KR0178206B1 - Adaptive image coding system - Google Patents

Abstract

The present invention can be implemented with less computation and simple hardware by using decimation and interoperation, and it is possible to generate a plurality of compressed video signals having at least different screen sizes corresponding to different receivers with different aspect ratios. The present invention relates to a video encoding system. To this end, the present invention divides a current frame into 8 × 8 DCT blocks, and then uses a cosine function to time-domain video signals for each of the divided DCT blocks. A DCT block for converting a DCT conversion coefficient value in the frequency domain of the block into a DCT conversion coefficient value of 8 × 8 of the converted current frame and each DCT conversion coefficient value of the corresponding 8 × 8 corresponding frame of the previous frame on the time axis. The coded specificization is performed after the compression coding by applying vector quantization using the first code table to the first difference signal obtained by using A first encoding path for providing a video signal of a size to the transmitter, and each DCT transform coefficient value of a low-pass portion of a predetermined size extracted from each DCT transform coefficient value of 8 × 8 of the converted current frame and a corresponding previous frame on the time axis The compression-coded specific screen output from the first encoding path by applying vector quantization using a second code table to second differential signals obtained based on respective DCT transform coefficient values of corresponding low-pass portions of a predetermined size of? And a second encoding path for providing the transmitter with an encoded video signal having a screen size at least smaller than the size.

Accordingly, the video encoding system of the present invention can effectively generate a compressed coded video signal having at least a different screen size (number of lines and number of pixels) that can meet the demands of the channel without unnecessarily excessive calculation amount and configuration complexity. will be.

Description

Adaptive Video Coding System

1 is a block diagram of an adaptive video encoding system according to a preferred embodiment of the present invention.

2 is a diagram illustrating an example of decimating an 8x8 DCT block into 4x4 blocks and interpolating the same into an 8x8 DCT block in adaptive image encoding according to the present invention.

* Explanation of symbols for main parts of the drawings

10, 28, 44: Frame memory 12: DCT block

14, 14, 18, 34 Subtractor 16 Interpolation Block

22, 38: vector quantization block 24, 40: inverse vector quantization block

26, 42: adder 30, 46: motion prediction block

32: decimation block

The present invention relates to an encoding system for compressing and encoding an image signal. More particularly, the present invention relates to an adaptive image encoding method for generating an image signal applicable to monitors having different sizes using decimation and interoperation. It is about the system.

As is well known in the art, the transmission of discrete video signals can maintain better image quality than analog signals. When a video signal composed of a series of image frames is represented in a digital form, a considerable amount of data must be transmitted, particularly in the case of high-definition television (HDTV). However, since the usable frequency range of the conventional transmission channel is limited, in order to transmit a large amount of digital data, it is necessary to compress the data to be transmitted and reduce its transmission amount. In addition, the compressed video signal and the audio signal are encoded through different coding techniques due to the characteristics of the signals. In this encoding, a video signal compression technique in which a larger amount of digital data is generated than an audio signal is particularly important. It can be said to take part.

On the other hand, as the various compression methods mainly used for encoding the video signal, the hybrid coding method combining the stochastic coding method and the temporal and spatial compression method is known to be the most efficient.

Most of hybrid coding schemes, which are one of the efficient coding schemes described above, use motion compensated DPCM (differential pulse code modulation), two-dimensional discrete cosine transform (DCT), quantization of DCT coefficients, VLC (variable modulation coding), and the like. Here, the motion compensation DPCM determines a motion of the object between the current frame and the previous frame, and predicts the current frame according to the motion of the object to generate a differential signal representing the difference between the current frame and the prediction value. Such methods are described, for example, in Staffan Ericsson's Fixed and Adaptive Predictors for Hybrid Predictive / Transform Coding, IEEE Transactions on Communication, COM-33, NO. 12 (Dec. 1985), or A motion Compensated Interframe Coding Scheme for Television Pictures, IEEE Transactions on Communication, COM-30, NO.1 (January, 1982) by Ninomiy and Ohtsuka.

More specifically, the motion compensation DPCM predicts the current frame from the previous frame according to the motion of the object estimated between the current frame and the previous frame. Here, the estimated motion may be represented by a two-dimensional motion vector representing the displacement between the previous frame and the current frame.

In general, there are various approaches to estimating the pixel displacement of an object, and they are generally classified into two types, one of which is a motion estimation method in block units and the other is a motion estimation method in pixel units. In motion estimation, the block of the current frame is compared with the blocks of the previous frame to determine an optimal matching block. From this, the interframe displacement vector (how much the block has moved between frames) for the entire transmitted frame is then determined. It is estimated.

Therefore, when transmitting the video signal, the transmitting side compresses and encodes the image signal in the buffer of the output side in order by taking into account the spatial and temporal correlation of the image signal in block units or pixel units through the above-described encoding technique. The received video data will be transmitted to the decoding system on the receiving side through the transmission channel at a desired bit rate in response to the channel request.

More specifically, the encoding system on the transmitting side removes spatial redundancy of the video signal by using transform coding such as discrete cosine transform (DCT), and further uses temporal encoding of the video signal by using differential encoding through motion estimation and prediction. By eliminating redundant redundancy, the video signal can be efficiently compressed.

On the other hand, in developing high-definition television (HDTV) broadcasting, which has recently been developed and entered into the commercialization stage, the HDTV receiver capable of receiving high-definition broadcasting at the present time, which can be referred to as the early stage of the development of the receiver and the start of broadcasting. It is very poorly distributed, and it will take at least a few years for the HDTV receiver to be widely spread (ie universalized) considering the full-scale mass production at the production site and securing the popularity of supply prices. It is estimated.

Therefore, as described above, during the period required for the diffusion and diffusion of HDTV receivers, that is, during a transition period where a general TV receiver (or a computer monitor) and an HDTV receiver coexist, a substantial high-definition television broadcast can also be received by a general TV receiver (or a computer monitor). Technology to ensure that this is inevitably involved.

In general, in the case of transmitting a high quality television broadcast from a broadcasting station or the like during the transition period, a user equipped with a monitor such as a general television receiver or a computer can have a decoding system capable of restoring a compressed coded video signal to an original signal. By providing a built-in inverter (built-in or external TV), it is possible to watch high-definition television broadcasts. In this case, the screen size of high-definition television broadcasting is higher than that of ordinary TV receivers (or computer monitors). With the size of 16 × 9, the undesirable result would be that some screens of high quality television broadcasts received on a normal TV receiver (or computer monitor) are displayed in a cut form.

On the other hand, in eliminating the undesirable cut-off phenomenon as described above, the desired size of the image using a technique such as sampling or decimation at the time of encoding the image in the transmitting side (e.g., broadcasting station) encoding system. The user can adjust the image to a desired size by using a technique such as sampling or decimation, similarly to the above-described encoding system, when the image is decoded by the receiving system (receiver). When you want to watch high-definition television broadcasts (or computer monitors), you will be able to reproduce (or display) a complete picture without stuttering.

As described above, a conventional TV receiver (or computer monitor) is a prior art that allows high-definition television broadcasts having different screen sizes without being cut off, for example, from July 4, 1988 to Nippon Telegraph and Telephone Co., Ltd. There is a picture coding apparatus filed in the country of the United States and published in Japanese Patent No. 16887 on January 19, 1990.

The above-described Japanese Patent Laid-Open Patent Publication includes two encoding paths, one encoding path being low-pass filtered through a low pass filter on an input video signal to remove high frequency components, and then, again, one pixel every two pixels horizontally and horizontally through a sampling circuit. By generating a reduced video signal by extracting each other and compressing and encoding the differential reduced video signal generated by the motion compensation circuit, the video signal of the HDTV screen size is converted into a screen size suitable for a general TV screen, and then compressed and encoded by the encoded TV. And a second encoding path for performing discrete orthogonal transformation, quantization, and the differential signal between the original input image signal and the image signal reconstructed through the interpolation magnification provided from the one encoding path. Compressed HDTV video signal generation using variable length coding It is the path.

Accordingly, the Japanese Laid Open Patent allows adaptive generation of a compressed and encoded general TV image and an HDTV image through a configuration having two encoding paths, thereby obtaining a result capable of simultaneously accommodating HDTV broadcasting and general TV broadcasting.

However, as described above, the Japanese Laid-open Patent discloses that low-pass filtering is applied to each entire image to remove high frequency components of the image, which not only requires a large amount of computation, but also requires a large amount of frame memory during filtering. As a result, the hardware is complicated.

Accordingly, the present invention is to solve the above-mentioned problems of the prior art, it is possible to implement with a smaller amount of calculation and simple hardware by using decimation and interoperation, at least different sizes each corresponding to different receivers with different aspect ratios An object of the present invention is to provide a video encoding system capable of generating a plurality of compressed video signals.

In order to achieve the above object, the present invention provides a code having a plurality of codewords in a predetermined subblock unit for a difference signal between a current input frame and a prediction frame obtained through motion estimation and compensation between the current frame and a previous frame. In an image encoding system for performing vector quantization using a table, the current frame is divided into 8 × 8 DCT blocks, and then a time domain image signal for each of the divided DCT blocks is determined using a cosine function. A DCT block for converting to a DCT transform coefficient value in a frequency domain of 占 8; A first code table for a first difference signal obtained based on each DCT transform coefficient value of 8x8 of the converted current frame and the corresponding DCT transform coefficient value of corresponding 8x8 of the previous frame corresponding to the previous frame on a time axis; A first encoding path which is subjected to compression encoding by using the vector quantization and provides an encoded image signal having a specific screen size to a transmitter; And each DCT transform coefficient value of a low pass portion of a predetermined size extracted from each DCT transform coefficient value of 8 × 8 of the converted current frame, and each DCT transform coefficient value of a corresponding low pass portion of a predetermined size of a previous frame corresponding to that on the time axis. By applying vector quantization using a second code table to a second differential signal obtained based on the coded video signal, a coded video signal having a screen size of at least smaller than the compressed coded specific screen size output from the first encoding path is obtained. A second encoding path provided to the transmitter, wherein the first encoding path corresponds to an 8 × 8 DCT transform coefficient value of the current frame provided from the DCT block and a corresponding previous frame on the time axis; A first sense of generating the first differential signal through subtraction with a prediction signal obtained by using each DCT transform coefficient value of 8x8; It means; A first vector quantization for re-dividing the first differential signal from the first subtracting means into subblocks of a predetermined size, and then vector quantizing the divided subblocks using the first code table having a plurality of codewords block; Inverse vector quantization is performed on the vector quantized video signals by using a code table having the same codewords as the codewords of the first code table, and then restored to each DCT transform coefficient value of 8x8. First previous frame generating means for generating the previous frame existing on the time axis of the current frame by adding the respective DCT conversion coefficient values and the first prediction signal; And a first motion prediction block generating the first prediction signal through motion estimation and compensation between the current frame and the current frame and the previous frame, and providing the first prediction signal to the first subtraction means and the first previous frame generation means. The second encoding path may include: a decimation block for extracting each DCT transform coefficient value of a low-pass portion of a predetermined size from the DCT transform coefficient values of 8 × 8 provided from the DCT block; Each DCT transform coefficient value of 4x4 of the lower portion of each block of the current frame provided from the decimation block, and the corresponding 4x4 angle of the lower portion of each block of the previous frame corresponding thereto on the time axis. Second subtracting means for generating a second differential signal by subtracting from a second predictive signal obtained by using a DCT conversion coefficient value; A second vector quantization for re-dividing the second difference signal from the second subtracting means into subblocks of a predetermined size, and then vector quantizing the divided subblocks using the second code table having a plurality of codewords block; Inverse vector quantization is performed on the vector quantized video signals by using a code table having the same codewords as the codewords of the second code table, and then restored to the 4 × 4 DCT transform coefficient values of the low-pass portion. Second previous frame generating means for generating the previous frame existing on the time axis of the current frame by adding each DCT transform coefficient value of the restored low-pass portion and a second prediction signal; And a second motion prediction block generating the second prediction signal through motion estimation and compensation between the current frame and the current frame and the previous frame, and providing the second prediction signal to the second subtracting means and the second previous frame generating means. An adaptive video encoding system is provided.

The above and other objects and various advantages of the present invention will become more apparent from the preferred embodiments of the present invention described below with reference to the accompanying drawings by those skilled in the art.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

First, the adaptive video encoding system to be realized in the present invention has at least two coding paths, and the compression-coded video signal output from each coding path has different sizes (number of lines and number of pixels). That is, the image size compressed and output through the first encoding path indicated by reference numeral A in FIG. 1 is, for example, about 768 × 1024 suitable for a screen of a typical TV receiver (or computer monitor). The image size (number of lines and the number of pixels) which is compressed and output through the second encoding path indicated by the reference sign B in the drawing may be, for example, a small general TV receiver (that is, a desktop general TV receiver or a portable general TV receiver, etc.). It is about 384 × 512 suitable for the screen of a computer monitor.

On the other hand, the coding system of the present invention does not apply MC-DCT to the difference signal (or error signal) between the current frame continuous and the predicted previous frame on the time axis as in the conventional coding system. After applying DCT of 8, motion estimation and compensation are performed using DCT transform coefficients to generate differential signals for DCT transform coefficients.

To this end, the first encoding path A of the present invention is primarily a 8 × 8 DCT transform coefficient block of the current frame and a corresponding block of the previous frame predicted for motion estimation and compensation in the second encoding path B. The first differential signal is obtained by subtracting the interpolated blocks, i.e., the prediction blocks that interpolate the 4x4 blocks of the predicted previous frame into 8x8 blocks, and secondarily obtain the first differential signal obtained as described above. And the second difference signal is obtained by subtracting the predicted difference signal of the previous frame predicted for motion estimation and compensation corresponding to the first difference signal, and the obtained second difference signal is encoded by vector quantization and then transmitted to a transmitter not shown. Is sent. Here, the second encoding path of the present invention is to perform encoding through motion estimation and compensation on the low-frequency 4x4 DCT blocks obtained by decimating the DCT blocks.

Therefore, by applying the two differential signal generation processes as described above, the present invention can further reduce the bit generation amount. That is, the present invention will be able to obtain a very large compression rate.

Referring to FIG. 1, the DCT block 12 in the first encoding path A is divided into 8 x 8 DCT blocks for the current frame signal provided from the frame memory 10 on the input side, and then divided. The video signal (pixel data) in the time domain for each of the DCT blocks is converted into a DCT conversion coefficient in the frequency domain of 8x8 using a cosine function. Then, the DCT transform coefficients transformed in 8 × 8 block units are provided to the next subtracter 14 and the decimation block 32 through the line L11.

On the other hand, the subtractor 14 predicts a 8x8 prediction signal of 4x4 prediction signal of the DCT transform coefficient block provided from the DCT block 12 through the line L11 and the second encoding path B described later in detail. A first differential signal is generated on the line L15 by subtracting the corresponding side transform coefficient block provided through the line L13 from the interpolation block 16 which interworks with the signal and provided to the next subtractor 18. .

Accordingly, the subtractor 18 receives the first difference signal of the transform coefficient value for each DCT transform coefficient block of the current frame provided from the subtractor 14 through the line L15 and the motion prediction block 30 described later in detail. Each difference signal is generated by subtracting the transform coefficient values of the corresponding DCT transform coefficient block of the predicted previous frame provided through line L17.

Next, vector quantization (VQ) is performed at block 22 for each subblock (e.g., 4x4 or 2) obtained by subdividing the difference signal of each DCT coefficient of variation block provided from subtractor 18 described above. Vector quantization is performed through pattern matching using a code table (not shown) having a plurality of codewords for subblocks of × 2). That is, the vector quantization block 22 searches a codeword (reference pattern) having a value most similar to the input pattern of each subdivided subblock from the code table to determine a reference pattern, that is, the corresponding input pattern corresponding to the subblock. Then determine an optimal reference pattern with the smallest error value, then generate an index value for the determined reference pattern on line L19 and send it to the transmitter, not shown.

At this time, the index values generated on the line L19 are input to an inverse vector quantization (IVQ) block 24. In the inverse quantization block 24, motion estimation and compensation between frames to remove redundancy on the time axis of the video signal are performed. In order to use the code table (not shown) having the same codewords as those used in the above-described vector quantization block 22, the inverse vector is converted into the second difference signal (DCT transform coefficient value) before the vector quantization. The quantized and thus inverse vector quantized 8x8 secondary differential signals are provided to the adder 26 of the next stage.

The adder 26 then adds the reconstructed current frame signal (second order difference signal) provided from the inverse vector quantization block 24 and the predicted frame signal provided from the motion prediction block 30 via line L17. By providing the frame memory 28, the frame memory 28 is stored as a frame immediately before the restored current frame signal, that is, the current frame signal currently input for encoding. Therefore, through this process, the previous frame signal stored in the frame memory 28 is continuously updated with the image data immediately before the currently encoded input image data.

Accordingly, the motion prediction block 30 moves the motion to a predetermined search range of the previous frame stored in the frame memory 28, for example, 16 × 16, for each DCT block of the current frame signal on the line L15. Estimation, that is, determining the DCT block of the previous frame most similar to each DCT block of the current frame, that is, generating a plurality of motion vectors using a block matching algorithm, and using these motion vectors, predicting differential signals into blocks of the previous frame. (Prediction signal) is generated and generated on the line L17, and the predicted difference signal generated in this way is provided to the above-described subtractor 18 and adder 26, respectively. In addition, although not shown in FIG. 1, the motion vector detected by the motion prediction block 30 is encoded after a predetermined encoding process to be transmitted to the receiving side decoding system, and then sent to a transmitter (not shown).

As a result, the subtractor 18 subtracts the difference signal (transform coefficient values) provided from the subtractor 14 through the line L15 and the predicted difference signal provided from the motion prediction block 30 through the line L17. A second difference signal (a difference signal of transform coefficient values) is obtained, and the second difference signal thus obtained is compressed and encoded through the vector quantization block 22, and then output through line L19 for transmission to a transmitter not shown (OUTPUT1). Is provided. In this case, the compression-coded video signal (actual index values) output from the first encoding path A according to the present invention, as described above, is suitable for the screen of a conventional general TV receiver (or computer monitor). It is a video signal having a size of about 1024.

On the other hand, in the encoding system of the present invention, 384 suitable for a screen such as a small general TV receiver (or computer monitor), for example, a desktop general TV receiver (or computer monitor) or a portable general TV receiver (or computer monitor). The second encoding path B for generating a compressed coded video signal having a screen size of about 512 pixels is an 8 × 8 output from the DCT block 12 included in the first encoding path A described above as an input thereof. DCT transform coefficient values of the block are provided.

That is, in the decimation block 32, when each of the 8x8 DCT conversion coefficient values ((a) of FIG. 2) on the line L11 is input, as shown in (b) of FIG. Among the 8x8 DCT transform coefficient values, only the DC component of the low-pass portion, that is, 4x4 transform coefficient values, which are important information sensitive to human visual characteristics, is extracted and displayed on the line L22. In other words, the DCT conversion coefficient values of each 8x8 may be converted into 4x4 DCT conversion coefficient values having only a low pass portion through the decimation block 32. This decimation process can be considered to play a role of generating a video signal having a screen size suitable for a small general TV receiver (or computer monitor) that is intended to be obtained in the present invention.

Then, the 4 × 4 DCT transform coefficient values extracted from the decimation block 32 and generated on the line L22 are subtracted by the subtractor 34 for generating the difference signal with the predicted previous frame signal, and for motion estimation and compensation. Input to the motion prediction block 46, respectively.

In FIG. 1, the second encoding path B has an 8 × 8 DCT transform coefficient value so as to generate an encoded video signal having a screen size at least smaller than the output image of the first encoding path A. FIG. The only difference is that the motion estimation, compensation, and vector quantization are performed only on the 4 × 4 DCT transform coefficient values of the lower portion of each block, but in terms of the configuration, the adder 34, the vector quantization block 38, The functions of the inverse vector quantization block 40, the adder 42, the frame memory 44, and the motion prediction block 46 are included in the first encoding path A shown at the top of the figure, and correspond to each other. The adder 18, the vector quantization block 20, the inverse vector quantization block 24, the adder 26, the frame memory 28, and the motion prediction block 30 are components that perform substantially the same functions. .

Therefore, since each function of these constituent members has already been described in detail in the operation description of the first encoding path A described above, the detailed description thereof will be omitted.

On the other hand, the 4x4 prediction signal generated on the line L24 from the motion prediction block 46 is interpolated so that the second encoding path B interpolates the first differential signal generation in the first encoding path A. It may be different from that provided by the migration block 16. That is, the 4x4 prediction signal on the line L22 as shown in FIG. 2 (b) is converted into an 8x8 prediction signal through the interpolation block 16, as shown in FIG. The interpolated 8x8 prediction signal is provided to the subtractor 14 via the line L13 to generate a difference signal with the current frame.

In addition, since the difference signals provided from the subtractor 34 at the time of vector quantization in the vector quantization block 38 are DCT transform coefficients having a magnitude of 4x4, unlike in the above-described first encoding path A, In the vector quantization block 38, the size of the subblock (input pattern) that is subdivided for vector quantization may be 2 × 2.

Therefore, the compressed coded video signal OUTPUT2 outputted from the vector quantization block 38 in the second encoding path B according to the present invention to the transmitter not shown through line L26 is a desktop or portable general TV receiver ( Or a video monitor having a screen size of about 384 × 512, which is suitable for a compact TV receiver.

On the other hand, in the preferred embodiment of the present invention as described above has been described as quantizing the input image by employing a vector quantizer (vector quantization block 38) in the second encoding path (B), the present invention is not limited thereto. That is, it will be easily understood by those skilled in the art that the scalar quantizer may be applied in consideration of the amount of bits generated in the second encoding path B of the present invention. It may be necessary to add a conventional variable length coder for variable length coding of scalar quantized DCT transform coefficients and motion vectors generated during motion estimation and compensation.

As described above, the video encoding system according to the present invention employs a decimation technique and an interpolation technique for DCT transform coefficients in a hybrid encoding system using MC-DCT, thereby unnecessarily excessive computation and its configuration. It is possible to effectively generate a compressed coded video signal having at least different screen sizes (number of lines and number of pixels) that can meet the needs of the channel without the complexity of the image.

Claims (6)

Image coding for performing vector quantization using a code table having a plurality of codewords in a predetermined subblock unit for a difference signal between a current input frame and a prediction frame obtained through motion estimation and compensation between the current frame and a previous frame. In the system, after dividing the current frame into 8 × 8 DCT blocks, the DCT transform coefficient value of the 8 × 8 frequency domain by using a cosine function for the time-domain video signal for each of the divided DCT blocks. A DCT block for converting the data into a transform; A first code table for a first difference signal obtained based on each DCT transform coefficient value of 8x8 of the converted current frame and the corresponding DCT transform coefficient value of corresponding 8x8 of the previous frame corresponding to the previous frame on a time axis; A first encoding path which is subjected to compression encoding by using the vector quantization and provides an encoded image signal having a specific screen size to a transmitter; And each DCT transform coefficient value of a low pass portion of a predetermined size extracted from each DCT transform coefficient value of 8 × 8 of the converted current frame, and each DCT transform coefficient value of a corresponding low pass portion of a predetermined size of a previous frame corresponding to that on the time axis. By applying vector quantization using a second code table to a second differential signal obtained based on the coded video signal, a coded video signal having a screen size of at least smaller than the compressed coded specific screen size output from the first encoding path is obtained. A second encoding path provided to the transmitter, wherein the first encoding path corresponds to an 8 × 8 DCT transform coefficient value of the current frame provided from the DCT block and a corresponding previous frame on the time axis; A first sense of generating the first differential signal through subtraction with a prediction signal obtained by using each DCT transform coefficient value of 8x8; It means; A first vector quantization for re-dividing the first differential signal from the first subtracting means into subblocks of a predetermined size, and then vector quantizing the divided subblocks using the first code table having a plurality of codewords block; Inverse vector quantization is performed on the vector quantized video signals by using a code table having the same codewords as the codewords of the first code table, and then restored to each DCT transform coefficient value of 8x8. First previous frame generating means for generating the previous frame existing on the time axis of the current frame by adding the respective DCT conversion coefficient values and the first prediction signal; And a first motion prediction block generating the first prediction signal through motion estimation and compensation between the current frame and the current frame and the previous frame, and providing the first prediction signal to the first subtraction means and the first previous frame generation means. The second encoding path may include: a decimation block for extracting each DCT transform coefficient value of a low-pass portion of a predetermined size from the DCT transform coefficient values of 8 × 8 provided from the DCT block; Each DCT transform coefficient value of 4x4 of the lower portion of each block of the current frame provided from the decimation block, and the corresponding 4x4 angle of the lower portion of each block of the previous frame corresponding thereto on the time axis. Second subtracting means for generating a second differential signal by subtracting from a second predictive signal obtained by using a DCT conversion coefficient value; A second vector quantization for re-dividing the second difference signal from the second subtracting means into subblocks of a predetermined size, and then vector quantizing the divided subblocks using the second code table having a plurality of codewords block; Inverse vector quantization is performed on the vector quantized video signals by using a code table having the same codewords as the codewords of the second code table, and then restored to the 4 × 4 DCT transform coefficient values of the low-pass portion. Second previous frame generating means for generating the previous frame existing on the time axis of the current frame by adding each DCT transform coefficient value of the restored low-pass portion and a second prediction signal; And a second motion prediction block generating the second prediction signal through motion estimation and compensation between the current frame and the current frame and the previous frame, and providing the second prediction signal to the second subtracting means and the second previous frame generating means. Adaptive video encoding system, characterized in that. The method of claim 1, wherein the first encoding path interpolates the 4x4 second prediction signal provided from the second motion prediction block and restores each DCT transform coefficient value of 8x8 of a previous frame. Interporation block; And a third difference signal generated by subtracting each DCT transform coefficient value of 8x8 of the current frame provided from the DCT block with each DCT transform coefficient value of 8x8 of the interpolated previous frame. And subtracting means, and providing the generated difference signal to the first subtractor as each DCT transform coefficient value of 8x8 of the current frame. The adaptive video encoding system according to claim 1 or 2, wherein the first vector quantization block re-divides the 8x8 DCT transform coefficients into four 4x4 subblocks. The adaptive video encoding system according to claim 1 or 2, wherein the first vector quantization block repartitions each of the 8x8 DCT transform coefficients into 16 2x2 subblocks. The adaptive video encoding system of claim 1, wherein a size of each DCT transform coefficient value of the low-pass portion extracted from each DCT transform coefficient value of 8x8 of the current frame is 4x4. 6. The adaptive video encoding system according to claim 1 or 5, wherein the second vector quantization block repartitions the 4x4 DCT transform coefficients into four 2x2 subblocks.