KR20080007060A - Video encoding and decoding apparatus and methods using separation of amplitude and sign of a differential image signal - Google Patents

Abstract

An apparatus and a method for encoding/decoding a video by using separation of amplitude and sign of a differential signal are provided to take an absolute value for a differential signal generated after motion prediction, thereby improving the correlation of pixels within a block where DCT(Discrete Cosine Transform) will be performed and accordingly increasing compression efficiency. A video prediction unit generates a prediction video with respect to an original video as an object for current encoding. A differential signal generating unit generates a differential signal between the original video and the prediction video. An amplitude/sign separating unit generates a sign bit(sign bit for compensation) for performing restoring of the differential signal and amplitude from the differential signal. An encoding unit performs transform encoding, quantization and entropy encoding with respect to the differential signal amplitude, and performs entropy encoding of the sign bit for compensation. A decoding unit performs inverse-transform of the differential signal amplitude, transform-encoded and quantized in the encoding unit, so as to perform restore into the original differential signal. By performing sign compensation by using the sign bit for compensation and performing motion compensation for the sign-compensated differential signal, the decoding unit restores the original video and stores the restored original video as a new reference video.

Description

Video Encoding / Decoding Apparatus and Method Using Separation of Code and Size of Differential Signal Separation of Amplitude and Sign of a differential image signal

1 is a configuration diagram of a first embodiment of a video encoding apparatus using separation of magnitude and code of a difference signal according to the present invention;

2 is an explanatory diagram of a scanning sequence of codes used for code transmission in the present invention;

3 is a diagram illustrating a code transmission method using a correlation between codes in a code transmission unit of FIG. 1 according to the present invention;

4 is a configuration diagram of a second embodiment of a video encoding apparatus using separation of magnitude and code of a difference signal according to the present invention;

5 is a configuration diagram of a first embodiment of a video decoding apparatus for the encoding apparatus of FIG. 1 according to the present invention;

6 is a configuration diagram of a second embodiment of a video decoding apparatus for the encoding apparatus of FIG. 4 according to the present invention.

* Explanation of symbols on the main parts of the drawing

10: image predicting unit 11: subtractor (difference signal generator)

12, 40: size / sign separator 13: encoder

14: decoder 42: mode determiner

44: mode selector 50, 60: decoder

51, 62: code compensator 52, 63: video restoring unit

61: mode selector

The present invention relates to a video encoding / decoding apparatus and a method thereof, and more particularly, to video encoding / decoding using separation of a magnitude and a code of a differential signal, which can improve the efficiency of video encoding by improving the performance of transform encoding. An apparatus and a method thereof are provided.

That is, the present invention separates the difference signal (the difference between the original video to be encoded currently and the predicted video for the original video) into a magnitude and a sign, performs transform encoding and quantization only on the magnitude signal, and receives the sign information. The present invention relates to a video encoding / decoding apparatus using a separation of a magnitude and a code of a differential signal, which can be transmitted separately for use in decoding, thereby improving the efficiency of video encoding by improving the performance of transform encoding.

Many standards, such as H.261, H.263, MPEG-1, MPEG-2, MPEG-4, and H.264, have been studied by ISO / IEC MPEG and ITU-T's VCEG, the standardization bodies for video compression. Is developed. Although these standards differ slightly, they generally consist of structures that include motion estimation and compensation, transform coding, and entropy encoding.

In particular, in relation to transform coding, each standard proposes a variety of block sizes, but employs many Discrete Cosine Transform (DCT) schemes in common. Pixel values for representing an image in a video or pixel values in a differential image generated after motion prediction for image compression have a real value. In the video image, spatially and temporally adjacent pixels have similar characteristics. It is known that the basis functions of the discrete cosine transform have a good performance for a given block of an image having a real value and a high correlation.

Therefore, many conventional video compression standards use discrete cosine conversion to obtain a high signal compression effect, thereby increasing the compression rate.

However, in the conventional video compression standard, since the differential signal obtained after performing the motion prediction has both positive and negative values, the signal concentration is low so that the basis functions of a large number of discrete cosine transforms should be used, and the coding after quantization is performed. Since the number of transform coefficients to be present is also significant, there is a problem in that the compression efficiency is lowered.

The present invention has been proposed to solve the above problems, and provides an apparatus and method for encoding / decoding a video using a difference signal size and code separation, which can improve video encoding efficiency through performance of transform encoding. Its purpose is to.

That is, according to the present invention, by taking an absolute value on the difference signal generated after the motion prediction, the video using the size of the difference signal and the separation of the code increases the correlation between the pixels in the block to be transformed and increases the compression efficiency. An object of the present invention is to provide an encoding device and a method thereof.

In addition, the present invention restores the original image by compensating the code of the difference signal that is damaged by taking an absolute value of the difference signal before encoding, as a result of performing inverse transform encoding or the like to decode the video signal. Another object of the present invention is to provide a video decoding apparatus and a method using separation of a magnitude and a code of a difference signal.

Other objects and advantages of the present invention can be understood by the following description, and will be more clearly understood by the embodiments of the present invention. It will also be appreciated that the objects and advantages of the present invention may be realized by the means and combinations thereof indicated in the claims.

 According to an aspect of the present invention, there is provided a video encoding apparatus using separation of a magnitude and a code of a difference signal, comprising: image prediction means for generating a predicted image with respect to an original image currently to be encoded; Difference signal generation means for generating a difference signal between the original image and the prediction image; Magnitude / code separation means for generating a magnitude and a code bit (compensation code bit) required for reconstruction of the difference signal from the difference signal; Encoding means for performing transcoding, quantization, and entropy encoding on the difference signal magnitude, and the compensation code bits for entropy encoding; And inversely converting the transform-coded and quantized differential signal magnitudes into the original differential signal, and performing code compensation using the compensation code bits, and performing motion compensation on the coded differential signal to restore the original image. The decoding means for storing as a new reference image.

On the other hand, the present invention provides a video encoding apparatus using separation of magnitude and code of a differential signal, comprising: image predicting means for generating a predicted image of an original image currently to be encoded; Difference signal generation means for generating a difference signal between the original image and the prediction image; A coding mode for the difference signal is determined on the basis of the compression efficiency, and in the normal mode, the difference signal is transmitted to an encoding means. In the separation mode, a code necessary for reconstructing the magnitude and the difference signal from the difference signal is determined. Size / code separation means for generating a bit (compensation code bit) and transmitting the difference signal magnitude to the encoding means; The encoding means for performing transform encoding, quantization, and entropy encoding on the difference signal or the difference signal magnitude in accordance with an encoding mode, and performing entropy encoding on the compensation code bits in the separation mode; And inversely converting the transform-coded and quantized signal by the encoding means to restore the original differential signal, and then compensate for the motion of the restored differential signal to reconstruct the original image and store it as a new reference image. And decoding means for reconstructing the original image after compensating the code of the reconstructed differential signal using the compensation code bit.

On the other hand, the present invention provides a video decoding apparatus using separation of magnitude and code of a differential signal, comprising: decoding means for entropy decoding an input bit string, inverse quantization, and inverse transform encoding to restore an original difference signal; Code compensating means for compensating a code of the recovered differential signal using the corresponding compensating code bit; And image reconstruction means for reconstructing the original image by performing motion compensation on the coded difference signal.

On the other hand, the present invention provides a video decoding apparatus using separation of magnitude and code of a differential signal, comprising: decoding means for entropy decoding an input bit string, inverse quantization, and inverse transform encoding to restore an original difference signal; Mode selection for determining whether to compensate the code according to the encoding mode information, to deliver the reconstructed difference signal to the image recovery means in the normal mode, and to transmit the reconstructed difference signal to the code compensation means in the separated mode Way; The code compensating means for compensating the code of the difference signal transmitted from the mode selecting means by using the compensating code bit; And the image restoring means for restoring the original image by compensating for motion of the coded difference signal or the difference signal transmitted from the mode selection means.

Meanwhile, the present invention provides a video encoding method using separation of magnitude and code of a difference signal, comprising: an image prediction step of generating a prediction image for an original image currently being encoded; Generating a difference signal between the original image and the prediction image; A magnitude / code separation step of generating a magnitude and a code bit (compensation code bit) required to recover the difference signal from the difference signal; An encoding step of transform encoding, quantization, and entropy encoding on the magnitude of the difference signal, and entropy encoding on the compensation code bits; And reconstructing the original differential signal by inversely transforming the transform-coded and quantized differential signal magnitudes in the encoding step, and performing code compensation using the compensation code bits, and restoring the original image by motion compensation to the coded differential signal. And storing the data as a new reference image.

On the other hand, the present invention, a video encoding method using the separation of the magnitude and the code of the difference signal, the video prediction step of generating a prediction image for the original image to be encoded; Generating a difference signal between the original image and the prediction image; A size / code separation step of determining an encoding mode for the difference signal based on a compression efficiency, and generating a magnitude and a code bit (compensation code bit) required to recover the difference signal from the difference signal; A normal mode encoding step of performing transform encoding, quantization, and entropy encoding on the difference signal if the encoding mode is a normal mode; In the normal mode, in the normal mode encoding step, the transform-coded and quantized signals are inversely transformed to be restored to the differential signal, and then the normal mode decoding is performed to restore the original image and store the new reference image by motion compensation to the restored differential signal. step; And a separation mode encoding step of performing transform encoding, quantization, and entropy encoding on the difference signal size and entropy encoding on the compensation code bits when the encoding mode is a separation mode. In the split mode, the transform coded and quantized signals are inversely transformed in the split mode encoding step to be restored to the original differential signal, and then the code is compensated using the corresponding compensation code bits. And a split mode decoding step of reconstructing the original image by performing motion compensation and storing the original image as a reference image.

On the other hand, the present invention, a video decoding method using the separation of the magnitude and code of the differential signal, Entropy decoding the input bit stream, inverse quantization and inverse transform encoding to restore the original differential signal; A code compensating step of compensating a code of the restored differential signal by using a corresponding compensation code bit; And an image restoration step of restoring the original image by performing motion compensation on the coded difference signal.

On the other hand, the present invention, a video decoding method using the separation of the magnitude and code of the differential signal, Entropy decoding the input bit stream, inverse quantization and inverse transform encoding to restore the original differential signal; A code compensating step of verifying encoding mode information and compensating a code of a difference signal reconstructed in the decoding step by using a corresponding compensation code bit; And image reconstruction means for reconstructing the original image by performing motion compensation on the differential signal reconstructed in the decoding step or the differential signal reconstructed in the code compensation step.

The present invention relates to transform coding, which is one of the core techniques of video compression, and takes an absolute value for a predetermined block and increases spatial redundancy (that is, an absolute value for a predetermined block) before performing a discrete cosine transform during video encoding. By taking a spatial redundancy and then performing discrete cosine transforms) to further concentrate the signal and thereby increase the compression efficiency.

In order to decode the video signal in the decoding apparatus according to the present invention, the result of performing the inverse discrete cosine transform is to compensate for the 'sign of the difference signal damaged when the absolute value of the difference signal is taken before encoding'. Obtain Therefore, in the present invention, the code of the original differential signal is separately coded and transmitted.

The above objects, features and advantages will become more apparent from the following detailed description taken in conjunction with the accompanying drawings, whereby those skilled in the art may easily implement the technical idea of the present invention. There will be. In addition, in describing the present invention, when it is determined that the detailed description of the known technology related to the present invention may unnecessarily obscure the gist of the present invention, the detailed description thereof will be omitted. Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a configuration diagram of a first embodiment of a video encoding apparatus using separation of magnitude and code of a difference signal according to the present invention.

As shown in FIG. 1, the video encoding apparatus according to the present invention includes an image predictor 10, a subtractor (difference signal generator) 11 that generates a difference signal, and a size that separates the magnitude and code of the difference signal / A code separator 12, an encoder 13, and a decoder 14 are included.

The image predictor 10 includes a motion predictor (ME) 101, a motion compensator (MC) 102, an intra prediction selector 103, and an intra predictor 104. The prediction image is generated with respect to the original image. Then, the subtractor 11 subtracts the predicted image generated by the image predictor 10 from the original image to be currently encoded to generate a difference signal.

That is, the image predictor 10 performs motion estimation or estimation to encode one block of a predetermined size in the current frame F _n to be encoded, and the subtractor 11 performs the difference signal D for the current block. _n )

The magnitude / code separator 12 separates the difference signal D _n into magnitude and sign, and includes an absolute value calculator 121 and a code transmitter 122. Here, the absolute value calculator 121 calculates the magnitude of the difference signal by performing an absolute value operation on the difference signal D _n . That is, the absolute value calculator 121 takes an absolute value with respect to the difference signal in order to increase the correlation between the pixels in the block. The code transmitter 122 extracts a sign of the difference signal. That is, the code transmitting unit 122 generates a code bit necessary for restoring the differential signal from the difference signal (this is a compensation code bit, hereinafter simply referred to as a 'code bit').

The encoding unit 13 includes a discrete cosine transform unit (DCT) 131, a quantization unit 132, a reordering unit 133, and an entropy encoder 134, and the difference output from the absolute value calculating unit 121. The signal magnitude is encoded by the discrete cosine transform 131, the quantization 132, the reordering 133, and the entropy encoding process 134, and the difference signal code bit transmitted from the code transmitter 122 is obtained. Only entropy encoding 134 is performed. The encoder 13 entropy encodes 134 the bit bits of the original difference signal D _n before taking an absolute value and transmits the encoded bit to the receiving side so that the receiving side decoding apparatus can use it to reconstruct the image.

)(142), 부호 보상부(143), 가산기(144), 필터(145)를 포함하여 이루어지는 것으로서, 다음 블록 또는 다음 프레임을 부호화하기 위해서 현재 블록 또는 현재 프레임을 부호화하는 과정에서도 원래 블록의 부호를 보상하여 영상을 복원하는 것이다.Meanwhile, the decoding unit 14 includes an inverse quantization unit (Q ⁻¹ ) 141 and an inverse discrete cosine transform unit (

142, the sign compensator 143, the adder 144, and the filter 145, and the sign of the original block even when the current block or the current frame is encoded to encode the next block or the next frame. To restore the image.

In more detail, the decoder 14 applies the inverse discrete cosine transform and inverse quantization process to the discrete cosine transform and quantized differential signal magnitude (output of the quantizer 132) in the encoder 13. After restoring the magnitude of the difference signal, the original difference signal is restored using the code bit (compensation code bit) transmitted from the code transmitter 122. Then, the decoder 14 restores the original image (the original image) by performing motion compensation on the reconstructed differential signal (difference image) D _n ′, and then converts it to a new reference image for encoding the next image. Save it.

Hereinafter, the code transmitter 122 will be described in more detail (see FIGS. 2 and 3).

Code transmission in the code transmission unit 122 may be transmitted in various ways as follows according to the purpose. The code transmitter 122 may transmit all of the codes of the differential signal D _n generated during the encoding process.

On the contrary, the code transmitter 122 transmits the corresponding code only to all values whose pixel values of the differential signal block reconstructed through the "141" and "142" of the differential signal D _n are not "0". It may be. Because, the signal reconstructed after going through the discrete cosine transform (T) 131 and the quantization process (Q) 132 may have less differential signal than the original image, and the signal used for reconstruction in the decoding apparatus is actually This is because the restored differential signal obtained by performing inverse quantization and inverse discrete cosine transform.

In other words, the code transmitter 122 transmits the code of the difference signal D _n , but not all of them, but the pixel value of the differential signal block restored through "141" and "142" is "0". Only the code of the position corresponding to the pixel is transmitted. In other words, with respect to coefficient values other than "0" of the differential signal restored through "141" and "142", the code (original code of the original differential signal) is transmitted in a predetermined order by referring to the code of the original differential signal. .

Originally, since the encoder 13 performs discrete cosine transform only on positive values that take an absolute value on the differential signal, the pixel value should only have a positive value when it is restored through the inverse discrete cosine transforms 141 and 142. . Therefore, if the pixel value of the difference signal reconstructed through "141" and "142" is a negative value, an incorrect result is obtained. In this case, the value is set to "0" in order to reduce an error of an image. Therefore, the sign for the pixel whose pixel value is negative will not be transmitted.

2 is an explanatory diagram of a scanning sequence of codes used for code transmission in the present invention.

Unlike the frequency axis, the spatial axis representing the differential image (differential signal) has no meaning or weight according to its position with respect to the pixel. In one embodiment of the present invention, the code of the coefficient to be sent to the decoding apparatus (decoder) is transmitted through sequential scan.

The video encoding apparatus according to the present invention performs entropy encoding (134) on Context Adaptive Binary Arithmetic Coding (CABAC) used in the H.264 standard for sequential scans. Transmit to the decoding device (decoder) side. In the present invention, sequential scanning is used. However, if the scanning order is applied in the same manner in the encoding apparatus (encoder) and the decoding apparatus (decoder), the zigzag scanning or the other direction scanning sequence used in H.264 may be used.

Meanwhile, in the present invention, transmitting the code bits for all the pixels other than "0" generates a lot of additional bits, and thus, the code bits may be adaptively transmitted according to network conditions or transmission bandwidths. In another embodiment, pixels having a value of 1 or 2 among the pixel values of the difference signal reconstructed through "141" and "142" mean an error 1 or 2 in the reconstructed image. Since these pixel values have a relatively small influence on the image quality of the image compared with other large pixel values, these values may be replaced with "0" and may not transmit a code when a small number of bit bits is required. have. That is, the code bit may be transmitted only for pixels whose pixel value is greater than or equal to a threshold.

In the case of code transmission using a threshold, the value below the same threshold (threshold) is discarded for the difference signal reconstructed by the decoding apparatus (decoder) without any additional information, and only the remaining coefficients are encoded. Code compensation is performed using the pixel value sent from the device (encoder).

In Equation 1 below, R (x, y) is the magnitude of the recovered differential signal, and Tm Is a threshold value. The code bit to be transmitted to the decoding apparatus (decoder) can be reduced by changing the threshold (threshold) value within a range that does not significantly degrade the quality of the image.

In other words, the code transmitter 122 sequentially scans the blocks of the difference signals reconstructed through the "141" and "142" by the decoder 14, and the difference corresponding to the position of the pixel whose pixel value is not "0". The code bit of the signal may be generated as the compensation code bit and transmitted, or the code bit of the differential signal corresponding to the position of the pixel whose pixel value is greater than or equal to a predetermined threshold among pixels whose pixel value is not "0" may be used as the compensation code bit. You can also create and send.

3 is a diagram illustrating an embodiment of a code transmission method using a correlation between codes in the code transmission unit of FIG. 1 according to the present invention, a code transmission method described with reference to FIG. 2 (a method of transmitting codes in a scanning order). Different way.

Unlike the sequential scanning of FIG. 2, in FIG. 3, a code is transmitted by using a correlation between codes of a difference signal (a difference signal reconstructed through “141” and “142”) to be transmitted. Pixels having a large value in the reconstructed differential signal block mean that a difference between the prediction image and the image currently being encoded is large.

In general, there is a high probability that a difference with a predicted value greatly occurs at the boundary of an image, and the difference values (differential signals) are likely to have similar signs. Accordingly, the code transmitter 122 sorts the pixel values of the reconstructed differential signal (differential image) in descending order of the magnitude of the absolute value, and the code of the aligned pixel values is taken from the original differential signal and arranged. do. The codes listed in this way are entropy encoded and transmitted to the video decoding apparatus at the receiving end.

As an example, if the trailing_sign_flag value is "1", it means that at least N (successive_sign_count = N) signs of the difference signal (the restored difference signal) are successively displayed. Only the sign (successive_sign_flag) is transmitted, and the remaining codes are sent by entropy encoding in the sorted order (remain_sign_coding ()).

On the contrary, if the trailing_sign_flag value is "0", it means that there are less than N consecutive codes of the difference signal. In this case, all the codes arranged in descending order are entropy coded and transmitted (sign_coding ()).

If the value of N is large, the coding efficiency is high but the occurrence probability is relatively decreased. If the value of N is small, the coding efficiency is low but the probability of occurrence is increased. If a predefined value is used between the decoding device (decoder) and the encoding device (encoder) as a value of N that defines the number of consecutive codes, the process of separately encoding and transmitting to the decoding device (decoder) may not be performed. Alternatively, efficient code transmission may be performed by transmitting variable N according to the characteristics of the video.

In summary, the code transmitting unit 122 includes the code bits of the differential signal corresponding to the position of the pixel whose pixel value of the differential signal restored by the decoding unit 14 through "141" and "142" is not "0", Alternatively, the code bit of the difference signal corresponding to the position of the pixel whose pixel value is greater than or equal to a predetermined threshold is transmitted to the entropy encoder 134 and the code compensator 143, or as described with reference to FIG. Instead of transmitting all of the corresponding codes, only the number and codes of the corresponding codes may be transmitted.

FIG. 4 is a configuration diagram of a video encoding apparatus according to a second embodiment using separation of magnitude and code of a difference signal according to the present invention, in which a discrete cosine transform is performed on an absolute value of a difference signal (separation mode) (see FIG. 1). And a video encoding apparatus for selectively using an existing encoding scheme (normal mode).

The encoding apparatus (FIG. 4) selectively using the coding scheme is the same as the encoding apparatus of FIG. 1, except that the size / code separator 40 includes an absolute value calculator 41, a mode determiner 42, and There is a difference in that it includes the code transmitter 43, and there is a difference even if the mode selector 44 is further included. Hereinafter, only the configuration means having a difference will be described.

The encoding scheme performed in the video encoding apparatus as shown in FIG. 1, that is, the discrete cosine transform process on the absolute value of the differential signal has a higher compression efficiency than the conventional method, and thus has an advantage in entropy encoding. However, in this case, there is a limitation that an additional code bit must always be transmitted to the decoding device (decoder).

In the video encoding apparatus illustrated in FIG. 4, a coding scheme using a discrete cosine transform having an absolute value and a method using a conventional discrete cosine transform method are selectively used. If all or most of the pixel values of the difference signal generated after the motion prediction or estimation have the same sign before taking the absolute value, there is no or no improvement in compression through the discrete cosine transform taking the absolute value. Therefore, since unnecessary code bits are added, the compression efficiency may be lowered as a result. Accordingly, in the present invention, a block having a code having the same code may be encoded without transmitting a code bit by applying a conventional method as it is.

The mode determining unit 42 takes the absolute value of the differential signal and then performs a discrete cosine transform or the like (separation mode) and an existing method (the method of performing the discrete cosine transform without separating the magnitude and the sign, that is, the normal mode). ) Is determined by the compression efficiency of the method. That is, the mode determiner 42 determines the 'coding mode' based on the compression efficiency using the magnitude of the difference signal and the difference signal output from the absolute value calculator 41. Here, the determination of the compression efficiency is determined by a bit-distortion optimization (RDO) method commonly used in mode selection in H.264.

In the case of using a discrete cosine transform whose absolute value is selectively used in the encoding apparatus according to the present invention, the fact of use should be reported to the video decoding apparatus (decoder). In other words, the code transmitter 43 transmits one bit called a mode flag to the decoder together with the code bit. If the mode flag value is "0", it indicates that the discrete cosine transform is performed without separating the magnitude and sign of the existing difference signal. If the mode flag value is "1", the absolute value is taken to the difference signal and then discrete. Performs a cosine transform and indicates that the sign is transmitted separately.

Meanwhile, the encoding apparatus performs a process of decoding (decoding) the current block or the current frame in order to encode the next block or the next frame. In this decoding process, the differential coarse method is selectively used according to the mode flag value. Restore That is, the mode selector 44 determines whether to compensate the code according to the mode flag received from the code transmitter 43, and if the mode flag is "1", uses the code bits transmitted together with the mode flag. To recover the differential signal.

5 is a configuration diagram of a first embodiment of a video decoding apparatus for the encoding apparatus of FIG. 1 according to the present invention, and includes a decoder 50, a code compensator 51, and an image reconstruction unit 52. Is done. That is, the decoding apparatus of FIG. 5 is a decoding apparatus for the encoding apparatus (FIG. 1) which has performed discrete cosine transform or the like on the absolute value of the differential signal.

The decoding unit 50 includes an entropy decoder 501, a reordering unit 502, an inverse quantization unit 503, and an inverse discrete cosine transform unit 504, and includes a bit string received from an encoding device (FIG. 1). The decoding process 50 performs entropy decoding 501, data reordering 502, inverse quantization 503, and inverse discrete cosine transform 504, like the existing video compression standards.

Then, the code compensator 51 restores the signed difference signal D _n ′ with the magnitude and the signed using the 'sign bit of the difference signal' transmitted by the encoding device (encoder) (FIG. 1). That is, the code compensator 51 compensates for pixels other than "0" in the output value of the decoder 50 by using the codes received from the encoder (encoder) (FIG. 1).

The image reconstructor 52 includes a motion compensator (MC) 521, an intra predictor 522, an adder 523, and a filter 524, and is reconstructed by the code compensator 51. The original image is reconstructed using the difference signal D _n ′ and the reference image (previous image).

6 is a configuration diagram of a video decoding apparatus according to a second embodiment of the encoding apparatus of FIG. 4 according to the present invention, and includes a decoder 60, a mode selector 61, a code compensator 62, and an image reconstruction. It comprises a portion 63. That is, the decoding apparatus of FIG. 6 is a decoding apparatus for the encoding apparatus of FIG. 4.

The decoder 60, like the decoder 50 of FIG. 5, entropy decoding, data reordering, inverse quantization, inverse discrete cosine transform, and the like, of the bit stream received from the encoding apparatus (FIG. 4) as in the existing video compression standards. Performs a decoding process.

Then, the mode selector 61 determines whether to compensate the code according to the mode flag received from the encoding apparatus of FIG. 4, and accordingly determines the output of the decoder 60 to the image reconstruction unit 63 or the code compensation. It passes to any of the parts 62.

That is, if the mode flag value is "0", the mode selector 61 determines that the image is encoded according to the conventional scheme, and reconstructs the image only through motion compensation. If the mode flag value is "1", the discrete cosine using the absolute value is used. The code compensator 62 compensates the code of the differential signal by determining that the data has been encoded according to the conversion scheme.

As described above, the method of the present invention may be implemented as a program and stored in a recording medium (CD-ROM, RAM, ROM, floppy disk, hard disk, magneto-optical disk, etc.) in a computer-readable form. Since this process can be easily implemented by those skilled in the art will not be described in more detail.

The present invention described above is capable of various substitutions, modifications, and changes without departing from the technical spirit of the present invention for those skilled in the art to which the present invention pertains. It is not limited by the drawings.

As described above, the present invention, unlike the existing video compression standard, in which a difference signal obtained after performing motion prediction has both positive and negative values, generates a new differential signal by making all the differential signals positive and generates the new differential signal. By performing the discrete cosine transform on the differential signal, the compaction of the signal is increased to use the basis function of the smaller number of discrete cosine transforms, and the number of transform coefficients to be encoded after quantization is also significantly reduced. There is an effect of increasing the compression efficiency.

That is, the present invention has an effect of increasing the compression efficiency by increasing the correlation of the pixels in the block to be subjected to the discrete cosine transform by taking the absolute value to the difference signal generated after the motion prediction.

Claims (29)

In a video encoding apparatus using separation of magnitude and code of a difference signal, Image prediction means for generating a predicted image with respect to the original image currently to be encoded; Difference signal generation means for generating a difference signal between the original image and the prediction image; Magnitude / code separation means for generating a magnitude and a code bit (compensation code bit) required for reconstruction of the difference signal from the difference signal; Encoding means for performing transcoding, quantization, and entropy encoding on the difference signal magnitude, and the compensation code bits for entropy encoding; And The encoding means inversely transforms the transform-coded and quantized differential signal to restore the original differential signal, and then performs code compensation using the compensation code bits, and restores the original image by motion compensation to the coded differential signal. The decoding means for storing as a new reference image Video encoding apparatus comprising a. The method of claim 1, The size / code separation means, Absolute value calculation means for obtaining an magnitude of the difference signal by taking an absolute value to the difference signal; And Code transmission means for generating said compensation code bit from said difference signal and transmitting it to said encoding means and said decoding means Video encoding apparatus comprising a. In a video encoding apparatus using separation of magnitude and code of a difference signal, Image prediction means for generating a predicted image with respect to the original image currently to be encoded; Difference signal generation means for generating a difference signal between the original image and the prediction image; A coding mode for the difference signal is determined on the basis of the compression efficiency, and in the normal mode, the difference signal is transmitted to an encoding means. In the separation mode, a code necessary for reconstructing the magnitude and the difference signal from the difference signal is determined. Size / code separation means for generating a bit (compensation code bit) and transmitting the difference signal magnitude to the encoding means; The encoding means for performing transform encoding, quantization, and entropy encoding on the difference signal or the difference signal magnitude in accordance with an encoding mode, and performing entropy encoding on the compensation code bits in the separation mode; And The encoding means inversely transforms the transform-coded and quantized signal to restore the original differential signal, and then restores the original image as a new reference image by performing motion compensation on the restored differential signal. Decoding means for restoring the original image after compensating the code of the restored differential signal using Video encoding apparatus comprising a. The method of claim 3, wherein The size / code separation means, Absolute value calculation means for obtaining an magnitude of the difference signal by taking an absolute value to the difference signal; And Mode determination means for determining an encoding mode on the basis of a compression efficiency, for transmitting the difference signal to the encoding means in the normal mode and for transmitting the magnitude of the difference signal to the encoding means in the separation mode; And In the separation mode, a magnitude and the compensation code bit are generated from the difference signal, and the difference signal magnitude is transmitted to the encoding means, and the compensation code bit is transmitted together with the encoding mode. Code transmission means for transmission by means of Video encoding apparatus comprising a. The method of claim 4, wherein The mode determining means, And a coding mode is determined based on a compression efficiency determined according to a rate-distortion optimization (RDO) method. The method of claim 4, wherein The encoding mode information, A video encoding device, characterized in that the mode flag of one bit. The method according to claim 2 or 4, The code transmission means, And the code bit of the difference signal corresponding to the position of the pixel whose pixel value of the difference signal reconstructed through the inverse transform by the decoding means is a compensation code bit. The method according to claim 2 or 4, The code transmission means, And the code bit of the difference signal corresponding to the position of the pixel whose pixel value of the difference signal reconstructed through the inverse transformation by the decoding means is the compensation code bit. The method according to claim 2 or 4, The code transmission means, And the compensation code bit is generated by using the correlation between the codes of the difference signal reconstructed by the inverse transform in the decoding unit. The method of claim 9, The code transmission means, In the decoding means, the pixel values of the differential signal reconstructed through the inverse transformation are arranged in descending order of the magnitude of the absolute value, and the original code for the aligned pixel values is transmitted in the sorting order as the compensation code bits. A video encoding device, characterized in that for transmitting consecutive codes, transmits only the number of codes and the code. In a video decoding apparatus using separation of magnitude and code of a difference signal, Decoding means for entropy decoding the input bit stream and inverse quantizing and inverse transform encoding to restore the original differential signal; Code compensating means for compensating a code of the recovered differential signal using the corresponding compensating code bit; And Image restoring means for restoring the original image by performing motion compensation on the coded difference signal Video decoding apparatus comprising a. In a video decoding apparatus using separation of magnitude and code of a difference signal, Decoding means for entropy decoding the input bit stream and inverse quantizing and inverse transform encoding to restore the original differential signal; Mode selection for determining whether to compensate the code according to the encoding mode information, to deliver the reconstructed difference signal to the image recovery means in the normal mode, and to transmit the reconstructed difference signal to the code compensation means in the separated mode Way; The code compensating means for compensating the code of the difference signal transmitted from the mode selecting means by using the compensating code bit; And The image restoring means for restoring the original image by compensating for the motion of the coded difference signal or the difference signal transmitted from the mode selection means; Video decoding apparatus comprising a. The method of claim 12, The encoding mode information, The video decoding apparatus, characterized in that the mode flag received from the transmitting side. The method according to any one of claims 11 to 13, The compensation code bit, The video decoding apparatus, characterized in that the code bit of the differential signal generated and transmitted by the video encoding apparatus of the transmitting side. The method of claim 14, The code compensation means, And a non-zero pixel of the reconstructed differential signal is compensated for by using a compensation code bit transmitted from a transmitting side. In the video encoding method using the separation of the magnitude and code of the difference signal, An image prediction step of generating a prediction image with respect to the original image to be currently encoded; Generating a difference signal between the original image and the prediction image; A magnitude / code separation step of generating a magnitude and a code bit (compensation code bit) required to recover the difference signal from the difference signal; An encoding step of transform encoding, quantization, and entropy encoding on the magnitude of the difference signal, and entropy encoding on the compensation code bits; And In the encoding step, the transformed and quantized differential signal is inversely transformed to restore the original differential signal, and then code-compensated using the compensation code bits. Decoding step to save as a new reference image Video encoding method comprising a. In the video encoding method using the separation of the magnitude and code of the difference signal, An image prediction step of generating a prediction image for the original image to be encoded; Generating a difference signal between the original image and the prediction image; A size / code separation step of determining an encoding mode for the difference signal based on a compression efficiency, and generating a magnitude and a code bit (compensation code bit) required to recover the difference signal from the difference signal; A normal mode encoding step of performing transform encoding, quantization, and entropy encoding on the difference signal if the encoding mode is a normal mode; In the normal mode, in the normal mode encoding step, the transform-coded and quantized signals are inversely transformed to be restored to the differential signal, and then the normal mode decoding is performed to restore the original image and store the new reference image by motion compensation to the restored differential signal. step; And A separation mode encoding step of performing encoding, quantization, and entropy encoding on the difference signal size and entropy encoding on the compensation code bits when the encoding mode is a separation mode; And In the separation mode, the transform-coded and quantized signals are inversely transformed in the separation mode encoding step to be restored to the original differential signal, and then the code is compensated using the corresponding compensation code bits. Separate mode decoding step of compensating and restoring original image and storing as reference image Video encoding method comprising a. The method according to claim 16 or 17, The size / code separation step, And a magnitude of the difference signal is obtained by taking an absolute value of the difference signal. The method according to claim 16 or 17, The size / code separation step, The video encoding method of claim 1, wherein the encoding mode is determined based on a compression efficiency. The method of claim 19, The size / code separation step, A video encoding method characterized in that the compression efficiency is determined according to a rate-distortion optimization (RDO) method. The method according to claim 16 or 17, The size / code separation step, And a code bit of the difference signal corresponding to a position of a pixel whose pixel value of the difference signal reconstructed through the inverse transform in the decoding step is not "0" as the compensation code bit. The method according to claim 16 or 17, The size / code separation step, And a code bit of the difference signal corresponding to a position of a pixel whose pixel value of the difference signal reconstructed through inverse transformation in the decoding step is a compensation code bit. The method according to claim 16 or 17, The size / code separation step, And the compensating code bit is generated by using correlation between codes of the difference signal reconstructed through inverse transform in the decoding step. The method of claim 23, The size / code separation step, In the decoding step, the pixel values of the difference signal reconstructed through the inverse transformation are arranged in descending order of the magnitude of the absolute value, and the original code for the aligned pixel values is transmitted in the sorting order as the compensation code bits. The video encoding method according to claim 1, wherein only the number of codes and the number of codes are transmitted for consecutive codes. In the video decoding method using the separation of the magnitude and code of the difference signal, A decoding step of entropy decoding the input bit stream, inverse quantization and inverse transform encoding to restore an original differential signal; A code compensating step of compensating a code of the restored differential signal by using a corresponding compensation code bit; And An image restoration step of restoring an original image by performing motion compensation on the coded difference signal; Video decoding method comprising a. In the video decoding method using the separation of the magnitude and code of the difference signal, A decoding step of entropy decoding the input bit stream, inverse quantization and inverse transform encoding to restore an original differential signal; A code compensating step of verifying encoding mode information and compensating a code of a difference signal reconstructed in the decoding step by using a corresponding compensation code bit; And Image reconstruction means for reconstructing the original image by motion compensation on the differential signal reconstructed in the decoding step or the differential signal reconstructed in the code compensation step Video decoding method comprising a. The method of claim 12, The encoding mode information, A video decoding method, which is a mode flag received from a transmitting side. The method according to any one of claims 25 to 27, The compensation code bit, And a code bit necessary for restoring the difference signal is transmitted by a video encoding apparatus on a transmission side. The method of claim 28, The code compensation step, And using the compensation code bits to compensate for pixels other than "0" of the reconstructed differential signal.

Images