing device 40, the first image encoded data 220 obtained by encoding the digital moving image signal which is the digitized input image signal 200 is encoded by the image encoding data conversion device 30. It is input to the image coded data analysis unit 310. The image coded data analysis unit 310 uses the first image coded data 220.
Is provided with a variable-length decoding process for the above, an inverse quantization process for the decoded quantization index, and the like. When the first image coded data 220 is input, The image encoded data 220 is analyzed. That is, the input first image coded data 220 is subjected to the first digital signal processing by using the functions of the variable length decoding processing and the inverse quantization processing, and the first digital signal processing is performed. The resulting encoded data 221 after signal processing, and a plurality of pieces of information 222 regarding the first image encoded data extracted from the first image encoded data 220 in the process of performing the first digital signal processing.
Is output to the image coded data combination unit 320.

On the other hand, the image coded data synthesizing section 320 deletes / adds / adds / converts the transform coefficient of the coded data 221 after the signal processing from the image coded data analysis section 310.
It has a function to perform coefficient correction processing such as correction, a function to perform quantization processing on the data for which coefficient correction has been performed, and a function to perform variable length coding processing on the quantization index. . Encoded data 2 after signal processing
21, a plurality of pieces of information 2 regarding the first image coded data 2
22 and the information 223 instructing the conversion, the image coded data synthesizing unit 320 performs the coefficient correction process, the quantization process, and the variable length coding process on the coded data 221 after the signal processing. The coded data combining process is performed by using such a function. That is, the coded data 221 after the signal processing is subjected to the second digital signal processing based on the plurality of information 222 regarding the first image coded data and the information 223 instructing the conversion, and the second digital signal processing is performed. The second image coded data 240, which is the result of the digital signal processing, is generated. Here, the plurality of pieces of information 222 regarding the first image encoded data extracted from the first image encoded data 220 in the process of the image encoded data analysis unit 310 performing the first digital signal processing is the first information 222. It is used when converting the image coded data 220 into a second image coded data 240 different from the image coded data 220 without decoding the image coded data 220. The second image coded data 240 thus generated is transmitted via a communication line or a CD.
-Recorded on a storage medium such as a ROM or a video tape, and sent to the decoding processing device 50.

In this way, the second image coded data 240 as transmission / recording information, which is transmitted through the communication line or recorded in the storage medium such as the CD-ROM or the video tape, is used as the decoding processing device 50. Is decoded and output as a decoded image signal 250. That is, the decoding processing device 50 is the same as the above-described encoding processing device 40.
It constitutes the reverse process of.

FIG. 4 is a block diagram showing an example of a concrete configuration for explaining the operation of the decoding processing device 50. In addition, 501 is a variable length decoder, 502 is an inverse quantizer, and 503 is an inverse transformer, and realizes a basic function of decoding the second image coded data 240 into a decoded image signal 250. The other part is added to realize the function of motion compensation prediction. In the figure, 241
Is the second image coded data 2 by the variable length decoder 501.
Quantization index obtained by the variable length decoding process of 40, 242 is also a quantization parameter, 243 is also a decoding motion vector, and 244 is an inverse quantizer 502.
Is the transform coefficient generated by dequantizing the quantization index 241 and 245 is the transform coefficient 2 generated by the inverse transformer 503.
44 is a prediction error frame image generated by performing an inverse DCT calculation on 44. 504 is this prediction error frame image 245
And a prediction frame image 24 from a motion compensation predictor described later.
8 is an adder for performing addition with 8, a current frame image output from the adder 504, and 505 is a current frame memory to be decoded in which the current frame image 246 is stored and is read out as a decoded image signal 250. Is. 506 stores the current frame image 246,
It is a decoded preceding frame memory which is read out as the preceding frame image 247 in the next cycle, and 507 indicates this preceding frame image 247 and the current frame image 2
46 is a motion compensation prediction circuit for generating a prediction frame image 248 to the adder 504 based on the decoded motion vector 243 from the variable length decoder 501.

In the decoding processing device 50 configured as shown in FIG. 4, the image coded data conversion device 30 generates the first image coded data 220 by subjecting the first image coded data 220 to first and second digital signal processing. Then, the second image coded data 240, which is transmitted via a communication line or recorded on a storage medium such as a CD-ROM or a video tape, is input to the variable length decoder 501. The variable length decoder 501 performs variable length decoding processing on the input second image coded data 240 and decodes it into the quantization index 241, and at the same time, the quantization parameter 24
2 and the decoded motion vector 243 are generated. Next, the quantization index 241 decoded by the variable length decoder 501
Then, the inverse quantizer 502 performs an inverse scalar quantization process in accordance with the decoding quantization parameter 242 from the variable length decoder 501, and obtains the inverse quantized transform coefficient 244. Next, the transform coefficient 24 inversely quantized by the inverse quantizer 502
The inverse transformer 503 performs an inverse DCT operation on 4 to generate a prediction error frame image 245. Next, the prediction error frame image 245 inversely transformed by the inverse transformer 503 is input to the adder 504 and added to the predicted frame image 248 from the motion compensation predictor 507 to generate a current frame image 246 to be decoded. . The current frame image 246 is accumulated in the current decoding target frame memory 505, read out from the current decoding target frame memory 505 as a decoded image signal 250, and output. Further, the current frame image 246 is also stored in the decoded preceding frame memory 506, which is read out as the decoded preceding frame image 247 at the time of motion compensation prediction of the next frame and sent to the motion compensation predictor 507. . In the motion compensation predictor 507, the decoded preceding frame image 247 and the adder 50 are added.
A predicted frame image 248 to the adder 504 is generated from the current frame image 246 output from No. 4 according to the decoded motion vector 243 from the variable length decoder 501.

In the above description, the encoding processing device 40 is used.
The case where the motion compensation prediction process is performed in the decoding processing device 50 has been described, but the motion compensation prediction process may be omitted.

As described above, according to the first embodiment, when the second digital signal processing for generating the second image coded data 240 is performed, the first image coded data 220 is firstly processed. Since the plurality of signals 222 relating to the first image coded data extracted in the process of performing the digital signal processing are used, it is unnecessary to attach special information for the second digital signal processing, which is wasteful. There is an effect that it becomes efficient in the amount of information that does not require such information. Further, compared with the case where the first image coded data is always decoded into the decoded image and then the second image coded data is re-encoded regardless of the decoded image, the deterioration of the image quality after conversion is caused. It is possible to obtain the image coded data conversion device that can be reduced in number and the processing delay can be shortened and the device scale can be reduced. Moreover, a plurality of pieces of information regarding the first image coded data are extracted in the process of generating the coded data after the signal processing from the first image coded data, and the plurality of pieces of information regarding the first image coded data are extracted based on the extracted information. Second digital signal processing
Image encoded data is generated, it is not necessary to attach special information to the second digital signal processing for generating the second encoded data, and wasteful information is not required. There is an effect that a good image coded data conversion device can be obtained.

Embodiment 2. In the first embodiment, a plurality of signals 222 relating to the first image coded data extracted in the process of performing the first digital signal processing on the first image coded data 220 are used to perform the signal processing after the signal processing. The encoded data 221 is subjected to the second digital signal processing to obtain the second
The coded image data used by the coding processing apparatus 40 to code the input image signal 200 to generate the first digital coded signal 220 has been described. Information that cannot be extracted in the process of performing the first digital signal processing on the first image coded data 220 in the data conversion device 30 is combined with the first image coded data 220 and sent to the image coded data conversion device 30. The image coded data conversion device 30 separates the image coded data 220 from the first image coded data 220 to obtain a plurality of pieces of information about the first image coded data.
It may be used when performing the second digital signal processing on the first image coded data 220.

FIG. 5 shows such a second embodiment of the present invention.
1 is a block diagram showing an image encoded data conversion device 30 according to the present invention together with an encoding processing device 40 and a decoding processing device 50.
The same reference numerals are given and their explanations are omitted. In the figure,
Reference numeral 404 is provided in the encoding processing device 40, and when the encoding processing device 40 performs the encoding process on the input image signal 200, the image encoded data conversion used by the converter 401 and the quantizer 402. Information that cannot be extracted in the process of performing the first digital signal processing on the first image encoded data 220 in the device 30 is combined with the first image encoded data 220 output from the variable length encoder 403. And 260 is the combined data output from the combining unit 404. Reference numeral 340 is provided in the image coded data conversion device 30, and the first image coded data 2 is obtained from the combined data 260 sent from the coding processing device 40.
20 is a separation unit that separates 20 from information that cannot be extracted in the first digital signal processing process, and reference numeral 224 relates to the first image coded data based on information that cannot be extracted in the first digital signal processing process. It is a plurality of information. The first image coded data 220 separated by the separation unit 340 is the image coded data analysis unit 31.
0, a plurality of pieces of information 2 regarding the first image coded data 2
24 are input to the image encoding / synthesizing unit data 320, respectively.

Next, the operation will be described. Here, the encoding processing device 40 basically operates in the same manner as that in the first embodiment, and the digitized input image signal 20
0 is encoded to obtain the first image encoded data 220.
To get However, in the encoding process, the converter 401
Information that cannot be extracted in the process of performing the first digital signal processing on the first image coded data 220 in the image coded data conversion device 30 used in the quantizer 402 is sent to the synthesizing unit 404 and the synthesizing unit 404. In (1), it is combined with the first image encoded data 220 output from the variable length encoder 403 to generate combined data 260,
The combined data 260 is converted into the image coded data conversion device 30.
It is different from the case of the first embodiment. Information that cannot be extracted in the process of performing the first digital signal processing on the first image coded data 220 in the image coded data conversion device 30 output from the converter 401 and the quantizer 402 is applied. It depends on the encoding system used, such as MPEG.
When 1 is assumed, the converter 401 outputs the parameter used when the type of the converter is determined, and the quantizer 402 outputs the parameter used when the quantization characteristic is determined. It

When the image coded data conversion apparatus 30 receives the combined data 260 sent from the coded processing apparatus 40 in this way, the separation section 340 converts it into the first image coded data 220 and the first image coded data 220. A plurality of pieces of information 224 relating to the first image coded data are separated by information that cannot be extracted in the process of one digital signal processing. Then, the separated first image coded data 220
Is input to the image coded data analysis unit 310 from the separation unit 340, and the image coded data analysis unit 310 performs the first digital signal processing on the first image coded data 220 to generate the first image data. The encoded data 221 after the signal processing, which is the result of the digital signal processing, is output to the image encoded data synthesizing unit 320. Therefore, also in this case,
The image coded data analysis unit 310 can be considered as a part of the image decoder. On the other hand, the plurality of pieces of information 224 regarding the first image encoded data separated by the separating unit 340 are input to the image encoded data combining unit 320. The image coded data synthesizing unit 320 includes an image coded data analysis unit 31.
When the coded data 221 after the signal processing and a plurality of pieces of information 224 relating to the first image coded data are input, the data amount to be converted and the image size to be converted are specified. Information 223 instructing conversion
Then, the second digital signal processing is performed in accordance with the above to perform the synthesis processing of the encoded data. Therefore, also in this case, the image coded data synthesizing unit 320 can be considered as a part of the image coder. It should be noted that this coded data combination processing in the image coded data combination unit 320 is performed by the separation unit 340 instead of the plurality of pieces of information 222 regarding the first image coded data extracted by the image coded data analysis unit 310. A plurality of pieces of information 224 regarding the first image encoded data separated from the combined data 260 are used, and the same operation as in the case of the first embodiment is performed. Here, the plurality of pieces of information 224 regarding the first image encoded data is converted into the second image encoded data 240 which is different from the first image encoded data 220 without decoding the first image encoded data 220 into an image. This information is used when the second encoded image data 240 is generated, but is information that can be used to improve the image quality over the plurality of pieces of information 222 regarding the first encoded image data. The second image coded data 240 generated as a result of performing the second digital signal processing is transmitted via a communication line or recorded on a storage medium such as a CD-ROM or a video tape. The second image coded data 240 is input to the decoding processing device 50,
The decoded image signal 250 is decoded in exactly the same manner as in the first embodiment.

As described above, in the second embodiment, because of the second digital signal processing in the image coded data synthesizing section 320, the first information based on the information that cannot be extracted in the process of the first digital signal processing is used. Since the plurality of pieces of information 224 regarding the image coded data are used, there is an effect that efficient conversion may be performed as compared with the case where the information 224 is not used. In addition, the separation unit combines the first image coded data and sends it, which cannot be extracted from the first image coded data that was used when the first image coded data was generated. Since a plurality of pieces of information regarding one piece of image coded data are separated, and the coded data after signal processing is subjected to the second digital signal processing based on the information, the second image coded data is generated. In the second digital signal processing for generating the encoded data, it becomes possible to use the information that cannot be extracted in the process of the first digital signal processing, and the conversion is performed more efficiently than when the information is not used. There is an effect that an image coded data conversion device that can possibly be obtained can be obtained.

Embodiment 3. In each of the above-described embodiments, the image coded data analysis unit 310 is part of the image decoder, and the image coded data synthesis unit 320 is part of the image encoder. A plurality of pieces of information 222 regarding the first image coded data extracted in the process of performing the digital signal processing, or the first digital signal combined with the first image coded data 220 in the coding processing device 40. The second digital signal processing is performed on the coded data 221 after the signal processing by using the plurality of pieces of information 224 regarding the first image coded data based on the information that cannot be extracted in the process of processing, and the second image coding is performed. Although the one that generates the data 240 has been described, the image coded data analysis unit 310 is the image decoder itself, and the image coded data synthesis unit 320 is the image encoder. The encoded data 221 after the signal processing is obtained from the encoded data 221 after the signal processing obtained by performing the first digital signal processing on the first image encoded data 220 and decoding the reconstructed encoded data 221. Information necessary for encoding is extracted or estimated, and is used as a plurality of pieces of information regarding the first image encoded data when the second image signal processing is performed on the first image encoded data 220. You may do it.

FIG. 6 shows such a third embodiment of the present invention.
1 is a block diagram showing an image encoded data conversion device 30 according to the present invention together with an encoding processing device 40 and a decoding processing device 50.
The same reference numerals are given and their explanations are omitted. In the figure,
Reference numeral 350 denotes the coded data 221 after the signal processing, which is decoded by the image coded data analysis unit 310 by performing the first digital signal processing on the first image coded data 220. The image coded data synthesizing unit 3 extracts or estimates information necessary for re-encoding by performing second digital signal processing on the image data as a plurality of pieces of information regarding the first image coded data.
Reference numeral 225 denotes a plurality of pieces of information about the first image coded data output from the information extraction / estimation section 350.

Next, the operation will be described. The process of the encoding processing device 40 performing the encoding process on the input image signal 200 to generate the first image encoded data 220, and the decoding processing device 50 decoding the second image encoded data 240. The process up to the generation of the decoded image signal 250 is the same as in the case of the first embodiment, and therefore its explanation is omitted here, and the operation of the image coded data conversion device 30 will be explained here.

When the image coded data conversion device 30 receives the first image coded data 220 from the coding processing device 40, the input first image coded data 220 is input in the image coded data analysis unit 310. Is subjected to first digital signal processing and is decoded, and the decoded image data is coded data 221 after signal processing as the information extraction estimation section 350 and the image coded data synthesizing section 3
Output to 20. In the information extraction estimation unit 350, the image coded data synthesis unit 32 converts the signal-processed coded data 221 input from the image coded data analysis unit 310.
A plurality of pieces of information 225 relating to the first image coded data necessary for the second digital signal processing at 0 are extracted or estimated and output to the image coded data synthesizing section 320. The image coded data synthesizing unit 320 has the first
A plurality of pieces of information 225 regarding the image coded data and the coded data 221 after the signal processing from the image coded data analysis unit 310 are instructed, together with the conversion that defines the amount of data to be converted, the image size to be converted, and the like. Information 22
3 is input, and the image coded data synthesizing unit 320, based on these pieces of information, coded data 2 after signal processing.
21 is subjected to the second digital signal processing and the image coding processing is performed to generate the second image coded data 240.
Here, the plurality of pieces of information 225 regarding the first image coded data extracted or estimated from the image coded data 221 after the signal processing by the information extraction estimation unit 350 is the first image coded data 220 decoded into an image. Then, it is used again when constructing another second image coded data 240 different from that. The second image coded data 240 thus generated is output from the image coded data conversion device 30 to the decoding processing device 50.

As described above, according to the third embodiment,
The first image encoding necessary for the second digital signal processing in the image encoded data synthesizing unit 320 is obtained from the encoded data 221 after the signal processing which is once decoded from the input first image encoded data 220. Since a plurality of pieces of information 225 regarding the data 220 are extracted or estimated, it is not necessary to perform special processing at the time of decoding, so that the configuration of the image coded data conversion device 30 can be simplified. Further, the information extraction / estimation unit extracts a plurality of pieces of data related to the first image coded data necessary for the second digital signal processing from the image data as the coded data after the signal processing decoded from the first image coded data. Individual pieces of information are extracted or estimated, and the coded data after the signal processing is subjected to the second digital signal processing based on the extracted information to generate the second image coded data. Therefore, special processing is required when performing decoding. Since there is no need to perform this, there is an effect that an image coded data conversion device that can simplify the device configuration can be obtained. Further, since the second image coded data image having a data amount different from the data amount of the first image coded data input to the image coded data analysis unit is generated by the coded data synthesizing unit, conversion is performed. There is an effect that an image coded data conversion device that can perform conversion into image coded data having different data amounts with less deterioration of image quality and processing delay afterwards can be obtained. Furthermore, when generating the second image coded data having a data amount different from that of the first image coded data, a coding process for generating the first image coded data, and the second image Since the decoding process for decoding the coded data is the coding process based on the transform coding including the transform process and the quantization process, the conversion of the data amount is made efficient by specializing the transform and the quantization. There is an effect that can be done.

Fourth Embodiment FIG. 7 shows an image coded data conversion device 30 according to Embodiment 4 of the present invention which reduces the data amount between the first image coded data 220 and the second image coded data 240. 3 is a block diagram showing the internal configurations of a data analysis unit 310 and an image coded data combination unit 320. FIG. In the figure, 311
Is a variable length decoder for performing a variable length decoding process of the first image encoded data 220 input from the encoding processing device 40 (not shown), and 226 is a result of the decoding process by the variable length decoder 311. This is the output quantization index. Reference numeral 312 denotes an inverse quantizer that performs inverse quantization processing on this quantization index 226, and 227 is a transform coefficient to which the result of the inverse quantization processing by this inverse quantizer 312 is output. The transformed coefficient 227 is output as the coded data 221 after the signal processing. The image coded data analysis unit 310 according to the fourth embodiment includes the variable length decoder 311 and the inverse quantizer 312.

Further, reference numeral 321 denotes coded data 2 after signal processing by the image coded data analysis unit 310 based on the information 223 instructing conversion which defines the amount of data to be converted.
Dequantized transform coefficient 227 sent as 21
Is a coefficient deletion / addition / correction unit that performs a correction such as deletion, addition, or correction of a part thereof. Reference numeral 228 indicates the result of the correction process by the coefficient deletion / addition / correction unit 321. Is a conversion coefficient of. Reference numeral 322 is a quantizing unit that performs a quantizing process on the transform coefficient 228 after the correction process, and 229 is a quantizing index to which the result of the quantizing process by the quantizing unit 322 is output. Reference numeral 323 denotes the second image coded data 24 obtained by performing the coding process on the quantized index 229.
It is a variable length encoder that outputs 0 to the decoding processing unit 50 (not shown). The image coded data synthesizing unit 320 according to the fourth embodiment includes the coefficient deleting / adding / correcting unit 32.
1, a quantizer 322 and a variable length encoder 323 are included.

Next, the operation will be described. Here, an encoding processing device 40 (not shown) that encodes the input image signal 200 to generate the first image encoded data 220 and inputs it to the image encoded data conversion device 30,
The second image coded data 240 output from the image coded data conversion device 30 is decoded to obtain the decoded image signal 25.
A decoding processing device 50 (not shown) that generates 0 adopts a coding method based on transform coding including transform processing and quantization processing, and the image coded data analysis unit 310 uses part of the decoding processing. , The coded image data analysis unit 310
In step 1, the first image coded data 220 is inversely quantized to extract a transform coefficient or a quantized index, and the inverse quantized transform coefficient or quantized index is partially deleted, and data to be transformed is obtained. The remaining transform coefficient or quantization index is corrected according to the amount ratio. As a result, compared to simply deleting a part of the transform coefficient or the quantization index, it becomes possible to transform the data amount in consideration of the image quality after the inverse transform.

A detailed description will be given below with reference to the drawings. The variable length decoder 311 uses the first image coded data 22
When 0 is input, the variable length decoder 311 refers to a table provided in advance to execute the variable length decoding process, and the quantization index 260 corresponding to the input first image coded data 220 is obtained. Output. An example of this table is shown below.

[0042]

[Table 1]

In the above table, for 6 data,
Codes are assigned according to their occurrence probabilities. Data with a higher probability of occurrence is assigned a shorter code, and data with a lower probability of occurrence is assigned a longer code. As a result, coded data having different lengths are associated with each other in accordance with the probability of occurrence of the quantization index, and the variable length decoding process by the variable length decoder 311 corresponds from the code (coded data) in such a table. It is done by extracting the data.

The quantization index 226 output from the variable length decoder 311 in this way is the inverse quantizer 312.
Entered in. The inverse quantizer 312 performs the inverse quantization process on the input quantization index 226 according to the quantization parameter (not shown) output from the variable length decoder 311 to obtain the inverse quantized transform coefficient 227. It is generated and output as the coded data 221 after the signal processing to the image coded data synthesizing unit 320. It should be noted that quantization in this digital processing can be generally realized by division processing, whereas inverse quantization can be realized by multiplication processing.

Information 223 instructing conversion is input to the coefficient deletion / addition / correction unit 321 of the image coded data synthesizing unit 320, and the amount of data to be converted transmitted by the information 223 instructing conversion is input. Based on the coded data analysis unit 310, coded data 2 after signal processing
A part of the transform coefficient 227 inversely quantized by the inverse quantizer 312 sent as 21 is deleted, and according to the ratio of the amount of data to be transformed defined by the information 223 instructing the transformation. Then, the remaining conversion coefficients are corrected.

In the fourth embodiment, the data amount is converted into a small amount by the image coded data conversion device 30, but when the data amount is reduced on the conversion coefficient, the low frequency component is lower on the conversion coefficient. Considering the quality of the decoded image, it has a greater meaning than the high frequency component,
It is advantageous to remove the transform coefficients from the high frequency components. Here, simply deleting the data at the position with the transform coefficient is effective for the purpose of reducing the amount of data, but affects the quality of the image decoded from the reduced transform coefficient. . In order to prevent this, in the fourth embodiment, part of the transform coefficient 227 is deleted, and
The other part of the conversion coefficient 227 is corrected according to the ratio of the amount of data to be converted specified by the information 223 instructing the conversion. At that time, the deleted conversion coefficient 227
Predict the effect of a part of
The rest of the conversion coefficient 227 is corrected so as to minimize the influence.

In this way, the coefficient deletion / addition / correction unit 321 deletes a part of the input conversion coefficient 227,
Conversion coefficient 22 after correction processing generated by correcting other parts
8 is input to the quantizer 322. The quantizer 322 quantizes the modified transform coefficient 228 to generate a quantized index 229, which is input to the variable-length encoder 323. The variable-length encoder 323 performs the variable-length encoding process on the input quantization index 229 in accordance with the quantization parameter (not shown) output from the quantizer 322, so that the second image encoded data 240 Is generated and output to the decoding processing device 50.

As described above, according to the fourth embodiment,
Since a part of the transform coefficient 227 sent as the coded data 221 after the signal processing from the image coded data analysis unit 40 is deleted, the data amount can be reduced, and a part is deleted. Since the remaining part of the conversion coefficient 227 is corrected according to the ratio of the data amount to be converted, there is an effect that the quality of the decoded image can be improved as compared with the case where the conversion coefficient is simply thinned.

Embodiment 5. In the fourth embodiment, when the data amount is reduced between the first image coded data 220 and the second image coded data 240, a part of the dequantized transform coefficient or the quantization index is used. While the case where the conversion coefficient is corrected according to the ratio of the amount of data to be converted has been described, the conversion coefficient to be deleted or the conversion coefficient in the vicinity of the quantization index is extracted from the extracted conversion coefficient or the quantization index. A part of the transform coefficient or the quantization index may be deleted by weighting with the quantization index.

In the case of such an image coded data conversion apparatus according to the fifth embodiment, the image coded data analysis unit 310 shown in FIG. 7 has the first image coded data 22.
The quantization index 226 obtained by decoding 0 by the variable length decoder 311 is further dequantized by the dequantizer 312, and the obtained transform coefficient 227 is image coded data 221 as coded data 221 after signal processing. Send to the combining unit 320.
In the image coded data synthesizing unit 320, the received transform coefficient 227 is input to the coefficient deleting / adding / correcting unit 321,
When a part of the conversion coefficient 227 is deleted, a certain weight is applied to a conversion coefficient in the vicinity of the conversion coefficient to be deleted, and then a part of the conversion coefficient 227 received based on the weight is deleted. That is, the coefficient deletion / addition / correction unit 321 checks the relationship between the conversion coefficient to be deleted and the conversion coefficients in the vicinity thereof before deleting a part of the conversion coefficient 227, and determines the image quality of the decoded image. The conversion coefficient to be deleted is deleted after the conversion coefficient is changed and weighted so that the influence is minimized.

As described above, according to the fifth embodiment,
Similar to the case of the fourth embodiment, there is an effect that the quality of the decoded image can be improved as compared with the case where the transform coefficients are simply thinned. Further, since the coefficient deleting / adding / correcting unit adds a new transform coefficient or quantization index corrected according to the ratio of the amount of data to be converted to the extracted transform coefficient or quantization index,
Since the new transform coefficient or quantization index to be added is corrected according to the ratio of the amount of data to be transformed, the image quality after inverse transform is considered compared to the case of simply adding the transform coefficient or quantization index. The amount of data can be added, which has the effect of enabling higher quality conversion.

Sixth Embodiment In the above-described Embodiments 4 and 5, the first image coded data 220
The case where the data amount between the second image coded data 240 and the second image encoded data 240 is reduced has been described, but the data amount can be increased conversely. That is, when a new transform coefficient or quantization index is added to the inversely quantized transform coefficient or quantization index to increase the amount of data, the additional transform coefficient is corrected according to the ratio of the amount of data to be transformed. .

In the case of the image coded data conversion apparatus according to the sixth embodiment, the image coded data analysis unit 310 shown in FIG. 7 has the first image coded data 22.
The variable length decoder 311 decodes 0, the obtained quantization index 226 is inversely quantized by the inverse quantizer 312, and the obtained transform coefficient 227 is sent to the image coded data combining unit 320. In the image coded data synthesis unit 320, the coefficient deletion / addition / correction unit 321 uses the received transform coefficient 227 and the image type obtained in the image coded data decoding procedure to convert the second image code after conversion. Data 240
When a new conversion coefficient is added in order to increase the amount of data, the influence on the image quality of the decoded image is minimized according to the ratio of the amount of data to be converted defined by the information 223 for instructing conversion. Then, the conversion coefficient newly added is corrected.

Here, when the data amount between the first image coded data 220 and the second image coded data 240 is to be increased, the decoded image of the decoded image is to be generated in the same manner as when the data amount is decreased. Considering the influence on the image quality, it is advantageous to add the transform coefficient or the quantization index for the high frequency component. Further, in general, the addition of the transform coefficient or the quantization index in the high frequency region causes a larger change in the data amount than the change in the transform coefficient or the quantization index in the low frequency region.

Thus, according to the sixth embodiment,
Since it is corrected according to the ratio of the amount of data to be converted for the new conversion coefficient or quantization index to be added,
Compared with the case of simply adding the transform coefficient or the quantization index, it is possible to add the data amount in consideration of the image quality after the inverse transform, and it is possible to perform the conversion with higher image quality. The coefficient deleting / adding / correcting unit predicts a transform coefficient or a quantization index including the transform coefficient or the quantization index of the addition target and its vicinity, and then extracts a new transform coefficient or a quantization index. Since it is added to the transform coefficient or quantization index, the prediction is performed so that the image quality at the time of decoding can be improved when adding the transform coefficient or quantization index. Alternatively, as compared with the case where the quantization index is added, it is possible to make the image visually easier to see, and there is an effect that higher quality conversion can be performed.

Embodiment 7. In the sixth embodiment, in order to increase the amount of data between the first image coded data 220 and the second image coded data 240, a new dequantized transform coefficient or quantization index is added. When adding a transform coefficient or quantization index, we explained the case of correcting the added transform coefficient or quantization index according to the ratio of the amount of data to be converted, but the transform coefficient or quantization index to be added and its neighborhood The conversion coefficient or the quantization index prediction including the conversion coefficient or the quantization index may be performed, and the conversion coefficient or the quantization index may be added.

In the case of such an image coded data conversion apparatus according to the seventh embodiment, the image coded data analysis unit 310 shown in FIG. 7 has the first image coded data 22.
The variable length decoder 311 decodes 0, the obtained quantization index 226 is inversely quantized by the inverse quantizer 312, and the obtained transform coefficient 227 is sent to the image coded data combining unit 320. In the image coded data synthesizing section 320, when the coefficient deleting / adding / correcting section 321 receives the received coefficient and the received transform coefficient 227 is used to increase the data amount of the converted second image coded data 240. The conversion coefficient is added by predicting the conversion coefficient including the newly added conversion coefficient and the conversion coefficient in the vicinity of the conversion coefficient to be added.

As described above, according to the seventh embodiment,
Since the transform coefficient or quantization index is added by predicting the transform coefficient or quantization index of the addition target and its neighborhood, it is easier to see the image visually than when simply adding the transform coefficient or quantization index. Therefore, there is an effect that higher quality conversion can be performed.

Embodiment 8. FIG. 8 shows an image coded data conversion device 30 according to the eighth embodiment of the present invention which reduces the data amount between the first image coded data 220 and the second image coded data 240. 3 is a block diagram showing the internal configurations of a data analysis unit 310 and an image coded data combination unit 320. FIG. In the figure, 230
Is the first input from the encoding processing device 40 (not shown).
The result of the decoding process is output from the variable length decoder 311 that performs the variable length decoding process of the image encoded data 220 of
This is one piece of encoded picture information as an encoding parameter indicating the image type, and the other portions are denoted by the same reference numerals as those of the corresponding portions in FIG. 7, and description thereof will be omitted. The variable length decoder 311 outputs this coded picture information 230 together with the quantization index 226, and the coefficient deletion / addition / correction unit 321 converts the amount of data to be converted based on the information 223 instructing conversion. At the same time, the coded picture information 230 is also used to convert the transform coefficient 227.
Are different from those in the fourth embodiment in that some of them are deleted, added, or corrected such as correction. The coded picture information 230 is a parameter that is multiplexed on a picture-by-picture basis, and includes the size of the picture and the coding type of the picture (intra-frame coding, unidirectional motion compensation inter-frame prediction coding in MPEG1). And bidirectional motion-compensated interframe predictive coding), the maximum value of the vector used in the picture, and the adaptive parameter used in the picture.

Next, the operation will be described. In the image coded data conversion apparatus according to the eighth embodiment, in the image coded data analysis unit 310, the transform coefficient or the quantization index obtained by dequantizing the first image coded data 220 and the first Of the transform coefficient or the quantization index included in the image type not used for prediction in the future encoding by using the image type obtained during the decoding of the image encoded data 220 As a result, it is possible to perform conversion while suppressing deterioration in image quality as a whole.

The image coded data analysis unit 31 shown in FIG.
At 0, the first image coded data 220 is decoded by the variable length decoder 311, and the quantization index 226 of the decoding result is obtained.
To the dequantizer 312, and the coded picture information 230 obtained in the decoding process is used as the image coded data synthesizing unit 32.
Send to 0. The inverse quantizer 312 inversely quantizes the received quantization index 226, and obtains the obtained transform coefficient 227.
Is sent to the image coded data synthesizing section 320 as coded data 221 after signal processing. On the other hand, the image coded data synthesizing unit 320 deletes the inversely quantized transform coefficient 227 and the coded picture information 230 together with the information 223 for instructing the conversion that defines the amount of data to be converted / coefficient deleted.
The conversion coefficient 22 received by the addition / correction unit 321
Delete part of 7. At this time, when a coded picture that is not used for prediction in future coding is shown, the conversion coefficient deletion ratio is increased. Here, when the received coded picture information 230 includes the picture coding type as described above, and the coding type indicates bidirectional motion compensation inter-frame predictive coding in MPEG1, for example. , The picture information is not used for subsequent encoding, so even if the image quality deteriorates due to the reduction of the data amount,
There is no fear that the deterioration will continue over time. Therefore, in such picture information, the efficiency of reducing the data amount can be improved by increasing the deletion ratio of the transform coefficient.

Thus, according to the eighth embodiment,
If the coded frame is not used in the next time unit coding, by increasing the reduction in the amount of data,
As a result, it is possible to effectively reduce the total amount of data.

Ninth Embodiment In the above-described Embodiment 8, the case where the reduction rate of the transform coefficient is increased when the coded picture that is not used in future prediction is shown has been described, but the coded picture used in future prediction is shown. In this case, the reduction rate of the conversion coefficient may be lowered. That is, in the image coded data analysis unit 310 shown in FIG. 8, the transform coefficient or the quantization index obtained by dequantizing the first image coded data 220 and the first image coded data 220. When the image type is used for prediction in future encoding, the reduction ratio of the transform coefficient or the quantization index is reduced by using the image type obtained in the middle of decoding. As a result, it is possible to perform a conversion that suppresses deterioration from continuing in the future by prediction.

As described above, according to the ninth embodiment,
When the coded frame is used in the next time-based coding, the reduction in the amount of data is reduced, so that there is an effect that the data can be efficiently reduced as a whole.

Embodiment 10. FIG. 9 shows an image coded data conversion device 30 according to the tenth embodiment of the present invention which reduces the data amount between the first image coded data 220 and the second image coded data 240. 3 is a block diagram showing the internal configurations of a data analysis unit 310 and an image coded data combination unit 320. FIG. In the figure, 2
Reference numeral 31 is an image block prediction in which the result of the decoding process is output from the variable length decoder 311 that performs the variable length decoding process of the first image encoded data 220 input from the encoding processing device 40 (not shown). This is coding block information as a coding parameter indicating the type, and the other parts are denoted by the same reference numerals as the corresponding parts in FIG. 8 and their description is omitted. The variable length decoder 311 outputs the coding block information 231 together with the quantization index 226 and the coded picture information 230, and the coefficient deletion / addition / correction unit 321 uses the information 2 for instructing conversion.
This coding block information 23 together with the amount of data to be converted defined by 23 and the coding picture information 230
1 is also used, and a part of the conversion coefficient 227 is deleted, added or corrected such as correction, which is different from those in the fourth embodiment. The coded block information 231 is information about a coded block (which may be called a micro block in MPEG1 or the like), which is the minimum unit of coding, and includes a quantization parameter, a motion vector, and a motion used in the coded block. It indicates the presence or absence of compensation prediction.

Next, the operation will be described. In the image coded data conversion device according to the tenth embodiment, in the image coded data analysis unit 310, the transform coefficient or the quantization index obtained by dequantizing the first image coded data 220 and the first Image coded data 220
By using the image type and the prediction type of the image block obtained during the decoding of, even if the image type used for prediction in future encoding, if the image block is not used in future prediction, By increasing the ratio of deleting the conversion coefficient or the quantization index included in the image block, it is possible to prevent the deterioration from continuing in the future and perform the conversion in which the deterioration of the total image quality is suppressed.

The image coded data analysis unit 31 shown in FIG.
At 0, the first image coded data 220 is decoded by the variable length decoder 311, and the quantization index 226 of the decoding result is obtained.
To the inverse quantizer 312, and the coded picture information 230 and the coded block information 23 obtained in the decoding process are transmitted.
1 is sent to the image coded data combination unit 320. The dequantizer 312 dequantizes the received quantization index 226, and sends the obtained transform coefficient 227 to the image coded data synthesizing unit 320 as the coded data 221 after signal processing. On the other hand, in the image coded data synthesizing section 320, the inversely quantized transform coefficient 227 and coded picture information 23
0 and the encoded block information 231 are input to the coefficient deletion / addition / correction unit 321 together with the information 223 instructing the conversion that defines the data amount to be converted, and a part of the received conversion coefficient 227 is deleted. At that time, based on the received encoded block information 231, whether or not to use the information of the image block in future encoding is determined not in encoded picture units but in image block units that allow finer control. Even if it is a coded picture that will be used for prediction in future encoding, if the image block indicates a block that is not used for prediction in future encoding, the conversion coefficient deletion ratio is set. Make it higher

As described above, according to the tenth embodiment, in addition to the coded picture information 230, the coded block information 231 is also used, and whether or not the ratio of deleting the transform coefficient or the quantization index is increased or not. Since it is determined, it is possible to perform control in finer units, and it is possible to perform conversion with higher quality. Further, the image coded data synthesis section may generate second image coded data having a decoding processing procedure different from the decoding processing procedure of the first image coded data input to the image coded data analysis section. Since it is configured as described above, when generating image coded data of different coding methods, it is possible to perform conversion with high image quality after conversion.

Eleventh Embodiment FIG. 10 shows an image coded data conversion device 30 according to the eleventh embodiment of the present invention, which converts the first image coded data 220 and the second image coded data 240 into a system in which the decoding processing procedure is different. 3 is a block diagram showing the internal configurations of a coded data analysis unit 310 and an image coded data combination unit 320. FIG. In the figure, 311 is an encoding processing device 40 not shown.
This is a variable length decoder that performs variable length decoding processing of the first image coded data 220 that is input further. 226 is a quantization index as an encoding parameter for outputting the result of the decoding process by the variable length decoder 311, 230 is the same encoded picture information, 231 is the same encoded block information, 232 is the same quantization parameter information, 233
Is also motion vector information, and these pieces of information are the image coding / synthesizing unit 3 as coded data 221 after signal processing.
It is output to 20. The image coded data analysis unit 310 according to the eleventh embodiment includes this variable length decoder 311. The quantization parameter information 232 and the motion vector information 233 are coded block information 231.
It is positioned as one of the.

Reference numeral 324 is the image coded data analysis unit 3
10 regarding the coding parameters such as the quantization index 226, the coded picture information 230, the coded block information 231, the quantization parameter information 232, and the motion vector information 233 sent as the coded data 221 after the signal processing. Is a coding parameter correction / conversion unit that transforms to match the coding system after conversion. Reference numeral 323 denotes the encoding parameter correction / conversion unit 32.
The second conversion result obtained by performing the encoding process on the converted output of No. 4
It is a variable length encoder that outputs the image encoded data 240 of 1. to the decoding processing unit 50 (not shown). The image coded data synthesizing unit 320 according to the eleventh embodiment includes the coding parameter correcting / converting unit 324 and the variable length coder 32.
Includes 3 and 3.

Next, the operation will be described. Here, the decoding processing procedure of the first image coded data 220 input to the image coded data analysis unit 310 and the second image coded data 24 output from the image coded data synthesis unit 320.
When an attempt is made to configure the decoding processing procedure of 0 to be different, the following processing is performed. Now, for example, if the input first image coded data 220 is MPE
The second image coded data 240 to be output is H.G.G1 which is a coding method for videophone / videoconference. Consider the case of H.261. Image coded data analysis unit 31
At 0, the MP input by the variable length encoder 311
From the first image coded data 220 by EG1, the quantization index 226, the coded picture information 230, the coded block information 231, and the quantization parameter information 232.
And the coding parameters such as the motion vector information 233 are extracted and output to the image coded data synthesizing section 320 as the coded data 221 after the signal processing. In the image coded data synthesizing unit 320, the received coded data 221 after the signal processing is input to the coding parameter correction / conversion unit 324, and the respective coding parameters are converted from the MPEG1 representation format into the H.264 format. It is converted into the expression format in H.261.

Specifically, the MPEG1 image coded data is converted into H.264. When converting the motion vector information 233, for example, when converting to the image coded data of H.261, in MPEG1, there is a motion vector of a half pixel unit, whereas in H.264. In H.261, since only a motion vector with integer pixel accuracy can be obtained, the coding parameter correction / conversion unit 324 extracts only the integer part of the motion vector considered to be the most appropriate in MPEG1, and the H.264 code is used. 261
Conversion processing such as using as a motion vector of the. Similarly, for other coding parameters such as the quantization parameter information 232 and the quantization index 226, MPEG1 and H.264 are used. H.261, unless used exactly the same as H.261. In step 261, the parameters are converted into the most appropriate parameter values. In this way, the H.264 coding scheme after coding by the coding parameter correction / transformation unit 310 is performed. The various encoding parameters converted so as to conform to H.261 are the variable length encoder 323.
At, the second image encoded data 240 is combined and output.

As described above, according to the eleventh embodiment, the first image coded data 220 input to the image coded data analysis unit 310 is converted into the second image coded having a different decoding processing procedure. Even when the data 240 is converted into the data 240 and output from the image coded data synthesizing unit 320, the first image coded data 220 is once decoded into an image, and then the image is coded by a corresponding coding method. Since there is no need to re-encode by the device, there is an effect that the device can be downsized and the cost can be reduced. Further, the second image coded data including the image size temporally or spatially different from the image size included in the first image coded data input to the image coded data analysis unit is converted into the image coded data. The image quality after conversion is high because it is generated by the synthesizing unit.
There is an effect that an image coded data conversion device that can easily convert image coded data having different image sizes temporally or spatially can be obtained. Furthermore, when generating second image coded data having an image size that is temporally or spatially different from the first image coded data,
The coding process for generating the first image coded data and the decoding process for decoding the second image coded data are based on transform coding including motion prediction compensation process, transform process, and quantization process. Since the encoding process is performed, there is an effect that it is possible to perform motion compensation prediction using a motion vector, or to convert an image size by changing a transform coefficient or a quantization step.

Twelfth Embodiment FIG. 11 shows an image coded data conversion device 30 according to a twelfth embodiment of the present invention, which converts image sizes different between the first image coded data 220 and the second image coded data 240. Encoded data analysis unit 310 and image encoded data synthesis unit 3
FIG. 20 is a block diagram showing the internal configuration of 20. Each unit is given the same reference numeral as the corresponding portion of FIG. 7, and description thereof will be omitted. In the twelfth embodiment, the image size to be changed is defined by the conversion instruction information 223, and the coefficient deletion / addition / correction unit 321 uses the conversion image size based on the conversion instruction information 223. , The transform coefficient 227 dequantized by the dequantizer 312
Is different from that in the fourth embodiment in that it is deleted or corrected.

Next, the operation will be described. Here, for example, the image size of the image signal included in the first image encoded data 220 input to the image encoded data analysis unit 310 and the second image code output from the image encoded data combining unit 320. As shown in FIG. 2, the encoding processing device 40 that generates the first image encoded data 220 is used when conversion is performed so that the image size of the image signal included in the encoded data 240 has a different size. Also, a decoding processing device 5 for decoding the second image coded data 240
If 0 is generated by the coding process based on the transform coding including the motion compensation prediction process, the transform process, and the quantization process, each of which uses a motion vector as shown in FIG. The dequantized transform coefficient or the quantization index is extracted by the quantized data analysis unit 310, and when the extracted transform coefficient or the quantized index is increased or decreased, the increase or decrease is performed according to the ratio of the image size to be converted. The transform coefficient or the quantization index to be corrected is corrected. As a result, it is possible to suppress the extreme deterioration or unnaturalness of the sense of resolution that is likely to occur when converting the image size.

A detailed description will be given below with reference to the drawings. The image coded data analysis unit 31 shown in FIG.
At 0, the quantization index 226 obtained by decoding the first image coded data 220 by the variable length decoder 311 is further dequantized by the dequantizer 312, and the obtained transform coefficient 227 is subjected to signal processing. The encoded data 211 is sent to the image encoded data synthesizing section 320. In the image coded data synthesizing unit 320, together with the information 223 instructing conversion,
The received conversion coefficient 227 is input to the coefficient deletion / addition / correction unit 321. When the coefficient deletion / addition / correction unit 321 increases / decreases the dequantized conversion coefficient 227, the conversion coefficient increasing / decreasing is deleted according to the ratio of the image size to be converted defined by the information 223 instructing the conversion. / Addition / correction such as correction. Here, for example, when considering the case of reducing the image size, when the image size is reduced, the number of samples is reduced as compared with the image of the original size, and the included frequency component becomes low. Therefore, it is necessary to limit the band in some way, but in this case, it is performed on the transform coefficient 227. That is, the correction for converting the high frequency component into the encoded data having a small image size is performed. The transform coefficient 228 after this correction processing is processed by the coefficient deletion / addition / correction unit 321 by the quantizer 3
22 and the quantization index 229 is generated by the quantization process, and the variable length encoder 323 performs the variable length coding process on the quantization index 229 to obtain the second index. The encoded image data 240 is output to the decoding processing device 50.

As described above, according to the twelfth embodiment, when the coded data is converted when the image size is changed, the conversion coefficient to be increased / decreased is corrected according to the image size to be converted. It is possible to suppress the extreme deterioration of the sense of resolution and the unnaturalness that are likely to occur when converting the size, and it is possible to prevent the deterioration of the image quality after the image size is changed.

Thirteenth Embodiment In the twelfth embodiment, when converting the first image coded data 220 and the second image coded data 240 into different image sizes, the conversion coefficient that increases or decreases according to the ratio of the image sizes to be converted. However, the magnitude of the motion vector used for the motion compensation extracted is corrected according to the ratio of the image sizes to be converted, and the first image code is corrected. The encoded data 220 may be converted into second image encoded data 240 having an image size different from that of the encoded data 220.

FIG. 12 shows an image coded data conversion device 30 according to the thirteenth embodiment of the present invention, which includes an image coded data analysis unit 310 and an image coded data synthesis unit 32.
3 is a block diagram showing the internal configuration of 0. FIG. In the figure, reference numeral 233 denotes motion compensation output from the variable length decoder 311 that performs variable length decoding processing of the first image encoded data 220 input from the encoding processing device 40 (not shown) in the process of the decoding processing. Motion vector information for
The other parts are given the same reference numerals as the corresponding parts in FIG. 11 and their explanations are omitted. The variable length decoder 311
Indicates that the motion vector information 233 is output together with the quantization index 226, and the coefficient deletion / addition / correction unit 321 determines the motion vector together with the image size to be converted defined by the information 223 instructing conversion. This is different from those in the twelfth embodiment in that the information 233 is also used to make corrections such as increase / decrease and correction of the conversion coefficient 227.

Next, the operation will be described. In the image coded data analysis unit 310 shown in FIG. 12, the first image coded data 220 is decoded by the variable length decoder 311, and the quantization index 226 of the decoding result is sent to the dequantizer 312 and the decoding thereof is performed. The motion vector information 233 obtained in the process is sent to the image coded data combining unit 320. Inverse quantizer 31
In 2, the received quantization index 226 is inversely quantized, and the obtained transform coefficient 227 is sent to the image coded data synthesizing unit 320 as the coded data 221 after the signal processing.
On the other hand, in the image coded data synthesizing unit 320, the dequantized transform coefficient 227 and the motion vector information 233, together with the information 223 instructing the transform that defines the image size to be transformed, the coefficient deleting / adding / correcting unit 321. To enter. The coefficient deletion / addition / correction unit 321 corrects the size of the motion vector according to the ratio of the image size to be converted specified by the information 223 instructing the conversion.
In this way, if the size of the motion vector is also changed according to the ratio of the change in the image size, the vector used before the conversion can be used almost as it is. Therefore, even when re-searching with the vector calculated here as the center, the search efficiency can be further improved as compared with the case of re-searching from the beginning.

As described above, according to the thirteenth embodiment, since the magnitude of the motion vector is corrected according to the ratio of the image size to be changed, the search efficiency of the motion vector in the converted image coded data is high. There is an effect that even if a motion compensation search in a narrower range using the corrected motion vector as a reference, a characteristic equivalent to a motion compensation search range in a wider range can be obtained.

Fourteenth Embodiment FIG. 13 converts the temporal sequence of images of the first image coded data 220 and the temporal sequence of images of the second image coded data 240,
12 is a block diagram showing an internal configuration of an image coded data analysis unit 310 and an image coded data synthesis unit 320 of an image coded data conversion device 30 according to a fourteenth embodiment of the present invention, and each unit corresponds to FIG. The same reference numerals as those of the parts are given and the description thereof will be omitted. In addition, this Embodiment 14
In the above, the arrangement information of the images to be converted is defined by the information 223 instructing the conversion.

In the image coded data analysis unit 310 in the figure, reference numeral 313 denotes an inverse transform for performing an inverse DCT operation or the like on the inversely quantized transform coefficient 227 output from the inverse quantizer 312 to perform an inverse transform process. 314 is an adder that adds the processing result of the inverse converter 313 and the processing result of the motion compensation unit described later. 234 is a decoded image output from the adder 314 as the coded data 221 after signal processing, and 315 is a frame memory in which the decoded image 234 is temporarily stored. 316 is a variable length decoder 3
11 performs motion compensation processing on the decoded image one cycle before read from the frame memory 315 based on the motion vector information 233 generated in the decoding process, and outputs the processing result to the adder 314. Is.

Further, in the image coded data synthesizing section 320, 325 is an inverse quantizer which inversely quantizes the transform coefficient 229 output from the quantizer 322, and 326 is an inverse output to this inverse quantizer 325. An inverse transformer that performs inverse transformation by performing DCT calculation or the like, and an adder 327 that adds the processing result of the inverse transformer 326 and a frame image output by a motion search unit described later. Reference numeral 328 is a frame memory in which the addition output from the adder 327 is temporarily stored,
329 is 1 read from this frame memory 328
Image data before cycle and image coded data analysis unit 3
The motion vector information 233 from 10 and the decoded image 234 as the coded data 221 after the signal processing, and the signal 22 instructing the conversion that defines the arrangement information of the images to be converted.
3 is input, the size of the motion vector extracted according to the arrangement information of the images to be converted is estimated, and the motion search unit performs the motion search based on the result, and 235 is the motion search unit. 329 is a frame image output from the adder 327 to the adder 327 and a subtracter described later. Reference numeral 330 denotes a subtracter that calculates the difference between the decoded image 234 sent from the image coded data analysis unit 310 as the coded data 221 after the signal processing and the image data 235 output from the motion search unit 329. Reference numeral 331 is a converter that performs a conversion process such as a DCT operation on the output of the subtractor 330 and inputs the processed result to the quantizer 322.

Next, the operation will be described. Here, for example, the temporal arrangement of the image signals included in the first image encoded data 220 input to the image encoded data analysis unit 310 and the image encoding output from the image encoded data combining unit 320. As shown in FIG. 2, the encoding processing device 40 that generates the first encoded image data 220 when the temporal arrangement of the image signals included in the data 240 is to be converted into a different order. Further, as shown in FIG. 4, a decoding processing device 50 for decoding the second image coded data 240 includes a transform code including a motion compensation prediction process, a transform process, and a quantization process, each of which uses a motion vector. If it is generated by a coding process based on encoding, the motion vector used in the motion compensation is extracted in the image coded data analysis unit 310, and the motion search unit 3
The size of the extracted motion vector is estimated according to the arrangement information of the images to be converted in 29. As a result, by performing a search based on the estimated motion vector when converting the coding image type, even if the motion search range involving a large amount of calculation is narrowed, the motion compensation search of a wider range is performed. The same efficiency as the range can be maintained.

The image coded data analysis unit 3 shown in FIG.
10, the first image coded data 220 is decoded by the variable length decoder 311 and the decoded result quantization index 22
6 is sent to the dequantizer 312, and the motion vector information 233 obtained in the decoding process is used as the image coded data synthesizing unit 32.
It is also sent to 0 and input to the motion compensation unit 316. In the inverse quantizer 312, the quantization index 226 is inversely quantized, and the obtained transform coefficient 227 is subjected to inverse transform processing such as inverse DCT operation in the inverse transformer 313, and the adder 314 is then applied.
Entered in. On the other hand, the motion compensation unit 316 performs motion compensation processing on the decoded image read one cycle before read from the frame memory 315 according to the motion vector information 233 received from the variable length decoder 311, and the result is added by the adder 314. To enter. The adder 314 adds the frame images sent from the inverse converter 313 and the motion compensation unit 316 to generate a decoded image 234, which is used as the coded data 221 after the signal processing, and the image coded data combining unit 320. Send to. As described above, the image coded data analysis unit 310 performs the same operation as the decoding operation of the normal motion-compensated transform coding method, but with the decoded image 234 as the coded data 221 after the signal processing, and the variable length. The motion vector information 233 extracted by the decoder 311 is also output to the coded image data combining unit 320.

In the image coded data synthesizing unit 320, the decoded image 234 and the motion vector information 233 are input to the motion searching unit 329 together with the information 223 instructing the conversion in which the arrangement information of the images to be converted is specified. . In this motion search unit 329, the transform coefficient 229 output from the quantizer 322 is inversely quantized by the inverse quantizer 325, and the inverse transformer 32 is applied to the output of the inverse quantizer 325.
Inverse conversion processing such as inverse DCT calculation is performed in 6 and further, in an adder 327, the addition result of the processing result of the inverse converter 326 and the processing result of the motion search unit 329 is read from the frame memory 328 in which the addition result is stored. The image data of one cycle before is also input. The motion search unit 329 generates a frame image 235 based on these pieces of information, and realizes an operation equivalent to a normal coding operation of a transform-encoding method with motion compensation. The motion vector for the current encoded data is estimated using the arrangement information of the images to be changed defined by the above and the motion vector information 233 from the image encoded data analysis unit 310,
Motion search is performed based on the result. Motion compensation unit 3
The frame image 235 output from 29 is a subtractor 330
And the difference from the decoded image 234 that has been input to the image encoded data analysis unit 310 and transmitted as the encoded data 221 after the signal processing is calculated. The operation result of the subtractor 330 is subjected to DCT operation and conversion processing in the converter 331, and further subjected to quantization processing in the quantizer 322, and the obtained conversion coefficient is converted into the first value in the variable length encoder 323. 2 is encoded into image encoded data 240. As described above, in the image coded data synthesis unit 320, the decoded image 234 and the motion vector information 2 from the image coded data analysis unit 310 are acquired.
33 is input and the same processing as the coding operation of the normal conversion-compensated coding method with motion compensation is performed. However, in the motion search unit 329, the size of the motion vector extracted by the image coded data analysis unit is converted. Is estimated according to the information on the arrangement of the images to be changed by the information 223 instructing, and the motion search is performed based on the result.

As described above, according to the fourteenth embodiment, by performing conversion for changing the temporal arrangement of images at the time of reproduction and the temporal arrangement of images after conversion, encoding delay can be reduced. In addition, in the conversion for changing the temporal arrangement of the image, the size of the motion vector is estimated according to the arrangement information of the image to be converted, so that the encoding by the converted motion vector is performed. There is an effect that efficiency can be improved. Further, there is an effect that an image coded data conversion device that can convert image coded data including a temporal sequence of different image signals with high image quality after conversion can be obtained. Furthermore, when generating the second image coded data in which the temporal arrangement of the image signals included in the first image coded data is different from the first image coded data, a coding process for generating the first image coded data, and Since the decoding process for decoding the image coded data of No. 2 is configured to be the coding process based on the transform coding with the motion prediction / compensation process, by estimating the rearranged motion vector, the efficiency can be improved. This has the effect of enabling conversion that can perform good motion-compensated prediction. In addition, since the motion search unit is configured to estimate the size of the motion vector extracted by the image coded data analysis unit according to the arrangement information of the images to be converted, the size of the motion vector to be converted By estimating according to the arrangement information of, the coding efficiency by the motion vector after conversion is improved.

Fifteenth Embodiment FIG. 14 is an image code in Embodiment 15 of the present invention for converting the number of image signals included in the first image coded data 202 and the number of image signals included in the second image coded data 20. FIG. 9 is a block diagram showing an internal configuration of an image coded data analysis unit 310 and an image coded data synthesis unit 320 of a coded data conversion device 30. Each unit is given the same reference numeral as the corresponding portion in FIG. Is omitted. In the fifteenth embodiment, the conversion instruction information 223 defines the image frame rate information, and the coefficient deletion / addition / correction unit 321 determines the image frame based on the conversion instruction information 223. Inverse quantizer 3 using rate information
This is different from that in the eighth embodiment in that the transform coefficient 227 dequantized by 12 is deleted or corrected.

Next, the operation will be described. Here, for example, the second output from the image coded data combination unit 320
The first image coded data 2 input to the image coded data analysis unit 310 is the number of decoded image signals included per time unit when the image coded data of 1 is decoded into an image.
When the number of decoded image signals included in a unit of time when 20 is decoded into an image is to be converted into a different number, the image coding is performed as information extracted from the first image coded data 220. If a mode is used and this image coding mode indicates an image that will not be used for future image coding, then the data for that image is deleted. As a result, the frame rate can be converted without greatly affecting the quality of future images.

The image coded data analysis unit 3 shown in FIG.
10, the first image coded data 220 is decoded by the variable length decoder 311 and the decoded result quantization index 22
6 is sent to the inverse quantizer 312, and the coded picture information 230 obtained in the decoding process is sent to the image coded data synthesizing unit 320. The inverse quantizer 312 inversely quantizes the quantization index 226 to obtain the obtained transform coefficient 2
27 is sent to the image coded data synthesizing section 320 as coded data 221 after signal processing. On the other hand, in the image coded data synthesizing unit 320, the inversely quantized transform coefficient 227
And the encoded picture information 230 together with the information 223 instructing the conversion that defines the image frame rate information to the coefficient deletion / addition / correction unit 321 to delete / add / coefficient the coefficient.
The correction unit 321 makes a determination based on the frame rate information of the image defined by the transform coefficient 227, the coded picture information 230, and the information 223 instructing the conversion, and a code that is not used for prediction in future coding. In the case where a picture is shown, the same concept as in the case of the eighth embodiment is applied to delete the data of the picture and equivalently convert the frame rate.

As described above, according to the fifteenth embodiment, the number of image signals included in the converted second image coded data 240 is set to the first image coded data 2 before conversion.
Setting the value different from the number of image signals included in 20 has an effect of facilitating conversion between television signals of different systems. In addition, since the size of the motion vector extracted by the image coded data analysis unit is estimated by the motion search unit according to the arrangement information of the images to be converted, the size of the motion vector is set to the arrangement information of the images to be converted. According to the estimation, there is an effect that the coding efficiency by the converted motion vector is improved. Further, since the image coded data synthesizing unit generates the second image coded data in which the number of decoded image signals included per unit time at the time of decoding is different from the first image coded data, the converted image is generated. There is an effect that it is possible to obtain an image coded data conversion device capable of converting image coded data having a high number of different numbers of images included in the included decoded image signal.

Sixteenth Embodiment FIG. 15 shows an embodiment of the present invention used for quantization when estimating a quantization parameter from a decoded image of first encoded image data 220 and using it to generate second encoded image data 240. 16 is a block diagram showing the internal structure of an image coded data analysis unit 310 and an image coded data combination unit 320 of the image coded data conversion device 30 in FIG. The description is omitted. In the figure, reference numeral 236 is a decoded image obtained by the inverse transformer 313 by performing an inverse DCT operation or the like on the transform coefficient 227 output from the inverse quantizer 312, and is decoded as the encoded data 221 after the signal processing. It is sent to the section 320. Reference numeral 332 is a quantization estimation unit that estimates the quantization parameter when the first image decoded data 220 is generated from the decoded image 236 sent as the coded data 221 after this signal processing, and 237 is the quantization estimation. Quantization parameter information output from the unit 332 to the quantizer 322, and 238 is a transform coefficient generated from the decoded image 236 by the transformer 331.

Next, the operation will be described. When converting the image rate of the first image encoded data 220 and the image rate of the second image encoded data 240, the image encoded data analysis unit 310 extracts from the first image encoded data 220. As the information to be stored, a quantization parameter in the quantization is extracted, and the image coded data synthesizing unit 32
At 0, quantization using the quantization parameter is performed. Accordingly, when the image coded data analysis unit 310 is the decoding process and the image coded data synthesis unit 320 is the coding process and the bit rate and the coding method before and after the conversion are the same, the best quality is obtained. It becomes possible to perform conversion. Further, even when the bit rates are different, optimum conversion can be performed by controlling the quantization parameter according to the ratio.

The image coded data analysis unit 3 shown in FIG.
In 10, the first image coded data 220 is variable length decoded by the variable length decoder 311 to generate the quantization index 226, and the dequantizer 312 dequantizes it to generate the transform coefficient 227. The inverse transformer 313 has its transform coefficient 2
The decoded image 236 is generated by performing an inverse DCT operation on 27 and is output to the image coded data synthesizing unit 320 as the coded data 221 after the signal processing. In this way, the image coded data analysis unit 310 performs the same operation as the normal decoding operation, and outputs the decoded image 236 as the signal-processed coded data 221. On the other hand, in the image coded data combination unit 320, the decoded image 236 sent as the coded data 221 after the signal processing is input to the converter and D
The transform processing is performed by performing a CT operation or the like, and the generated transform coefficient 238 is input to the quantizer 322. The decoded image 236 is also input to the quantizer estimator 332, and the quantizer estimator 332 receives the first decoded image 236 from the received decoded image 236.
The quantization parameter at the time of generating the image coded data 220 is estimated, and the quantization parameter information 237 as the estimation result is sent to the quantizer 322. Hereinafter, based on the quantization parameter information 237 from the quantizer estimation unit 332, the quantizer 322 causes the transform coefficient 2 from the transformer 331 to change.
38 is subjected to a quantization process, and the variable length encoder 323 performs a coding process on the quantization index to generate and output the second image coded data 240.

As described above, according to the sixteenth embodiment, it is possible to estimate the quantization parameter from the decoded image 236 and improve the image quality of the newly generated second image coded data 240. There is an effect that the image quality after conversion can be maintained in applications such as relay transmission,
In particular, if the rates before and after the conversion are the same, it is possible to maintain the image quality even when the encoding is repeated multiple times by using the estimated quantization parameter, and the rates are different. Even in such a case, there is an effect that optimum conversion can be performed by controlling the quantization parameter according to the ratio.

Seventeenth Embodiment It goes without saying that not only the configuration of the communication image shown in each of the above-described embodiments, but also an image conversion system between multiple points and a system for copying image data in a storage medium can be configured.

In each of the above embodiments, coefficient deletion /
Addition / correction unit 321 and encoding parameter correction / conversion unit 3
24, the quantizer estimator 332 and other parts that play a central role in conversion are arranged in the image coded data synthesizer 320, but they are not necessarily in the image coded data synthesizer 320. There is no need, and depending on the configuration method, even in the image coded data analysis unit 310,
They may be independently arranged outside them.

[0099]

As described above, according to the present invention, the first
The first coded image data is subjected to the first digital signal processing to generate coded data after the signal processing, and the coded data after the signal processing is generated based on a plurality of pieces of information about the first coded data Since the second image signal is subjected to the second digital signal processing to generate the second image encoded data, the first image encoded data is always decoded into the decoded image, and As compared with the case where re-encoding processing is performed on the image coded data of No. 2, the image quality after conversion can be reduced, the processing delay can be shortened, and the device scale can be reduced. There is an effect that the digitized data conversion device can be obtained.

According to the present invention, in the process of generating coded data after signal processing from the first image coded data,
Since a plurality of pieces of information regarding the image coded data of (1) are extracted and the coded data after the signal processing is subjected to the second digital signal processing based on the extracted information, the second image coded data is generated. There is no need to attach special information in the second digital signal processing for generating the second coded data, and an image coded data conversion device that does not require useless information and is efficient in terms of information quantity is obtained. It is effective.

According to the present invention, the separating section allows the first
A plurality of pieces of first image coded data that are combined and sent to the first image coded data and cannot be extracted from the first image coded data used when generating the first image coded data. Since the information is separated and the coded data after the signal processing is subjected to the second digital signal processing based on the information to generate the second image coded data, the second coded data is generated. Information that cannot be extracted in the process of the first digital signal processing can be used in the second digital signal processing for generation, and efficient conversion can be performed as compared with the case where the information is not used. There is an effect that an image coded data conversion device having a property can be obtained.

According to the present invention, the first data necessary for the second digital signal processing is calculated from the image data as the coded data after the signal processing which is decoded from the first image coded data by the information extraction estimation section. Of a plurality of pieces of information regarding the image coded data of (3) are extracted or estimated, and based on the extracted information, the coded data after the signal processing is subjected to the second digital signal processing to generate the second image coded data. Therefore, it is not necessary to perform special processing when performing decoding, and thus there is an effect that an image coded data conversion device that can simplify the device configuration can be obtained.

According to the present invention, the second image coded data image having a data amount different from the data amount of the first image coded data input to the image coded data analysis unit is coded data synthesized. Since it is configured to be generated by a unit, it is possible to obtain an image coded data conversion device that can perform conversion into image coded data having different data amounts with less deterioration in image quality and processing delay after conversion.

According to the present invention, when the second image coded data having a data amount different from that of the first image coded data is generated, a coding process for generating the first image coded data, Since the decoding process for decoding the second image coded data is the coding process based on the transform coding including the transform process and the quantizing process, the decoding process is specialized for transforming and quantizing. This has the effect of enabling efficient conversion of the data amount.

According to the present invention, the coefficient deletion / addition / correction unit reduces a part of the transform coefficient or the quantization index extracted by the image coded data analysis unit,
Since the remaining conversion coefficients or the quantization index are configured to be converted according to the ratio of the amount of data to be converted,
The amount of data can be reduced by deleting a part of the transform coefficient or quantization index, and the remaining part is corrected according to the ratio of the amount of data to be transformed, so it is possible to simply use the transform coefficient or quantization index. There is an effect that the quality of the decoded image can be improved as compared with the case of thinning out.

According to the present invention, the coefficient deletion / addition / correction unit weights the transform coefficient or the quantization index to be reduced with the neighboring transform coefficient or the quantization index to extract the image coded data analysis unit. Since it is configured to reduce a part of the transform coefficient or the quantization index, the data amount can be reduced, and the quality of the decoded image can be improved compared to the case where the transform coefficient is simply thinned. There is.

According to the present invention, the new transform coefficient or quantization index corrected by the coefficient deleting / adding / correcting unit according to the ratio of the amount of data to be converted is converted into the extracted transform coefficient or quantization index. Since it is configured to add a new transform coefficient or quantization index, the new added transform coefficient or quantization index is corrected according to the ratio of the amount of data to be converted. The amount of data can be added in consideration of the image quality after the inverse conversion, and there is an effect that conversion with higher image quality is possible.

According to the present invention, the coefficient deletion / addition / correction unit predicts a conversion coefficient or a quantization index including a conversion coefficient or a quantization index of the addition target and its neighborhood, and then a new conversion coefficient. Alternatively, since the quantization index is configured to be added to the extracted transform coefficient or quantization index, when adding the transform coefficient or quantization index, it is possible to improve the image quality when decoding Since it is performed, it is possible to form an image that is visually easy to see compared to a case where a transform coefficient or a quantization index is simply added, and there is an effect that a higher image quality conversion can be performed.

According to the present invention, if the image type obtained in the process of extracting the transform coefficient or the quantization index by the coefficient deleting / adding / correcting unit is not used for prediction in future encoding, the transform is performed. Since it is configured to increase the ratio of reducing the coefficient or the quantization index, the amount of data relating to the image is calculated for each coding of the time unit frame when the frame is not used in the next time unit coding. Therefore, there is an effect that it is possible to perform high quality conversion while maintaining the total image quality on the time axis.

According to the present invention, if the image type obtained in the process of extracting the transform coefficient or the quantization index by the coefficient deleting / adding / correcting unit is used for prediction in future encoding, Since it is configured to reduce the conversion coefficient or the quantization index reduction ratio, when the frame is used in the next time unit encoding, the reduction ratio of the data amount in the frame becomes small, This has the effect of enabling high-quality conversion while maintaining the total image quality on the time axis.

According to the present invention, the coefficient deletion / addition / correction unit, even if the image type is used for prediction in future encoding, as long as the image block is not used for prediction in future encoding, Since it is configured to increase the ratio of reducing the transform coefficient or the quantization index, whether to increase the deletion ratio of the transform coefficient or the quantization index by using not only the encoded picture information but also the encoded block information. Therefore, it is possible to control in finer units, and it is possible to perform conversion with higher quality when viewed on the time axis.

According to the present invention, the second image coded data having a decoding process procedure different from the decoding process procedure of the first image coded data input to the image coded data analysis unit is image coded. Since it is configured to be generated by the data synthesizing unit, it is possible to perform conversion with high image quality after conversion when generating image coded data of different coding methods.

According to the present invention, the coding parameter correcting / converting unit changes the expression format of the extracted coding parameters from the expression format in the decoding processing procedure of the first image coded data to the second image coding. Since it is configured to be converted into the data decoding processing procedure, the first image coding data input to the image coding data analysis unit is converted into the second image coding data having a different decoding processing procedure. Even when outputting from the image coded data synthesis unit,
Since it is not necessary to decode the first image encoded data into an image once and then re-encode it according to a corresponding encoding method, it is possible to reduce the size and cost of the device. .

According to the present invention, the second image code including the image size temporally or spatially different from the image size included in the first image coded data input to the image coded data analysis unit. Since the encoded data is configured to be generated by the image encoded data synthesizing unit, it is possible to easily convert image encoded data having a high image quality after conversion and temporally or spatially different image sizes. There is an effect that the digitized data conversion device can be obtained.

According to the present invention, when the second image coded data having the image size temporally or spatially different from the first image coded data is generated, the first image coded data is generated. The encoding process to be generated and the decoding process to decode the second image coded data are configured to be a coding process based on transform coding including a motion prediction compensation process, a transform process, and a quantization process. Therefore, there is an effect that it becomes possible to perform motion compensation prediction using a motion vector, or to change the image size by changing the transform coefficient or the quantization step.

According to the present invention, the coefficient deletion / addition / correction unit corrects the image size to be converted and then increases or decreases the extracted conversion coefficient or quantization index. Therefore, when the coded data is converted when the image size is changed, the conversion coefficient quantization index is corrected according to the image size to be converted, which is likely to occur when the image size is converted.
It is possible to suppress the extreme deterioration of the sense of resolution and the unnaturalness, and it is possible to reduce the deterioration of the image quality after the image size is changed.

According to the present invention, the coefficient deletion / addition / correction unit is configured to correct the size of the extracted motion vector according to the ratio of the image sizes to be converted. Is corrected according to the ratio of the image size to be changed, and the efficiency of the motion vector in the image coded data after conversion can be improved. There is an effect that it is possible to obtain characteristics equivalent to a wider range of motion compensation search range.

According to the present invention, the second sequence including the temporal sequence of the image signals different from the temporal sequence of the image signals included in the first encoded image data input to the image encoded data analysis unit. Since the image coded data of is generated by the image coded data synthesizing unit, it is possible to perform conversion into image coded data including a temporal sequence of different image signals with high image quality after conversion. There is an effect that the encoded data conversion device can be obtained.

According to the present invention, the first image coded data is generated when the second image coded data in which the temporal arrangement of the image signals included therein is different from the first image coded data is generated. Since the encoding process for decoding and the decoding process for decoding the second image encoded data are configured to be the encoding process based on the transform encoding with the motion prediction compensation process, the rearranged motion vector is By estimating,
This has the effect of enabling conversion that enables efficient motion compensation prediction.

According to the present invention, by the motion search section,
Since the size of the motion vector extracted by the image coded data analysis unit is configured to be estimated according to the arrangement information of the image to be converted, the size of the motion vector is estimated according to the arrangement information of the image to be converted. By doing so, there is an effect that the coding efficiency by the converted motion vector is improved.

According to the present invention, the image coded data synthesizing unit generates the second image coded data in which the number of decoded image signals included per unit time at the time of decoding is different from the first image coded data. With this configuration, it is possible to obtain an image coded data conversion device that has a high image quality after conversion and can perform conversion into image coded data in which the number of images included in the included decoded image signal is different.

According to the present invention, since the quantization parameter is extracted or estimated as a plurality of pieces of information regarding the first image coded data, the quantization parameter is estimated from the decoded image to newly Since the image quality of the generated second image coded data can be improved, an image coded data conversion device capable of maintaining the image quality after conversion in applications such as relay transmission can be obtained.

According to the present invention, the estimation quantizer estimation unit estimates the quantization parameter when the first image coded data is generated from the decoded image output from the image coded data analysis unit. Since it is configured, it is effective when the image coded data analysis unit performs an operation equivalent to a normal decoding operation, and it becomes possible to perform the highest quality conversion when the rates are the same before and after the conversion, Further, even if the rates are different, there is an effect that optimum conversion can be performed by controlling the quantization parameter according to the ratio.

[Brief description of drawings]

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

FIG. 2 is a block diagram showing a configuration example of an encoding processing device according to the above embodiment.

FIG. 3 is a conceptual diagram for explaining an encoding mode in the encoding processing device.

FIG. 4 is a block diagram showing a configuration example of a decoding processing device in the above embodiment.

FIG. 5 is a block diagram showing a configuration of an image coded data conversion device according to a second embodiment of the present invention.

FIG. 6 is a block diagram showing a configuration of an image coded data conversion device according to a third embodiment of the present invention.

FIG. 7 is a block diagram showing a detailed configuration of an image coded data conversion device according to a fourth to a seventh embodiment of the present invention.

FIG. 8 is an eighth embodiment and a ninth embodiment of the present invention.
2 is a block diagram showing a detailed configuration of an image coded data conversion device according to FIG.

FIG. 9 is a block diagram showing a detailed configuration of an image coded data conversion device according to a tenth embodiment of the present invention.

FIG. 10 is a block diagram showing a detailed configuration of an image coded data conversion device according to an eleventh embodiment of the present invention.

FIG. 11 is a block diagram showing a detailed configuration of an image coded data conversion device according to a twelfth embodiment of the present invention.

FIG. 12 is a block diagram showing a detailed configuration of an image coded data conversion device according to a thirteenth embodiment of the present invention.

FIG. 13 is a block diagram showing a detailed configuration of an image coded data conversion device according to a fourteenth embodiment of the present invention.

FIG. 14 is a block diagram showing a detailed configuration of an image coded data conversion device according to a fifteenth embodiment of the present invention.

FIG. 15 is a block diagram showing a detailed configuration of an image coded data conversion device according to a sixteenth embodiment of the present invention.

FIG. 16 is a block diagram showing a configuration of a conventional image coded data conversion device.

[Explanation of symbols]

30 image coded data conversion device, 40 coding processing device, 50 decoding processing device, 200 input image signal, 22
0 first image coded data, 221 coded data after signal processing, 222, 224, 225 plural pieces of information regarding the first image coded data, 226 quantization index (coding parameter), 227 transform coefficient, 230
Encoded picture information (encoding parameter) 231
Encoding block information (encoding parameter), 232 Quantization parameter information (encoding parameter), 233 Motion vector information (encoding parameter), 236 Decoded image, 240 Second image encoded data, 250 Decoded image signal, 310 Image coded data analysis unit, 320 Image coded data combination unit, 321 Coefficient deletion / addition / correction unit, 324 Coding parameter correction / conversion unit, 329
Motion search unit, 332 quantizer estimation unit, 340 separation unit, 350 information extraction estimation unit.

   ─────────────────────────────────────────────────── ─── Continued front page    (72) Inventor Atsumi Murakami             2-3 2-3 Marunouchi, Chiyoda-ku, Tokyo             Inside Ryo Electric Co., Ltd. F-term (reference) 5C059 KK41 MA00 MA05 MA14 MA23                       MC14 MC38 ME02 NN01 NN21                       PP05 PP06 PP07 RC12 RC16                       SS07 SS12 TA16 TA47 TC00                       TC12 UA02 UA05 UA32 UA33                 5C064 AA02 AB04 AD02 AD14                 5J064 AA02 BA09 BA16 BB01 BB03                       BC16

Claims (2)

[Claims] 1. A coding processing device for performing coding processing on a digitized input image signal based on transform coding including motion-compensated prediction processing using motion vectors, conversion processing, and quantization processing. An image coded data conversion device which receives image coded data No. 1 as an input and performs digital signal processing on the first image coded data to generate second image coded data, wherein the first image coded Image coded data analysis for performing decoding processing on the data, generating decoded image data after the decoding processing, and extracting a coding parameter including a motion vector used for motion compensation regarding the first image coded data And a decoding parameter after the decoding process from the image coded data analysis unit based on the coding parameter relating to the first image coded data from the image coded data analysis unit. When the second image coded data is generated by performing a coding process opposite to the decoding process on the No. image data, the magnitude of the motion vector, which is one of the extracted coding parameters, is set. The first image coded data included in the first image coded data input to the image coded data analysis unit includes a motion search unit that estimates according to the arrangement information of the images to be converted and performs a motion vector search based on the estimated value. Image coding data synthesizing section for generating the second image coded data including a temporal sequence of image signals different from the temporal sequence of image signals Data converter. 2. The first image coded data generated by a coding processing device for coding a digitized input image signal is input, and digital signal processing is performed on the first image coded data. In an image coded data conversion device that generates second image coded data, a decoding process is performed on the first image coded data, and the first image coded data is used as the coded data after the decoding process. From the image coded data analysis unit which outputs the decoded decoded image and the decoded image signal which the image coded data analysis unit outputs,
A quantizer estimation unit that estimates a quantization parameter when the first image coded data is generated is provided, and encoding that is the reverse of the decoding process is performed using the quantization parameter estimated from the decoded image signal. An image coded data conversion unit, comprising: an image coded data synthesizing unit that performs a process on a decoded image obtained by decoding the first image coded data to generate the second image coded data. apparatus.