JPH09233426A - Coding data conversion method, coding data recording method, coding data converter and coding data recorder - Google Patents

Abstract

PROBLEM TO BE SOLVED: To facilitate special reproduction while suppressing increase in a data amount without causing deterioration in the image by converting prediction coding data, which include an orthogonal transformation coefficient and a motion vector resulting from motion compensation of an object image with respect to a reference image and from orthogonal transformation, into in-frame coding data and recording the converted data. SOLUTION: Image data of each frame are processed by a variable length coding decoding circuit 21 and an inverse quantization processing circuit 22 or the like and the result is fed to an adder 24. On the other hand, data decoded by the variable length coding decoding circuit 21 are added to the processed data as above at the adder 24 via a motion vector extract circuit 25 and a motion compensation 26 and output data of the adder 24 are stored in a frame memory 27. The output of the frame memory 27 is fed to the motion compensation circuit 26. Furthermore, the output data from the adder 24 are coded again by a DCT circuit 28 and a quantization circuit 29 and the resulting data are fed to a data recording and reproducing device 37 via a variable length coding circuit 30 and an output buffer circuit 31.

Description

Detailed Description of the Invention

[0001]

[0001]

[0002]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to special features such as fast-forwarding, rewinding, stillness and slow reproduction when recording data transmitted or recorded / reproduced by compressing a moving image by predictive coding by the MPEG system or the like. The present invention relates to conversion of encoded data for converting data into a form suitable for reproduction.

[0003]

[0002]

[0004]

2. Description of the Related Art In recent years, various types of digital moving image compression / expansion methods have been proposed with the progress of multimedia technology. A high efficiency and high compression ratio of these moving image compression / decompression methods A typical MPEG (Motion
Picture Image Coding Experts Group) method has been proposed.

The data structure according to the MPEG system has a hierarchical structure, and a group of images including a video sequence layer, a GOP (Group Of Picture) layer that can be independently reproduced independently, and luminance / color difference image information are arranged from the top. Frame layer to configure, slice layer in scan order, left and right and top and bottom brightness,
The macro block layer includes color difference blocks, and the block layer includes 8 lines × 8 pixels adjacent to each other in luminance or color difference.

[0006]

Among these MPEG data structures, a part of the digital moving picture data contained in the bit stream (GOP layer) is coded with high efficiency and high compression rate and transferred. Digital moving image data may generally have high redundancy (reducible information). That is, there is orthogonal transformation as a method for reducing the redundancy in the spatial direction by utilizing the high correlation between the pixel data value in a certain image and the pixel data value in the vicinity, and the redundancy in the time direction is used. Interframe prediction is a method of reducing the degree. In the MPEG system, a hybrid coding method combining these orthogonal transforms and inter-frame prediction is used, and DCT (Discrete Cosine Transform:
Discrete Cosine Transform) is used, and motion-compensated interframe prediction is adopted in interframe prediction.

[0007]

In order to efficiently compress a digital moving image, first, reference image data, which is a standard image, and a target object are used for a plurality of image frames by a motion compensation inter-frame prediction method. Redundancy in the time direction is reduced by performing a calculation between image data, detecting a motion vector and difference data, and adopting a form capable of predicting an image motion between frames. Then, by performing DCT transformation and quantization on the obtained motion vector and difference data, the reducible information included in the high frequency term of the image data is rounded, and the redundancy in the spatial direction is reduced. Further, it is possible to further compress the information by using the variable length coding which assigns the code length according to the frequency of occurrence of the data.

[0008]

FIG. 6 shows the types of screens in the MPEG system, and there are three types: I frame, P frame and B frame. The subscript indicates the order of the original images. FIG. 6 shows an example in which the first screen is an I frame and two B frames are inserted between the I frame or P frame and the P frame. The I frame (intra-coded screen) is an intra-frame coded screen that encodes the entire screen using only image data in the same frame, and an image can be reconstructed using only its own image data.
The P frame (forward predictive coding screen) shows motion compensation, that is, a motion vector and a difference, with respect to an I frame (or P frame) that is temporally backward (past), as shown in the relationship in the figure. This is a screen for obtaining and encoding. Backward in time

[0009]

The B frame (bidirectional predictive coding screen) performs forward motion compensation with the image for the I (or P) frame that is temporally backward, as shown in the relationship between and in the figure. The backward motion compensation with respect to the image for the P (or I) frame that is temporally forward (future) is performed, the difference obtained by each motion compensation is compared and calculated, and the one having the smaller difference is selected and encoded. It is a screen. The MPEG method employs a block matching method for detecting a motion vector in units of macroblocks in a data hierarchy. As shown in FIG. 7, when the frame layer of one screen is 704 × 480 pixels in the MPEG system, the unit of macroblock used in the block matching method is 16 × 16 pixels.

[0010]

That is, in one macroblock unit, a motion vector and a motion vector for a reference image (reference image) for obtaining a moving destination of an image (target image) which is an operation target image and which has elapsed in time. Motion compensation is performed to find the difference. For example, when one macroblock is used as pixel data of a reference image as shown in FIG. 7B, an image predicted at the position of a motion vector obtained by motion compensation on the reference image (referred to as a prediction image) The image obtained by adding the differential image data (difference image) to is the target image.

A motion compensation method using this block matching method is disclosed in Japanese Patent Publication No. 6-54976 and Japanese Patent Publication No. 6-
Since it is disclosed in Japanese Patent Publication No. 95757, etc., detailed description thereof will be omitted.

[0012]

Next, a method of performing motion compensation for each frame of the bit stream, encoding this, and compressing the data will be described with reference to FIG. FIG.
Shows a state in which one macroblock (16 × 16 pixels) is divided into four equal parts, and one size in this is called a subblock (8 × 8 pixels), which is a DCT transform that is an encoding process. Image data compression processing by quantization is performed in units of this sub-block. The encoding process referred to here is that one image (for example, 704 × 480 pixels) is divided into 8 × 8 pixel square pixel blocks (sub-blocks).

The conversion is performed for each square sub-block.
→ DCT transformation Each coefficient after transformation is a divisor (called quantization step)
Divide by and round the remainder. → Says quantization.

[0014]

In FIG. 8, one macroblock is divided into four.
After performing DCT transformation on one of the equally divided sub-blocks, for example, the sub-block in the figure, quantization processing is performed to compress data. These processes are similarly performed from other sub-blocks, and thereafter, the process is returned to the place corresponding to the number of each sub-block to form 16 × 16 code data. FIG. 9 is a block diagram of a moving picture coding apparatus 10 that processes original images of each frame that are sequentially input. Each processing in the following description is performed in macroblock units as described above. In FIG. 9, reference numeral 11 is a frame A memory. The output of the frame A memory 11 is connected to the DCT conversion circuit 12, the data of a predetermined frame is DCT converted, and then stored in the frame B memory 18 via the quantization circuit 16 and the variable length coding circuit 17. Further, the data in the frame A memory 11 is supplied to the motion detection circuit 13, the motion vector and the difference between predetermined frames are detected, and the motion vector is stored in the MV memory 14.
The differences are stored in the difference memory 15, respectively.

[0015]

The pixel data of the difference image stored in the difference memory 15 is DCT-converted by the DCT conversion circuit 12 and then passed through the quantization circuit 16 and the variable-length coding circuit 17,
It is stored in the frame B memory 18. The motion vector is stored in the frame B memory 18 in the variable-length coding circuit 17 in the state of being added as a header of the pixel data of the coded difference image corresponding to the motion vector.

Now, a processing flow in the moving picture coding apparatus 10 will be described with reference to FIGS. 9 and 10. First, each frame of the bit stream is I ₁ frame, B
_The data is stored in the frame A memory 11 in the order of ₂ frames, B ₃ frame, and P ₄ frame.

First, the I ₁ frame is DC in step S1.
After the DCT conversion by the T conversion circuit 12, step S
In step 2, the quantization circuit 16 quantizes, and in step S3, the variable-length coding circuit 17 performs variable-length coding processing, and the variable-length coding circuit 17 stores it in the frame B memory 18.

[0018]

Here, a symbol "E" indicating that the image data of the I ₁ frame is treated as pixel data is added, and a symbol "C" indicating that the pixel data is encoded is added, which is referred to as CEI _1. . Next, the I ₁ frame and the B ₂ frame stored in the frame A memory 11 are called to the motion detection circuit 13 in step S4, and forward motion detection is performed by the block matching method. That is, the pixel data of the B ₂ frame is compared with the pixel data of the I ₁ frame which is a past image to obtain the motion vector and the difference, the motion vector is stored in the MV memory 14, and the difference is stored in the difference memory 15.
Memorize each. At this time, "V" is added as a symbol indicating the motion vector, which is referred to as VB ₂ .

Further, a symbol "EΔ" indicating the pixel data of the difference image is attached and is referred to as EΔB ₂ .

[0020]

Further, in steps S5 and S6, the same processing as that for the B ₂ frame is performed on the B ₃ frame and the P ₄ frame to obtain VB ₃ , EΔB ₃ and VP.
₄ and EΔP ₄ are obtained. The pixel data (EΔP ₄ ) of the difference image of the P ₄ frame is the DCT conversion circuit 1 of step S7.
It is converted in step 2, and is quantized by the quantization circuit 16 in step S8, and then supplied to the variable length coding circuit 17 in step S9. In step S9, the motion vector (VP ₄ ) stored in the MV memory 14 is called, the motion vector (VP ₄ ) is added as a header of the pixel data (EΔP ₄ ) of the difference image, and the variable length encoding circuit 17 After the variable-length coding processing is performed, it is stored in the frame B memory 18 (described as CEΔP ₄ similarly to the I ₁ frame).

[0021]

Next, in step S10, the P ₄ frame and the B ₃ frame stored in the frame A memory 11 are called to the motion detection circuit 13, and the backward motion detection is performed by the block matching method. That is, the image data of the B ₃ frame is compared with the image data of the P ₄ frame, which is a future image, and the backward motion vector and the difference are obtained. Next, the forward direction difference and the backward direction difference of the B ₃ frame stored in the difference memory 15 are respectively called and compared, and the pixel data (EΔ) of the difference image having the smaller numerical value is calculated.
B ₃ ) is selected. The pixel data (EΔB ₃ ) of the selected difference image is sent to the DCT conversion circuit 12 in step S11.
Are converted by the quantization circuit 16 and are quantized by the quantization circuit 16 in step S12, and then supplied to the variable length coding circuit 17 in step S13.

[0022]

Here, the motion vector (VB ₃ ) which has been selected as a difference from the motion vectors stored in the MV memory 14 is called, and the header of the pixel data (EΔB ₃ ) of the difference image of the B ₃ frame is called. As a result, the motion vector (VB ₃ ) is added, the variable length coding circuit 17 performs variable length coding processing in step S13, and then the frame B memory 18 stores it (CEΔB ₃ ). In addition, in steps S14 to S17, B ₂
The same processing as that for the B ₃ frame is performed on the frame, and the result is stored in the frame B memory 18 (CEΔB
₂ ). Each encoded image data stored in the frame B memory 18, CEI ₁ , CEΔP ₄ , CEΔB ₂ , C
EΔB ₃ is arranged in the order of I ₁ , P ₄ , B ₂ , and B ₃ and transmitted in the form of a bitstream.

[0023]

FIG. 11 is a block diagram of a moving image decoding apparatus and a recording / reproducing apparatus 30 for decoding the image of each frame layer. In FIG. 11, the received bit stream is subjected to variable length code decoding by a variable length code decoding circuit 21, and is supplied to an adder 24 via an inverse quantization circuit 22 and an inverse DCT (IDCT conversion) circuit 23. . On the other hand, the output of the variable length code decoding circuit 21 is the motion vector extraction circuit 2
5 and its output is supplied to the motion compensation circuit 26. The motion compensation circuit 26 performs motion compensation between the motion vector extracted by the motion vector extraction circuit 25 and the image data stored in the frame memory 27, and the obtained image data (predicted image) is sent to the adder 24. Supply and inverse DCT
The image data obtained from the circuit 23 is added. And
The image data output from the adder 24 is stored in the frame memory 27.

[0024]

Now, the processing flow of the moving picture decoding apparatus will be described with reference to FIG. First, the data of each frame is decoded. The decoding process referred to here is to multiply each coefficient by a quantization step for each of the received 8 × 8 square pixels.

Inverse Quantization Inverse transform (inverse DCT operation) is performed for each 8 × 8 square pixel. → Say reverse conversion.

That is, after the encoded data of the I ₁ frame is decoded by the variable length code decoding circuit 21 in step S1, the inverse quantization which is one of the decoding processes is performed by the inverse quantization circuit 22 in step S2. And in step S3, reverse DC
Inverse conversion is performed by the T conversion circuit 23.

[0027]

[0017] Accordingly, I _1-frame coded data (CEI ₁₎ is I ₁ frame pixel data is released is coded coefficient "C" (EI ₁₎ is obtained. Then, it is stored in the frame memory 27. Next, the encoded data of the P ₄ frame is decoded by the variable-length code decoding circuit 21 in step S4, dequantized in step S5, and IDCT-transformed in step S6. This output is used as the pixel data (EΔP ₄ ) of the difference image of the P ₄ frame by the adder 2
4 is supplied.

In step S7, the motion vector extraction circuit 25 extracts the motion vector (VP ₄ ) of the P ₄ frame from the output data of the variable length code decoding circuit 21, and this is output from the inverse DCT conversion circuit 23 in step S3. I ₁ frame pixel data and motion compensation circuit 2
6, and motion compensation is performed.

[0029]

[0018] That is, the pixel data of the I ₁ frame as a reference image, obtaining the pixel data of the predicted image P ₄ frame by causing corresponding motion vectors of the P ₄ frame. Then, in step S8, the predicted image and the differential image are added by the adder 24, P ₄ frames of pixel data (IE
It is stored in the frame memory 27 as P ₄ ). The symbol "I" attached here means "intra" type data in the sense that an image can be restored only by its own data. Next, in steps S9 to S11, the decoding processing of the B ₂ frame is similarly performed, and in step S11, the inverse DCT conversion circuit 23 outputs the pixel data (EΔB ₂ ) of the difference image of the B ₂ frame and the adder 24 Is supplied to. On the other hand, step S12
, The data supplied to the motion compensation circuit 26 is P
_{Of the} motion vectors for ₄ frames or I ₁ frame, the motion vector is for the smaller difference, and motion compensation is performed with the corresponding pixel data of the I ₁ or P ₄ frame output from step S3 or step S8. Be seen.

[0030]

Then, in step S13, the adder 2
In step 4, the difference image and the predicted image are added, and the output is stored in the frame memory 27 as pixel data (IEB ₂ ) of the B ₂ frame. Also, for the B ₃ frame, B
_{The same} process as the pixel data of ₂ frames is performed, and step S1
From step 4 through step S18, the pixel data (IEB ₃ ) of the B ₃ frame is stored in the frame memory 27.

As described above, the pixel data of each frame stored in the frame memory 27 through each step is
I ₁ , IB ₂ , IB ₃ , and IP ₄ are extracted in this order, converted into a TV signal form by the NTSC conversion circuit 33 of the recording / reproducing apparatus 30, and then recorded by an image recording / reproducing apparatus such as a VTR 34. When the received data is directly viewed, the moving image can be viewed by directly reproducing the output signal of the NTSC conversion circuit 33 on the TV receiver 36. Further, when reproducing and viewing the recorded moving image, the reproduced image can be viewed by switching the switch 35 to the VTR 34 side.

It is well known that such recording on a VTR or the like enables special reproduction of forward / reverse rotation, fast-forward, slow, still image, etc. in addition to normal reproduction of moving images.

[0033]

[0020]

[0034]

However, when a video signal is recorded and reproduced as an analog signal in a recording / reproducing apparatus such as a VTR, image quality is deteriorated. On the other hand, if the bit stream shown in FIG. 6 is recorded as a digital signal as it is, deterioration of image quality can be prevented, but it is inconvenient for special reproduction. That is, for example, when performing reverse fast forward reproduction, for example, when it is necessary to sequentially reproduce only B ₁₁ , B ₈ , B ₄ , and B ₂ frames, a procedure for decoding all the frames of the GOP is taken. There must be. To do this, first reproduce and store the data in GPO units,
Unless all frames are decoded, desired special reproduction cannot be supported at any time.

[0035]

The control and data processing at the time of reproduction as described above are complicated, and since the data of each frame is stored in the frame memory in the form of image data, the memory capacity becomes enormous.

The present invention has been made by paying attention to the above-mentioned problems, and when recording compressed moving image data, since it is recorded in a special reproduction mode with little deterioration in image quality and increase in data amount, the predictive coding is performed. An object of the present invention is to provide an encoded data conversion method, an encoded data recording method, an encoded data conversion device, and an encoded data recording device for converting data into intra-frame encoded data and recording it.

[0037]

[0022]

[0038]

The present invention has been made in view of the above-mentioned problems, and the coded data recording method according to claim 1 is obtained by motion compensation and orthogonal transformation of a reference image of a target image. It is characterized in that the predictive coded data including the motion vector and the orthogonal transform coefficient of the difference image are converted into intraframe coded data and recorded.

[0039]

Further, in the coded data conversion method according to the second aspect, the predictive coded data including the motion vector obtained by the motion compensation and the orthogonal conversion of the reference image of the target image and the orthogonal transform coefficient of the difference image are intra-framed. A coded data conversion method for converting into coded data, the step of obtaining pixel data of a reference image, and the step of extracting predicted image pixel data from the obtained reference image pixel data based on a motion vector of a target image And a step of performing inverse orthogonal transformation of the orthogonal transformation coefficient of the difference image to obtain pixel data of the difference image, and adding the obtained difference image pixel data and the extracted predicted image pixel data to obtain pixel data of the target image. And a step of orthogonally transforming the pixel data of the target image into an orthogonal transform coefficient of the target image as intraframe coded data. To.

[0040]

The coded data conversion apparatus according to a third aspect of the present invention includes, within a frame, predictive coded data including a motion vector obtained by motion compensation and orthogonal transformation of a reference image of a target image and an orthogonal transformation coefficient of a difference image. A coded data conversion device for converting into coded data, means for obtaining pixel data of a reference image, and means for extracting predicted image pixel data from the obtained reference image pixel data based on a motion vector of a target image. And means for inversely orthogonally transforming the orthogonal transformation coefficient of the difference image to obtain pixel data of the difference image, and the obtained reference image pixel data and the extracted predicted image pixel data to obtain pixel data of the target image. And a means for orthogonally transforming the pixel data of the target image to obtain an orthogonal transform coefficient of the target image as intraframe coded data. To.

[0041]

Further, in the coded data conversion method according to the fourth aspect, the predictive coded data including the motion vector obtained by the motion compensation and the orthogonal transform of the reference image of the target image and the orthogonal transform coefficient of the difference image are intra-framed. A coded data conversion method for converting into coded data, the step of obtaining pixel data of a reference image, and the step of extracting predicted image pixel data from the obtained reference image pixel data based on a motion vector of a target image And a step of orthogonally transforming the extracted predicted image pixel data to obtain an orthogonal transformation coefficient of the predicted image, and adding the obtained orthogonal transformation coefficient of the predicted image and the orthogonal transformation coefficient of the difference image to intra-frame encoded data. It is characterized in that it has a step of making an orthogonal transformation coefficient of the transformed target image.

[0042]

The coded data conversion apparatus according to a fifth aspect of the present invention includes, within a frame, predictive coded data including a motion vector obtained by motion compensation and orthogonal transformation of a reference image of a target image and an orthogonal transformation coefficient of a difference image. A coded data conversion device for converting into coded data, means for obtaining pixel data of a reference image, and means for extracting predicted image pixel data from the obtained reference image pixel data based on a motion vector of a target image. And means for orthogonally transforming the extracted predicted image pixel data to obtain an orthogonal transform coefficient of the predicted image, and a means for adding the obtained orthogonal transform coefficient of the predicted image and the orthogonal transform coefficient of the difference image to intraframe coded data. It is characterized in that it has means for making an orthogonal transformation coefficient of the transformed target image.

[0043]

Further, the coded data conversion recording device according to a sixth aspect of the present invention frame-codes predictive coded data including a motion vector obtained by motion compensation and orthogonal transformation of a reference image of a target image and an orthogonal transformation coefficient of a difference image. A recording device for converting into intra-encoded data and recording, and means for obtaining pixel data of a reference image and means for extracting predicted image pixel data from the obtained reference image pixel data based on a motion vector of a target image And means for orthogonally transforming the extracted predicted image pixel data to obtain an orthogonal transform coefficient of the predicted image, and a means for adding the obtained orthogonal transform coefficient of the predicted image and the orthogonal transform coefficient of the difference image to intraframe coded data. The present invention is characterized by having means for making an orthogonal transformation coefficient of the transformed target image, and recording means for recording the orthogonal transformation coefficient of the subject image on a predetermined recording medium. To.

[0044]

The coded data conversion method according to a seventh aspect of the present invention includes a step of inversely orthogonally transforming the orthogonal transform coefficient of the target image converted into the intraframe coded data in the fourth aspect to obtain pixel data of the target image. It is characterized by further having.

[0045]

[0029]

[0046]

[Action]

(1) In the invention of claim 1, since the predictive encoded data is converted into the intra-frame encoded data and recorded,
Special reproduction becomes easy.

(2) In the inventions of claims 2 and 3,
First, the predicted image pixel data is extracted from the reference image pixel data based on the motion vector of the target image, while the orthogonal transform coefficient of the difference image is inversely transformed into pixel data.
The predicted image pixel data and the difference image pixel data are added to obtain the pixel data of the target image. By orthogonally transforming this, it is possible to obtain intraframe coded data.

(3) In the inventions of claims 4 and 5,
First, the predicted image pixel data is extracted from the reference image pixel data based on the motion vector of the target image, and the predicted image pixel data is orthogonally transformed into the predicted image orthogonal transformation coefficient. By adding this and the orthogonal transformation coefficient of the difference image, it is possible to obtain intraframe coded data.

(4) In the sixth aspect of the invention, the orthogonal transformation coefficient of the target image converted into the intra-frame coded data is obtained in the same manner as in the above (3), and this is recorded on the predetermined recording medium by the recording means. Will be recorded.

(5) In the invention of claim 7, first,
Similar to the above item (3), the orthogonal transform coefficient of the target image converted into the intra-frame coded data is obtained, and the recording signal that can be specially reproduced is obtained. Next, this is subjected to inverse orthogonal transform to become pixel data of the target image. As a result, it is possible to obtain a signal that can be recorded or viewed by the conventional recording / reproducing apparatus.

[0051]

[0030]

[0052]

1 is a block diagram of a moving image reprocessing circuit and its reproducing circuit according to a first embodiment of the present invention. In the description of the embodiments of the present invention, the same parts as those of the conventional example are designated by the same reference numerals, and the duplicated description will be omitted. 2 is a processing flow chart for the moving image reprocessing circuit 20, and is a processing flow chart in which steps of the variable length code decoding circuit and the inverse quantization circuit shown in FIG. 1 are omitted.

In FIG. 1A, the moving image reprocessing circuit 2
The image data of each frame supplied to 0 as an input is decoded by the variable length code decoding circuit 21 and then supplied to the inverse quantization circuit 22 and the IDCT circuit 23 which are decoding processing.
The data subjected to the inverse DCT conversion by the IDCT circuit 23 is supplied to the adder 24.

[0054]

On the other hand, the data decoded by the variable length code decoding circuit 21 is supplied to the motion vector extraction circuit 25, and the motion vector data is extracted. The extracted motion vector data is supplied to the motion compensation circuit 26. The output data of the motion compensation circuit 26 and the output data of the IDCT circuit 23 are added by the adder 24, and the output data is stored in the frame memory 27. The output of the frame memory 27 is supplied to the motion compensation circuit 26. Further, the output data of the adder 24 is re-encoded by the DCT circuit 28 and the quantization circuit 29, and then supplied to the output buffer circuit 31 via the variable length encoding circuit 30. In addition, the output buffer circuit 31
Is connected to the input terminal of the moving image reprocessing circuit 20.
The output of the output buffer circuit 31 is the data recording / reproducing device 37.
Is supplied to. The data recording / reproducing device 37 records / reproduces this encoded data.

[0055]

In FIG. 1B, the reproduced coded data is supplied to the reproducing circuit 40, is decoded again by the variable length code decoding circuit 42 via the input buffer 41, and is dequantized by the inverse quantization circuit 43 and the IDCT circuit 44. The decryption process is performed. The decoded pixel data is converted into analog image data by the NTSC conversion circuit 33 and then supplied to the TV receiver to watch a moving image.

Here, the relationship of data processing for encoding a moving image, decoding image data sent in the form of a bit stream, and converting it again into encoded data will be described using mathematical expressions.

[0057]

[0033]

[0058]

[Equation 1]

[0059]

[0034]

[0060]

[Equation 2]

[0061]

[0035]

[0062]

(Equation 3)

[0063]

[0036]

[0064]

(Equation 4)

[0065]

Equations 1 and 2 show the relationship when the image data of the moving image is encoded. The left side of Expression 1 indicates a DCT coefficient of an I-picture image block, that is, DCT-transformed (CEI). Like the right side, pixel data (EI) of the I-picture is matrix-multiplied by a DCT coefficient. It shows that it is required.
Practically, the pixel data is encoded in 8 × 8 pixel sub-block units as shown in FIG. 7. At this time, the sub-blocks are calculated in the form of a matrix of 8 digits and 8 rows. I mean.

[0066]

Similarly, Expression 2 is described for a P picture or a B picture, and the left side is obtained by encoding differential pixel data of the P picture or the B picture (CE
ΔP, CEΔB), and the right side is a matrix subtraction of pixel data (EP, EB) of the P picture or B picture from the corresponding predicted image data (with a symbol “F” indicating the predicted image), that is, the difference. Pixel data (EΔP,
It shows that B) is DCT-transformed. As described with reference to FIG. 7B, for example, the pixel data of the P picture is the pixel data of the difference image and the pixel data of the prediction image that are matrix-added. However, when encoding, the pixel data of the P picture ( EP) is subjected to matrix subtraction of pixel data (EFP) of the predicted image, that is, pixel data (EΔP) of the difference image is encoded.

[0067]

Expression 3 is a mathematical expression for decoding the above-mentioned coded image data, and generally when decoding this on the decoder side, the following is performed.

That is, the image data (CEΔP, CEΔB) of the encoded P picture or B picture is subjected to IDCT conversion, and the pixel data of each prediction image is added to this by matrix addition to obtain the pixel data of the P picture or B picture. I am trying.

Now, in order to obtain the intra-type DCT coefficient of the P picture or B picture, both sides of the equation 3 are DCT
Considering the conversion, Equation 4 is obtained from this. The last side of Expression 4 is image data (CEΔP, CEΔB) of a P picture or B picture that is not an intra type.
"CIEP, B" to indicate that the
Is written. Therefore, from Equation 4, the image data of each intra-coded picture (P picture or B picture) is obtained by subjecting the pixel data of the predicted image to motion compensation, DCT-transformed this, and predictive coding of each picture. It can be seen that the image data of the DCT coefficient can be obtained by matrix addition.

[0070]

The operation of the moving picture reprocessing circuit 20 will be described below with reference to the processing flow chart of FIG.

First, the image data (CEI ₁ ) of the I ₁ frame supplied to the input terminal of the moving image reprocessing circuit 20 is inversely DCT-transformed (EI ₁ ) by the IDCT circuit 23 in step S1, and then step S2. With frame memory 2
7 is stored as I ₁ frame pixel data. Similarly, the image data (CEΔP ₄ ) of the P ₄ frame is subjected to inverse DCT conversion (EΔP) by the IDCT circuit 23 in step S3.
P ₄₎ after being supplied to the adder 24. In step S4, in the motion compensation circuit 26, the motion vector (VP ₄ ) of the P ₄ frame extracted by the motion vector extraction circuit 25,
Motion compensation is performed with the I ₁ frame image data stored in the frame memory 27. As a result, the pixel data (FEP ₄ ) of the P ₄ frame predicted image is obtained as the output of the motion compensation circuit 26.

[0072]

Then, in step S5, IDCT
The difference image (EΔP ₄ ) output from the circuit 23 and the pixel data (FEP ₄ ) of the predicted image obtained at the output of the motion compensation circuit 26 are added by the adder 24, and the pixel data of the P ₄ frame (EP ₄ ) Is obtained. The pixel data (EP ₄ ) of the P ₄ frame is stored in the frame memory 27 at step S6, and then the DCT at step S7.
It is re-encoded (CIEP ₄ ) in the conversion circuit 28. Then, quantization and variable length coding are performed, and the output buffer 3
1 is stored. As described above, the pixel data (EP ₄ ) of the P ₄ frame is the data that can reconstruct the image only with the pixel data of its own, and encoding this is as follows.
This corresponds to the intra-coded image that is the definition of the I frame. Therefore, the symbol "I" indicating that the pixel data is re-encoded and I-framed is added, and is described as CIEP ₄ .

[0073]

Next, the B ₂ frame image data (CEΔ
B ₂ ) is an inverse DC signal generated by the IDCT circuit 23 in step S8.
After being T-converted (EΔB ₂ ), it is supplied to the adder 24.

On the other hand, in step S9, the motion compensation circuit 26 performs motion compensation by the motion vector (VB ₂ ) of the B ₂ frame extracted by the motion vector extraction circuit 25 and the pixel data of the I ₁ frame or the P ₄ frame. Done. In step S10, the difference image (EΔB ₂ ) output from the IDCT circuit 23 and the predicted image (EFB ₂ ) output from the motion compensation circuit 26 are added by the adder 24, and the pixel data (EB ₂ of the B ₂ frame is added. ) Get. The output of the adder 24 is re-encoded in step S11 and stored in the output buffer 31 as I-framed image data (CIEB ₂ ).

[0075]

Similarly, the pixel data (CEΔB ₃ ) of the B ₃ frame is obtained from steps S12 to S15.
Then, the same processing as that for the B ₂ frame is performed, and the image data (CIEB ₃ ) converted into the I frame is stored in the output buffer 31. Further, the image data (CEI ₁ ) of the I ₁ frame is stored in the output buffer 31 as it is. As described above, the image data stored in the output buffer 31 is collected in frame units, and if necessary,
Information indicating the frame number and the type of the original frame is added as a header, for example. Then, CEI ₁ , CIEB ₂ , CIEB ₃ , and CIEP, which are the frame order of the original image,
₄ , output to the data recording / reproducing device 37 in this order,
Be recorded. This image data is recorded and reproduced by the data recording / reproducing device 37.
Is reproduced and converted into pixel data by the reproduction circuit 40. At this time, the data recording / reproducing device 37 may sequentially search and output pixel data of only a required frame based on the frame number or the like according to the reproduction mode. If the processing speed of the data recording / reproducing apparatus is sufficiently high, all the frames may be reproduced and only the image data of the required frames may be selected and output.

[0076]

FIG. 3 is a block diagram of a second embodiment according to the present invention, which is made based on the above-mentioned equation 4. FIG. 3 (a)
In, the image data of each frame supplied to the input of the moving image reprocessing circuit 50 is decoded by the variable length code decoding circuit 21, and then supplied to the inverse quantization circuit 22 and the IDCT circuit 23. The data subjected to the inverse DCT conversion by the IDCT circuit 23 is supplied to the adder 24. On the other hand, the data decoded by the variable length code decoding circuit 21 is supplied to the motion vector extraction circuit 25, and the motion vector data is extracted. The extracted motion vector data is supplied to the motion compensation circuit 26. One output data and ID of the motion compensation circuit 26
The output data of the CT circuit 23 is added by the adder 24, and the output data is stored in the frame memory 27. The output data of the frame memory 27 is the motion compensation circuit 2
6.

[0077]

The other output data of the motion compensation circuit 26 is D
The output data is supplied to the CT circuit 28, and the output data is supplied to one terminal of a newly provided adder 32. Further, the output data of the inverse quantization circuit 22 is supplied to the other terminal of the adder 34 and added to the output data of the DCT circuit 28. The output data of the adder 32 is supplied to the output buffer 31 via the quantization circuit 29 and the variable length coding circuit 30. The output of the output buffer 31 is connected to the input terminal of the moving image reprocessing circuit 50.

The operation of the moving image reprocessing circuit 50 will be described below with reference to the processing flow chart of FIG. First, the pixel data (CEI ₁ ) of the I ₁ frame supplied to the input terminal of the moving image reprocessing circuit 50 is inversely DCT-transformed (EI ₁ ) by the IDCT circuit 23 in step S1, and then the frame memory 27 in step S2. Memorized in.

[0079]

Further, the image data (CEΔP ₄ ) of the P ₄ frame difference image is inversely DCT-transformed (EΔP ₄ ) by the IDCT circuit 23 in step S3, and then added by the adder 2
4 is supplied. In step S4, the motion compensation circuit 26 stores the motion vector (VP ₄ ) of the P ₄ frame extracted by the motion vector extraction circuit 25 and the frame memory 27.
Motion compensation is performed with the pixel data of the I ₁ frame stored in. In step S5, the difference image (EΔP ₄ ) output from the IDCT circuit 23 and the P ₄ frame predicted image (E ΔP ₄ ) output from the motion compensation circuit 26 (E
FP ₄ ) is added by the adder 24. The output of the adder 24 is pixel data (EP ₄ ) of the P ₄ frame, and the step S
It is stored in the frame memory 27 at 6. On the other hand, in step S7, the pixel data (EFP ₄ ) of the P ₄ frame predicted image output from the motion compensation circuit 26 is re-encoded by the DCT circuit 28 (CEFP ₄ ).

[0080]

Then, in step 8, the image data (CEΔP ₄ ) of the P ₄ frame difference image supplied to the input terminal of the moving image reprocessing circuit 50 and the recoded P ₄ obtained from the DCT circuit 28. The image data (CEFP ₄ ) of the predicted image of the frame is added by the adder 32 (C
EΔP ₄ + CEFP ₄ ). The image data output from the adder 32 is DCT × (EΔP ₄ + EFP ₄ ) = DCT × E
In other P ₄ and calligraphy, found to be in a form obtained by encoding the pixel data of the P ₄ frame, corresponding to an intra coded picture of P ₄ frame. Then, it is stored in the output buffer 31 as intra-coded image data (CIEP ₄ ).

[0081]

Next, in step 9, the B ₂ frame motion vector (VB ₂ ) extracted by the motion vector extraction circuit 25 and the I ₁ frame pixel data (EI ₁ ) obtained in step 2 or step 6 Motion compensation is performed with the pixel data (EP ₄ ) of the P ₄ frame obtained in step 1, and the pixel data (EFB ₂ ) of the predicted image of the B ₂ frame
Is output. Then, in step 10, the DCT conversion circuit 2
Re-encoded by 8 (CEFB ₂ ). Step 11
Then, the image data (CEΔB ₂ ) of the difference image of the B ₂ frame input to the moving image reprocessing circuit 50 and the DCT circuit 2
The pixel data (CEFB ₂ ) of the re-encoded B ₂ frame predicted image obtained from 8 is added by the adder 32 (CEΔB ₂ + CEFB ₂ ).

[0082]

The image data obtained here corresponds to the intra-coded image of the B ₂ frame, as in the case of the P ₄ frame. Then, it is stored in the output buffer 31 as intra-coded image data (CIEB ₂ ).
Similarly, the B ₃ frame is also stored in the output buffer 31 as intra-coded image data (CIEB ₃ ) through steps 12 to 14.
Thereafter, the same processing as in the first embodiment is performed. Further, frames other than the above-mentioned frames, such as P ₇ frame and B
_{The 5} frames are converted into intra-coded data by the same method.

As can be seen by comparing the processing flow charts of the second embodiment with the first embodiment,
It requires less processing. That is, when the processing from the I ₁ frame to the B ₃ frame is simply viewed, the IDCT conversion is reduced twice.

[0084]

In the second embodiment described above, since there is no process in which the image data is inversely converted into pixel data during data recording, the image cannot be viewed during recording as it is.
FIG. 5 is a block diagram of a third embodiment that enables this.
In this figure, by switching the input of the reproducing circuit 40 to the output side of the moving image reprocessing circuit 50 by the switch 44 during recording, the video being recorded can be viewed. On the other hand, during reproduction, the switch 44 may be switched to the output side of the data recording / reproducing device 37.

As a data storage / reproduction device, DA
Magnetic tapes such as T and digital VTRs, hard disks,
Any device capable of recording and reproducing digital data, such as an optical disk such as a CD-R or VDR, or a solid-state memory may be used. In addition, at the time of recording, in order to enable special reproduction by random access, the data includes, for example, a control signal, a header, an ID number,
It is preferable to add some sort of searchable identification information suitable for each recording / reproducing apparatus such as a frame number and an address number.

[0086]

[0051]

[0087]

As described above, according to the present invention, the intra-coded prediction coded data including the motion vector obtained by the motion compensation and the orthogonal transform of the reference image of the target image and the orthogonal transform coefficient of the difference image is encoded. Since the converted data is converted and recorded, it is possible to facilitate special reproduction while suppressing an increase in the data amount without causing deterioration of the image. At this time, as a method of converting into intraframe coded data, first, predictive image pixel data is extracted from the reference image pixel data based on the motion vector of the target image, and this is orthogonally transformed into a predictive image orthogonal transform coefficient. The processing can be simplified by adding and the orthogonal transformation coefficient of the difference image.

[Brief description of drawings]

FIG. 1 is a block diagram of a moving image reprocessing circuit and a reproducing circuit according to a first embodiment of the present invention.

FIG. 2 is a processing flowchart for the moving image reprocessing circuit according to the first embodiment of the present invention.

FIG. 3 is a block diagram of a moving image reprocessing circuit and its reproducing circuit according to a second embodiment of the present invention.

FIG. 4 is a processing flowchart for a moving image reprocessing circuit according to a second embodiment of the present invention.

FIG. 5 is a block diagram of a moving image reprocessing circuit according to a third embodiment of the present invention.

FIG. 6 is a diagram illustrating a screen type in the MPEG system.

FIG. 7 is a diagram illustrating a macroblock in the MPEG system.

FIG. 8 is a diagram showing sub-blocks for encoding.

FIG. 9 is a block diagram of a moving picture coding circuit in the MPEG system.

FIG. 10 is a processing flow chart for a moving picture coding circuit in the MPEG system.

FIG. 11 is a block diagram of a moving image decoding device and a recording / reproducing device in a conventional example.

FIG. 12 is a processing flowchart for the moving picture decoding apparatus in the conventional example.

[Explanation of symbols]

 20, 50 ... Moving image reprocessing circuit 21, 42 ... Variable length code decoding circuit 22, 43 ... Inverse quantization circuit 23, 44, 38 ... IDCT circuit 24, 32 ... Adder 25 ... Motion vector extraction circuit 26 ... Motion compensation circuit 27 ... Frame memory 28 ... DCT circuit 29 ... Quantization circuit 30 ... Variable length coding circuit 31 ... Output buffer 33 ... NTSC conversion circuit 34 ... VTR 35,44 ... Switch 36 ... TV receiver 37 ... Data recording / reproducing device 40 ... Reproducing circuit 41 ... Input buffer

─────────────────────────────────────────────────── ───

[Procedure amendment]

[Submission date] January 16, 1997

[Procedure amendment 1]

[Document name to be amended] Statement

[Correction target item name] Detailed description of the invention

[Correction method] Change

[Correction contents]

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to special features such as fast-forwarding, rewinding, stillness and slow reproduction when recording data transmitted or recorded / reproduced by compressing a moving image by predictive coding by the MPEG system or the like. The present invention relates to conversion of encoded data for converting data into a form suitable for reproduction.

[0002]

2. Description of the Related Art In recent years, various types of digital moving image compression / expansion methods have been proposed with the progress of multimedia technology. A high efficiency and high compression ratio of these moving image compression / decompression methods A typical MPEG (Mot
ion Picture Image Coding
The Experts Group) method has been proposed. The data structure according to the MPEG system has a hierarchical structure, and a video sequence layer from the upper part and a GOP (Group of Pictures) capable of independent reproduction independently.
e) Layers, frame layers that make up a grouped image including luminance / color difference image information, slice layers in scan order, left and right and top and bottom luminance, macroblock layers consisting of color difference blocks, and 8 lines with adjacent luminance or color differences The block layer is composed of × 8 pixels.

Among these MPEG data structures, a part of the digital moving picture data contained in the bit stream (GOP layer) is coded with high efficiency and high compression rate and transferred. Digital moving image data may generally have high redundancy (reducible information). That is, there is orthogonal transformation as a method for reducing the redundancy in the spatial direction by utilizing the high correlation between the pixel data value in a certain image and the pixel data value in the vicinity, and the redundancy in the time direction is used. Interframe prediction is a method of reducing the degree. In the MPEG system, a hybrid coding method that combines these orthogonal transforms and interframe prediction is used, and DCT (Discrete Cosine) is used as the orthogonal transform.
(Transform: Discrete Cosine Transform) is used, and motion-compensated inter-frame prediction is incorporated in inter-frame prediction.

In order to efficiently compress a digital moving image, first, reference image data, which is a standard image, and a target object are used for a plurality of image frames by a motion compensation inter-frame prediction method. Redundancy in the time direction is reduced by performing a calculation between image data, detecting a motion vector and difference data, and adopting a form capable of predicting an image motion between frames. Then, by performing DCT transformation and quantization on the obtained motion vector and difference data, the reducible information included in the high frequency term of the image data is rounded, and the redundancy in the spatial direction is reduced. Further, it is possible to further compress the information by using the variable length coding which assigns the code length according to the frequency of occurrence of the data.

FIG. 6 shows the types of screens in the MPEG system, and there are three types: I frame, P frame and B frame. The subscript indicates the order of the original images. FIG. 6 shows an example in which the first screen is an I frame and two B frames are inserted between the I frame or P frame and the P frame. The I frame (intra-coded screen) is an intra-frame coded screen that encodes the entire screen using only image data in the same frame, and an image can be reconstructed using only its own image data.
The P frame (forward predictive coding screen) shows motion compensation, that is, a motion vector and a difference, with respect to an I frame (or P frame) that is temporally backward (past), as shown in the relationship in the figure. This is a screen for obtaining and encoding. Backward in time

The B frame (bidirectional predictive coding screen) performs forward motion compensation with the image for the I (or P) frame that is temporally backward, as shown in the relationship between and in the figure. The backward motion compensation with respect to the image for the P (or I) frame that is temporally forward (future) is performed, the difference obtained by each motion compensation is compared and calculated, and the one having the smaller difference is selected and encoded. It is a screen. The MPEG method employs a block matching method for detecting a motion vector in units of macroblocks in a data hierarchy. As shown in FIG. 7, when the frame layer of one screen is 704 × 480 pixels in the MPEG system, the unit of macroblock used in the block matching method is 16 × 16 pixels.

That is, in one macroblock unit, a motion vector and a motion vector for a reference image (reference image) for obtaining a moving destination of an image (target image) which is an operation target image and which has elapsed in time. Motion compensation is performed to find the difference. For example, when one macroblock is used as pixel data of a reference image as shown in FIG. 7B, an image predicted at the position of a motion vector obtained by motion compensation on the reference image (referred to as a prediction image) The image obtained by adding the differential image data (difference image) to is the target image. The motion compensation method using this block matching method is described in Japanese Patent Publication No. 6-5.
Since it is disclosed in Japanese Patent Publication No. 4976, Japanese Patent Publication No. 6-95757, etc., detailed description thereof will be omitted.

Next, a method of performing motion compensation for each frame of the bit stream, encoding this, and compressing the data will be described with reference to FIG. FIG.
Shows a state in which one macroblock (16 × 16 pixels) is divided into four equal parts, and one size in this is called a subblock (8 × 8 pixels), which is a DCT transform that is an encoding process. Image data compression processing by quantization is performed in units of this sub-block. The encoding process referred to here is that one image (for example, 704 × 480 pixels) is divided into 8 × 8 pixel square pixel blocks (sub-blocks). Conversion is performed for each square sub-block. → DCT transformation Each coefficient after transformation is a divisor (called quantization step)
Divide by and round the remainder. → Says quantization.

In FIG. 8, one macroblock is divided into four.
After performing DCT transformation on one of the equally divided sub-blocks, for example, the sub-block in the figure, quantization processing is performed to compress data. These processes are similarly performed from other sub-blocks, and thereafter, the process is returned to the place corresponding to the number of each sub-block to form 16 × 16 code data. FIG. 9 is a block diagram of a moving picture coding apparatus 10 that processes original images of each frame that are sequentially input. Each processing in the following description is performed in macroblock units as described above. In FIG. 9, reference numeral 11 is a frame A memory. The output of the frame A memory 11 is connected to the DCT conversion circuit 12, and after the data of a predetermined frame is DCT-converted, it is stored in the frame B memory 18 via the quantization circuit 16 and the variable length coding circuit 17. Further, the data in the frame A memory 11 is supplied to the motion detection circuit 13, the motion vector and the difference between predetermined frames are detected, and the motion vector is stored in the MV memory 14.
The differences are stored in the difference memory 15, respectively.

The pixel data of the difference image stored in the difference memory 15 is DCT-converted by the DCT conversion circuit 12 and then passed through the quantization circuit 16 and the variable-length coding circuit 17,
It is stored in the frame B memory 18. The motion vector is stored in the frame B memory 18 in the variable-length coding circuit 17 in the state of being added as a header of the pixel data of the coded difference image corresponding to the motion vector. Now, a processing flow in the moving picture coding apparatus 10 will be described here with reference to FIGS. 9 and 10. First, each frame of the bit stream is stored in the frame A memory 11 in the order of I ₁ frame, B ₂ frame, B ₃ frame, and P ₄ frame. First, after the I ₁ frame is DCT-converted by the DCT conversion circuit 12 in step S1,
Quantization is performed by the quantization circuit 16 in step S2, variable length encoding processing is performed by the variable length encoding circuit 17 in step S3, and the result is stored in the frame B memory 18.

Here, a symbol "E" indicating that the image data of the I ₁ frame is treated as pixel data is attached, and a symbol "C" indicating that the pixel data is encoded is attached, which is referred to as CEI _1. . Next, the I ₁ frame and the B ₂ frame stored in the frame A memory 11 are called to the motion detection circuit 13 in step S4, and forward motion detection is performed by the block matching method. That is, the pixel data of the B ₂ frame is compared and calculated with the pixel data of the I ₁ frame which is a past image to obtain a motion vector and a difference. The motion vector is stored in the MV memory 14, and the difference is stored in the difference memory 15.
Memorize each. At this time, "V" is added as a symbol indicating the motion vector, which is referred to as VB ₂ . In addition, a symbol “EΔ” indicating the pixel data of the difference image is attached and is referred to as EΔB ₂ .

Further, in steps S5 and S6, the B ₃ frame and the P ₄ frame are processed in the same manner as the B ₂ frame, and VB ₃ , EΔB ₃ and VP are processed.
₄ and EΔP ₄ are obtained. The pixel data (EΔP ₄ ) of the difference image of the P ₄ frame is the DCT conversion circuit 1 of step S7.
It is converted in step 2, and is quantized by the quantization circuit 16 in step S8, and then supplied to the variable length coding circuit 17 in step S9. In step S9, the motion vector (VP ₄ ) stored in the MV memory 14 is called, the motion vector (VP ₄ ) is added as a header of the pixel data (EΔP ₄ ) of the difference image, and the variable length encoding circuit 17 After the variable-length coding process is performed, it is stored in the frame B memory 18 (denoted as CEΔP ₄ similarly to the I ₁ frame).

[0013] Next, in step S10, calls the _{P 4} frames and _{B 3} frames stored in the frame memory A 11 to the motion detecting circuit 13 performs the backward motion detected by the block matching method. That is, the image data of the B ₃ frame is compared with the image data of the P ₄ frame, which is a future image, to obtain the backward motion vector and the difference. Next, the forward direction difference and the backward direction difference of the B ₃ frame stored in the difference memory 15 are respectively called and compared, and the pixel data (EΔ) of the difference image having the smaller numerical value is calculated.
B ₃ ) is selected. The pixel data (EΔB ₃ ) of the selected difference image is output to the DCT conversion circuit 12 in step S11.
Are converted by the quantization circuit 16 and are quantized by the quantization circuit 16 in step S12, and then supplied to the variable length coding circuit 17 in step S13.

Here, the motion vector (VB ₃ ) which has been selected as a difference from the motion vectors stored in the MV memory 14 is called, and the header of the pixel data (EΔB ₃ ) of the difference image of the B ₃ frame is called. As a result, a motion vector (VB ₃ ) is added, the variable length coding circuit 17 performs variable length coding processing in step S13, and then the frame B memory 18 stores it (CEΔB ₃ ). In addition, in steps S14 to S17, B ₂
The same processing as that for the B ₃ frame is performed on the frame, and the result is stored in the frame B memory 18 (CEΔB
₂ ). Each coded image data stored in the frame B memory 18, CEI ₁ , CEΔP ₄ , CEΔB ₂ , C
EΔB ₃ is arranged in the order of I ₁ , P ₄ , B ₂ , and B ₃ , and is transmitted in the form of a bitstream.

FIG. 11 is a block diagram of a moving image decoding apparatus and a recording / reproducing apparatus 30 for decoding the image of each frame layer. In FIG. 11, the received bit stream is subjected to variable length code decoding by a variable length code decoding circuit 21, and is supplied to an adder 24 via an inverse quantization circuit 22 and an inverse DCT (IDCT conversion) circuit 23. . On the other hand, the output of the variable length code decoding circuit 21 is the motion vector extraction circuit 2
5 and its output is supplied to the motion compensation circuit 26. The motion compensation circuit 26 performs motion compensation between the motion vector extracted by the motion vector extraction circuit 25 and the image data stored in the frame memory 27, and the obtained image data (predicted image) is sent to the adder 24. Supply and inverse DCT
The image data obtained from the circuit 23 is added. And
The image data output from the adder 24 is stored in the frame memory 27.

Now, the processing flow of the moving picture decoding apparatus will be described with reference to FIG. First, the data of each frame is decoded. The decoding process referred to here is to multiply each coefficient by a quantization step for each of the received 8 × 8 square pixels. Inverse quantization Inverse transformation (inverse DCT operation) is performed for each 8 × 8 square pixel. → Say reverse conversion. That is, after the encoded data of the I ₁ frame is decoded by the variable length code decoding circuit 21 in step S1,
In step S2, the inverse quantization circuit 22 performs the decoding process 1
The inverse quantization is performed, and the inverse DCT transform circuit 23 performs inverse transform in step S3.

[0017] Accordingly, _{I 1-frame} coded data (CEI ₁₎ is _{I 1} frame pixel data is released is coded coefficient "C" (EI ₁₎ is obtained. Then, it is stored in the frame memory 27. Then, the encoded data of P ₄ frame after being decoded by the variable-length code decoding circuit 21 in step S4, the inverse quantization in step S5, the IDCT converted in step S6. This output is added by the adder 2 as pixel data (EΔP ₄ ) of the P ₄ frame difference image.
4 is supplied. In step S7, the motion vector extraction circuit 25 outputs the output data of the variable length code decoding circuit 21.
The motion vector (VP ₄ ) of the P ₄ frame is extracted by this, and this and the pixel data of the I ₁ frame output from the inverse DCT conversion circuit 23 in step S3 are supplied to the motion compensation circuit 26 to perform motion compensation. .

[0018] That is, the pixel data of the I ₁ frame as a reference image, obtaining the pixel data of the predicted image P ₄ frame by causing corresponding motion vectors of the P ₄ frame. Then, in step S8, the predicted image and the differential image are added by the adder 24, P ₄ frames of pixel data (IE
It is stored in the frame memory 27 as P ₄ ). The symbol "I" attached here means "intra" type data in the sense that an image can be restored only by its own data. Next, in steps S9 to S11, the decoding processing of the B ₂ frame is similarly performed, and in step S11, the inverse DCT conversion circuit 23 outputs the pixel data (EΔB ₂ ) of the difference image of the B ₂ frame and the adder 24 Is supplied to. On the other hand, step S12
, The data supplied to the motion compensation circuit 26 is P
It is a motion vector for a smaller difference among the motion vectors for ₄ frames or I ₁ frame, and motion compensation is performed with the corresponding pixel data of I ₁ or P ₄ frame which is the output of step S3 or step S8. Be seen.

Then, in step S13, the adder 2
In 4, the difference image and the prediction image are added, and the output is stored in the frame memory 27 as pixel data (IEB ₂ ) of the B ₂ frame. Also, for B ₃ frames, B
_{The same} processing as the pixel data of ₂ frames is performed, and step S1
The pixel data (IEB ₃ ) of the B ₃ frame is stored in the frame memory 27 from 4 through step S18.
As described above, the frame memory 27
Pixel data of each frame stored in the, I _1, I
B ₂ , IB ₃ , and IP ₄ are extracted in this order, and the recording / reproducing device 3
After being converted into a TV signal form by the NTSC conversion circuit 33 of 0, it is recorded by an image recording / reproducing device such as a VTR 34. Also, if you want to directly view the received data, NTS
The moving image can be viewed by directly reproducing the output signal of the C conversion circuit 33 on the TV receiver 36. If you want to play back the recorded video, switch 3
5 can be switched to the VTR 34 side to view the reproduced image. It is well known that special recording such as forward / reverse rotation, fast forward, slow and still images can be performed in addition to normal moving image reproduction by recording on a VTR or the like.

[0020]

However, when a video signal is recorded and reproduced as an analog signal in a recording / reproducing apparatus such as a VTR, image quality is deteriorated. On the other hand, if the bit stream shown in FIG. 6 is recorded as a digital signal as it is, deterioration of image quality can be prevented, but it is inconvenient for special reproduction. That is, for example, when performing reverse fast forward reproduction, for example, when it is necessary to sequentially reproduce only B ₁₁ , B ₈ , B ₄ , and B ₂ frames, a procedure for decoding all the frames of the GOP is taken. There must be. To do this, first reproduce and store the data in GPO units,
Unless all frames are decoded, desired special reproduction cannot be supported at any time.

The control and data processing at the time of reproduction as described above are complicated, and since the data of each frame is stored in the frame memory in the form of image data, the memory capacity becomes enormous. The present invention has been made in view of the above-mentioned problems. When recording compressed moving image data, the predictive coded data is recorded in a frame so as to enable special reproduction with little deterioration in image quality and increase in data amount. An object of the present invention is to provide an encoded data conversion method, an encoded data recording method, an encoded data conversion device, and an encoded data recording device for converting and recording inner encoded data.

[0022]

The present invention has been made in view of the above-mentioned problems, and the coded data recording method according to claim 1 is obtained by motion compensation and orthogonal transformation of a reference image of a target image. It is characterized in that the predictive coded data including the motion vector and the orthogonal transform coefficient of the difference image are converted into intraframe coded data and recorded.

Further, in the coded data conversion method according to the second aspect, the predictive coded data including the motion vector obtained by the motion compensation and the orthogonal conversion of the reference image of the target image and the orthogonal transform coefficient of the difference image are intra-framed. A coded data conversion method for converting into coded data, the step of obtaining pixel data of a reference image, and the step of extracting predicted image pixel data from the obtained reference image pixel data based on a motion vector of a target image And a step of performing inverse orthogonal transformation of the orthogonal transformation coefficient of the difference image to obtain pixel data of the difference image, and adding the obtained difference image pixel data and the extracted predicted image pixel data to obtain pixel data of the target image. And a step of orthogonally transforming the pixel data of the target image into an orthogonal transform coefficient of the target image as intraframe coded data. To.

The coded data conversion apparatus according to a third aspect of the present invention includes, within a frame, predictive coded data including a motion vector obtained by motion compensation and orthogonal transformation of a reference image of a target image and an orthogonal transformation coefficient of a difference image. A coded data conversion device for converting into coded data, means for obtaining pixel data of a reference image, and means for extracting predicted image pixel data from the obtained reference image pixel data based on a motion vector of a target image. And means for inversely orthogonally transforming the orthogonal transformation coefficient of the difference image to obtain pixel data of the difference image, and the obtained reference image pixel data and the extracted predicted image pixel data to obtain pixel data of the target image. And a means for orthogonally transforming the pixel data of the target image to obtain an orthogonal transform coefficient of the target image as intraframe coded data. To.

Further, in the coded data conversion method according to the fourth aspect, the predictive coded data including the motion vector obtained by the motion compensation and the orthogonal transform of the reference image of the target image and the orthogonal transform coefficient of the difference image are intra-framed. A coded data conversion method for converting into coded data, the step of obtaining pixel data of a reference image, and the step of extracting predicted image pixel data from the obtained reference image pixel data based on a motion vector of a target image And a step of orthogonally transforming the extracted predicted image pixel data to obtain an orthogonal transformation coefficient of the predicted image, and adding the obtained orthogonal transformation coefficient of the predicted image and the orthogonal transformation coefficient of the difference image to intra-frame encoded data. It is characterized in that it has a step of making an orthogonal transformation coefficient of the transformed target image.

The coded data conversion apparatus according to a fifth aspect of the present invention includes, within a frame, predictive coded data including a motion vector obtained by motion compensation and orthogonal transformation of a reference image of a target image and an orthogonal transformation coefficient of a difference image. A coded data conversion device for converting into coded data, means for obtaining pixel data of a reference image, and means for extracting predicted image pixel data from the obtained reference image pixel data based on a motion vector of a target image. And means for orthogonally transforming the extracted predicted image pixel data to obtain an orthogonal transform coefficient of the predicted image, and a means for adding the obtained orthogonal transform coefficient of the predicted image and the orthogonal transform coefficient of the difference image to intraframe coded data. It is characterized in that it has means for making an orthogonal transformation coefficient of the transformed target image.

Further, the coded data conversion recording device according to a sixth aspect of the present invention frame-codes predictive coded data including a motion vector obtained by motion compensation and orthogonal transformation of a reference image of a target image and an orthogonal transformation coefficient of a difference image. A recording device for converting into intra-encoded data and recording, and means for obtaining pixel data of a reference image and means for extracting predicted image pixel data from the obtained reference image pixel data based on a motion vector of a target image And means for orthogonally transforming the extracted predicted image pixel data to obtain an orthogonal transform coefficient of the predicted image, and a means for adding the obtained orthogonal transform coefficient of the predicted image and the orthogonal transform coefficient of the difference image to intraframe coded data. The present invention is characterized by having means for making an orthogonal transformation coefficient of the transformed target image, and recording means for recording the orthogonal transformation coefficient of the subject image on a predetermined recording medium. To.

The coded data conversion method according to a seventh aspect of the present invention includes a step of inversely orthogonally transforming the orthogonal transform coefficient of the target image converted into the intraframe coded data in the fourth aspect to obtain pixel data of the target image. It is characterized by further having.

[0029]

(1) In the invention of claim 1, since the predictive encoded data is converted into the intra-frame encoded data and recorded,
Special reproduction becomes easy. (2) In the inventions of claims 2 and 3, first, the predicted image pixel data is extracted from the reference image pixel data based on the motion vector of the target image, while the orthogonal transform coefficient of the difference image is inversely transformed to obtain the pixel data. It is said that The predicted image pixel data and the difference image pixel data are added to obtain the pixel data of the target image. By orthogonally transforming this, it is possible to obtain intraframe coded data. (3) In the inventions of claims 4 and 5, first, the predicted image pixel data is extracted from the reference image pixel data based on the motion vector of the target image, and the predicted image pixel data is orthogonally transformed to obtain the predicted image orthogonal transformation coefficient. By adding this and the orthogonal transformation coefficient of the difference image, it is possible to obtain intraframe coded data. (4) In the invention of claim 6, the orthogonal transformation coefficient of the target image converted into the intra-frame coded data is obtained in the same manner as in the above (3), and this is recorded on the predetermined recording medium by the recording means. . (5) In the invention of claim 7, first, the orthogonal transform coefficient of the target image converted into the intra-frame coded data is obtained in the same manner as in the above (3), and a special reproducible recording signal is obtained. Next, this is subjected to inverse orthogonal transform to become pixel data of the target image. As a result, it is possible to obtain a signal that can be recorded or viewed by the conventional recording / reproducing apparatus.

[0030]

1 is a block diagram of a moving image reprocessing circuit and its reproducing circuit according to a first embodiment of the present invention. In the description of the embodiments of the present invention, the same parts as those of the conventional example are designated by the same reference numerals, and the duplicated description will be omitted. 2 is a processing flow chart for the moving image reprocessing circuit 20, and is a processing flow chart in which steps of the variable length code decoding circuit and the inverse quantization circuit shown in FIG. 1 are omitted. FIG.
In (a), the image data of each frame supplied to the input to the moving image reprocessing circuit 20 is the variable length code decoding circuit 2
Inverse quantization circuit 22 which is a decoding process after being decoded by 1.
And the IDCT circuit 23. IDCT circuit 23
The data subjected to the inverse DCT conversion in (1) is supplied to the adder 24.

On the other hand, the data decoded by the variable length code decoding circuit 21 is supplied to the motion vector extraction circuit 25, and the motion vector data is extracted. The extracted motion vector data is supplied to the motion compensation circuit 26. The output data of the motion compensation circuit 26 and the output data of the IDCT circuit 23 are added by the adder 24, and the output data is stored in the frame memory 27. The output of the frame memory 27 is supplied to the motion compensation circuit 26. The output data of the adder 24 is re-encoded by the DCT circuit 28 and the quantizing circuit 29, and then supplied to the output buffer circuit 31 via the variable length encoding circuit 30. In addition, the output buffer circuit 31
Is connected to the input terminal of the moving image reprocessing circuit 20.
The output of the output buffer circuit 31 is the data recording / reproducing device 37.
Is supplied to. The data recording / reproducing device 37 records / reproduces this encoded data.

In FIG. 1B, the reproduced coded data is supplied to the reproducing circuit 40, is decoded again by the variable length code decoding circuit 42 via the input buffer 41, and is dequantized by the inverse quantization circuit 43 and the IDCT circuit 44. The decryption process is performed. The decoded pixel data is converted into analog image data by the NTSC conversion circuit 33 and then supplied to the TV receiver to watch a moving image. Here, the relationship of data processing of encoding a moving image, decoding image data transmitted in the form of a bit stream, and converting it into encoded data again will be described using mathematical expressions.

[0033]

(Equation 1)

[0034]

(Equation 2)

[0035]

(Equation 3)

[0036]

(Equation 4)

Equations 1 and 2 show the relationship when the image data of the moving image is encoded. The left side of Expression 1 indicates a DCT coefficient of an I-picture image block, that is, DCT-transformed (CEI). Like the right side, pixel data (EI) of the I-picture is matrix-multiplied by a DCT coefficient. It shows that it is required.
Practically, the pixel data is encoded in 8 × 8 pixel sub-block units as shown in FIG. 7. At this time, the sub-blocks are calculated in the form of a matrix of 8 digits and 8 rows. I mean.

Similarly, Expression 2 is described for a P picture or a B picture, and the left side is obtained by encoding differential pixel data of the P picture or the B picture (CE
ΔP, CEΔB), and the right side is a matrix subtraction of pixel data (EP, EB) of the P picture or B picture from the corresponding predicted image data (with a symbol “F” indicating the predicted image), that is, the difference. Pixel data (EΔP,
It shows that B) is DCT-transformed. As described with reference to FIG. 7B, for example, the pixel data of the P picture is the pixel data of the difference image and the pixel data of the prediction image that are matrix-added. However, when encoding, the pixel data of the P picture ( EP) is subjected to matrix subtraction of pixel data (EFP) of the predicted image, that is, pixel data (EΔP) of the difference image is encoded.

Expression 3 is a mathematical expression for decoding the above-mentioned coded image data, and generally when decoding this on the decoder side, the following is performed. That is, the encoded P-picture or B-picture image data (CEΔ
(P, CEΔB) is subjected to IDCT conversion, and pixel data of each predicted image is added to this by matrix to obtain pixel data of P picture or B picture. Now, considering that DCT conversion is performed on both sides of Expression 3 in order to obtain an intra-type DCT coefficient of a P picture or a B picture, Expression 4 is obtained from this. The last side of Expression 4 is described as “CIEP, B” to indicate that the image data (CEΔP, CEΔB) of a P picture or B picture that is not an intra type has been converted to intra. Therefore, from Equation 4, the image data of each intra-coded picture (P picture or B picture) is obtained by subjecting the pixel data of the predicted image to motion compensation, DCT-transformed this, and predictive coding of each picture. It can be seen that the image data of the DCT coefficient can be obtained by matrix addition.

The operation of the moving picture reprocessing circuit 20 will be described below with reference to the processing flow chart of FIG. First, the image data (CEI ₁ ) of the I ₁ frame supplied to the input terminal of the moving image reprocessing circuit 20 is inversely DCT-transformed (EI ₁ ) by the IDCT circuit 23 in step S1, and then step S1.
In step 2, it is stored in the frame memory 27 as pixel data of I ₁ frame. Similarly, the image data (CEΔP ₄ ) of the P ₄ frame is inversely DCT-transformed (EΔP ₄ ) by the IDCT circuit 23 in step S3 and then supplied to the adder 24. In step S4, the motion compensation circuit 26 performs motion compensation based on the P ₄ frame motion vector (VP ₄ ) extracted by the motion vector extraction circuit 25 and the I ₁ frame image data stored in the frame memory 27. Done. As a result, P ₄ is output as the output of the motion compensation circuit 26.
Pixel data (FEP ₄ ) of the predicted image of the frame is obtained.

Then, in step S5, IDCT
The difference image (EΔP ₄ ) output from the circuit 23 and the pixel data (FEP ₄ ) of the predicted image obtained at the output of the motion compensation circuit 26 are added by the adder 24, and the pixel data of the P ₄ frame (EP ₄ ) Is obtained. The pixel data (EP ₄ ) of the P ₄ frame is stored in the frame memory 27 at step S6, and then the DCT at step S7.
It is re-encoded (CIEP ₄ ) in the conversion circuit 28. Then, quantization and variable length coding are performed, and the output buffer 3
1 is stored. As described above, the pixel data (EP ₄ ) of the P ₄ frame is the data that can reconstruct the image only with the pixel data of its own, and encoding it is as follows.
This corresponds to the intra-coded image that is the definition of the I frame. Therefore, a symbol "I" indicating that the pixel data is re-encoded and I-framed is added, and is referred to as CIEP ₄ .

Next, B ₂ frame image data (CEΔ
B ₂ ) is inverse DC by the IDCT circuit 23 in step S8.
After being T-converted (EΔB ₂ ), it is supplied to the adder 24. On the other hand, in step S9, in the motion compensation circuit 26,
Motion compensation is performed by the motion vector (VB ₂ ) of the B ₂ frame extracted by the motion vector extraction circuit 25 and the pixel data of the I ₁ frame or the P frame. In step S10, the difference image (EΔB ₂ ) output from the IDCT circuit 23 and the prediction image (EFB ₂ ) output from the motion compensation circuit 26 are added by the adder 24, and the pixel data (EB ₂ ) of the B ₂ frame is added. ) Get. The output of the adder 24 is re-encoded in step S11 and stored in the output buffer 31 as I-framed image data (CIEB ₂ ).

Similarly, the pixel data (CEΔB ₃ ) of the B ₃ frame is obtained from steps S12 to S15.
After that, the same processing as that for the B ₂ frame is performed, and the image data is converted to an I frame (CIEB ₃ ) and stored in the output buffer 31. Further, the image data (CEI ₁ ) of the I ₁ frame is stored in the output buffer 31 as it is. As described above, the image data stored in the output buffer 31 is collected in frame units, and if necessary,
Information indicating the frame number and the type of the original frame is added as a header, for example. Then, CEI ₁ , CIEB ₂ , CIEB ₃ , and CIEP, which are the frame order of the original image,
₄ , output to the data recording / reproducing device 37 in this order,
Be recorded. This image data is recorded and reproduced by the data recording / reproducing device 37.
Is reproduced and converted into pixel data by the reproduction circuit 40. At this time, the data recording / reproducing device 37 may sequentially search and output pixel data of only a required frame based on the frame number or the like according to the reproduction mode. If the processing speed of the data recording / reproducing apparatus is sufficiently high, all the frames may be reproduced and only the image data of the required frames may be selected and output.

FIG. 3 is a block diagram of a second embodiment according to the present invention, which is made based on the above-mentioned equation 4. FIG. 3 (a)
In, the image data of each frame supplied to the input of the moving image reprocessing circuit 50 is decoded by the variable length code decoding circuit 21, and then supplied to the inverse quantization circuit 22 and the IDCT circuit 23. The data subjected to the inverse DCT conversion by the IDCT circuit 23 is supplied to the adder 24. On the other hand, the data decoded by the variable length code decoding circuit 21 is supplied to the motion vector extraction circuit 25, and the motion vector data is extracted. The extracted motion vector data is supplied to the motion compensation circuit 26. One output data and ID of the motion compensation circuit 26
The output data of the CT circuit 23 is added by the adder 24, and the output data is stored in the frame memory 27. The output data of the frame memory 27 is the motion compensation circuit 2
6.

The other output data of the motion compensation circuit 26 is D
The output data is supplied to the CT circuit 28, and the output data is supplied to one terminal of a newly provided adder 32. Further, the output data of the inverse quantization circuit 22 is supplied to the other terminal of the adder 34 and added to the output data of the DCT circuit 28. The output data of the adder 32 is supplied to the output buffer 31 via the quantization circuit 29 and the variable length coding circuit 30. The output of the output buffer 31 is connected to the input terminal of the moving image reprocessing circuit 50. Now, the operation of the moving image reprocessing circuit 50 will be described with reference to the processing flow chart of FIG. First, the pixel data (CEI ₁ ) of the I ₁ frame supplied to the input terminal of the moving image reprocessing circuit 50 is subjected to inverse DCT conversion by the IDCT circuit 23 in step S1 (E
I ₁₎ after it stored in the frame memory 27 in step S2.

The image data (CEΔP ₄ ) of the P ₄ frame difference image is inversely DCT-transformed (EΔP ₄ ) by the IDCT circuit 23 in step S3, and then the adder 2 is added.
4 is supplied. In step S4, the motion compensation circuit 26 detects the motion vector (VP ₄ ) of the P ₄ frame extracted by the motion vector extraction circuit 25 and the frame memory 27.
Motion compensation is performed with the pixel data of the I ₁ frame stored in. In step S 5, the difference image (EΔP ₄ ) output from the IDCT circuit 23 and the P ₄ frame predicted image (E ΔP ₄ ) output from the motion compensation circuit 26.
FP ₄ ) is added by the adder 24. The output of the adder 24 is pixel data (EP ₄ ) of the P ₄ frame, and the step S
It is stored in the frame memory 27 at 6. On the other hand, in step S7, the pixel data (EFP ₄ ) of the P ₄ frame predicted image output from the motion compensation circuit 26 is re-encoded by the DCT circuit 28 (CEFP ₄ ).

In step 8, the image data (CEΔP ₄ ) of the P ₄ frame difference image supplied to the input terminal of the moving image reprocessing circuit 50 and the re-encoded P ₄ obtained from the DCT circuit 28. The image data (CEFP ₄ ) of the predicted image of the frame is added by the adder 32 (C
EΔP ₄ + CEFP ₄ ). The image data output from the adder 32 is DCT × (EΔP ₄ + EFP ₄ ) = DCT × E
In other P ₄ and calligraphy, found to be in a form obtained by encoding the pixel data of the P ₄ frame, corresponding to an intra coded picture of P ₄ frame. Then, it is stored in the output buffer 31 as intra-coded image data (CIEP ₄ ).

Next, in step 9, the motion vector (VB ₂ ) of the B ₂ frame extracted by the motion vector extraction circuit 25 and the pixel data (EI ₁ ) of the I ₁ frame obtained in step 2 or step 6 Motion compensation is performed with the pixel data (EP ₄ ) of the P ₄ frame obtained in step S 1, and the pixel data (EFB ₂ ) of the predicted image of the B ₂ frame
Is output. Then, in step 10, the DCT conversion circuit 2
Re-encoded by 8 (CEFB ₂ ). Step 11
Then, the image data (CEΔB ₂ ) of the difference image of the B ₂ frame input to the moving image reprocessing circuit 50 and the DCT circuit 2
The pixel data (CEFB ₂ ) of the re-encoded B ₂ frame predicted image obtained from 8 is added by the adder 32 (CEΔB ₂ + CEFB ₂ ).

The image data obtained here corresponds to the intra-coded image of the B ₂ frame, as in the case of the P ₄ frame. Then, it is stored in the output buffer 31 as intra-coded image data (CIEB ₂ ).
Similarly, the B ₃ frame is also stored in the output buffer 31 as intra-coded image data (CIEB ₃ ) through steps 12 to 14.
Thereafter, the same processing as in the first embodiment is performed. Further, frames other than the above-mentioned frames, for example, P ₇ frame and B
_{The 5} frames are also converted into intra-coded data by the same method. The second embodiment requires less processing than the first embodiment, as can be seen by comparing the processing flow charts. That is, when the processing from the I ₁ frame to the B ₃ frame is simply viewed, the IDCT conversion is reduced twice.

In the second embodiment described above, since there is no process in which the image data is inversely converted into pixel data during data recording, the image cannot be viewed during recording as it is.
FIG. 5 is a block diagram of a third embodiment that enables this.
In this figure, by switching the input of the reproducing circuit 40 to the output side of the moving image reprocessing circuit 50 by the switch 44 during recording, the video being recorded can be viewed. On the other hand, during reproduction, the switch 44 may be switched to the output side of the data recording / reproducing device 37. The data storage / reproduction device may be any device capable of recording / reproducing digital data, such as a magnetic tape such as DAT or a digital VTR, a hard disk, an optical disc such as a CD-R or VDR, or a solid-state memory. . In addition, at the time of recording, in order to enable special reproduction by random access, the control signal, header,
It is advisable to add some sort of searchable identification information suitable for each recording / reproducing apparatus, such as an ID number, a frame number, and an address number.

[0051]

As described above, according to the present invention, the intra-coded prediction coded data including the motion vector obtained by the motion compensation and the orthogonal transform of the reference image of the target image and the orthogonal transform coefficient of the difference image is encoded. Since the converted data is converted and recorded, it is possible to facilitate special reproduction while suppressing an increase in the data amount without causing deterioration of the image. At this time, as a method of converting into intraframe coded data, first, predictive image pixel data is extracted from the reference image pixel data based on the motion vector of the target image, and this is orthogonally transformed into a predictive image orthogonal transform coefficient. The processing can be simplified by adding and the orthogonal transformation coefficient of the difference image.

Claims (7)

[Claims] 1. A predictive coded data including a motion vector obtained by motion compensation and orthogonal transform of a reference image of a target image and an orthogonal transform coefficient of a difference image is converted into intraframe coded data and recorded. A method for recording encoded data. 2. A coded data conversion method for converting predictive coded data including a motion vector obtained by motion compensation and orthogonal transform of a reference image of a target image and an orthogonal transform coefficient of a difference image into intraframe coded data. There, the step of obtaining the pixel data of the reference image, the step of extracting the predicted image pixel data from the obtained reference image pixel data based on the motion vector of the target image, the orthogonal transformation coefficient of the difference image A step of performing inverse orthogonal transformation to obtain pixel data of the difference image; a step of adding the obtained difference image pixel data and the extracted predicted image pixel data to obtain pixel data of the target image; Encoding of the target image as the intraframe coded data by orthogonally transforming the pixel data of Over data conversion method. 3. A coded data conversion device for converting predictive coded data including a motion vector obtained by motion compensation and orthogonal transform of a target image of a target image and an orthogonal transform coefficient of a difference image into intraframe coded data. There, means for obtaining pixel data of the reference image, means for extracting predicted image pixel data from the obtained reference image pixel data based on the motion vector of the target image, orthogonal transformation coefficient of the difference image Means for performing inverse orthogonal transformation to obtain pixel data of the difference image; means for adding the obtained reference image pixel data and the extracted predicted image pixel data to obtain pixel data of the target image; Encoding means for orthogonally transforming the pixel data of the above to obtain an orthogonal transform coefficient of the target image as the intra-frame encoded data. Over data conversion device. 4. A coded data conversion method for converting predictive coded data including a motion vector obtained by motion compensation and orthogonal transform of a reference image of a target image and an orthogonal transform coefficient of a difference image into intraframe coded data. There is a step of obtaining pixel data of the reference image, a step of extracting predicted image pixel data from the obtained reference image pixel data based on a motion vector of the target image, the extracted predicted image pixel data And the step of obtaining the orthogonal transformation coefficient of the predicted image by orthogonal transformation, the orthogonal transformation coefficient of the obtained predicted image and the orthogonal transformation coefficient of the difference image are added to the target image transformed into the intra-frame coded data. A method for transforming coded data, characterized in that it has a step of making an orthogonal transform coefficient. 5. A coded data conversion device for converting predictive coded data including a motion vector obtained by motion compensation and orthogonal transform of a target image of a target image and an orthogonal transform coefficient of a difference image into intraframe coded data. There, means for obtaining pixel data of the reference image, means for extracting predicted image pixel data from the obtained reference image pixel data based on the motion vector of the target image, the extracted predicted image pixel data Means for obtaining the orthogonal transformation coefficient of the predicted image by orthogonal transformation of the target image transformed into the intra-frame coded data by adding the obtained orthogonal transformation coefficient of the predicted image and the orthogonal transformation coefficient of the difference image. An encoded data conversion device, characterized in that it has means for making an orthogonal transform coefficient. 6. A recording device for converting predictive coded data including a motion vector obtained by motion compensation and orthogonal transform of a reference image of a target image and an orthogonal transform coefficient of a difference image into intraframe coded data and recording the data. A means for obtaining pixel data of the reference image, a means for extracting predicted image pixel data from the obtained reference image pixel data based on a motion vector of the target image, and the extracted predicted image pixel data Means for obtaining the orthogonal transformation coefficient of the predicted image by orthogonal transformation, and adding the orthogonal transformation coefficient of the obtained predicted image and the orthogonal transformation coefficient of the difference image to obtain the orthogonality of the target image transformed into the intra-frame coded data. Encoded data, characterized by having means for using a transform coefficient and recording means for recording the orthogonal transform coefficient of the target image on a predetermined recording medium. Recording apparatus. 7. The coded data conversion method according to claim 4, further comprising the step of inversely orthogonally transforming the orthogonal transform coefficient of the target image converted into the intraframe coded data to obtain pixel data of the target image.