KR100380230B1 - Image codec system based on multi-resolution image - Google Patents

Abstract

The present invention relates to a method and apparatus for encoding and decoding digital information using an error detection and error concealment technique.

In the present invention, in the encoding process, the digital image data is divided into multi resolutions to configure n (n ≥ 2) multi-resolution images for one original image, and each of the multi-resolution image data is encoded into one bit stream. In the decoding process, the image of each resolution is separated and decoded from the bit stream of the multi-resolution image data, and the image signals of each resolution are compared with each other on a block-by-block basis. Proposes an image encoding and decoding method and apparatus for improving the quality of a reconstructed image by replacing and outputting the generated block by using different resolution image data (block) of the actual image without error. do.

Description

Method and apparatus for encoding and decoding multi-resolution based video signals {IMAGE CODEC SYSTEM BASED ON MULTI-RESOLUTION IMAGE}

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and apparatus for encoding and decoding digital information using error detection and error concealment techniques. In particular, the present invention relates to a method of processing an image signal by processing a digital image signal in units of macro blocks. In an image encoding and decoding system that performs encoding and decoding, the original image data is converted into a multi-resolution image and encoded, and when decoding, an error block is found by comparing the multi-resolution image to conceal the corresponding error. The present invention relates to a video encoding and decoding method and apparatus for improving the quality of reconstructed video.

More specifically, the present invention converts an input image into a multi-resolution image having different resolutions and transmits the multi-resolution image as one bit stream when compressing and encoding the moving image, and decodes the multi-resolution image by separating the multi-resolution image. Decode each image, compare each multiresolution image block by block, find the error block, and replace the error block with an image block of a different resolution corresponding to the corresponding position in the error concealer. The present invention relates to a multi-resolution based image encoding and decoding method and apparatus for improving the quality of a received image by decoding and displaying a signal.

A system that compresses, transmits, and encodes digital multimedia data. For example, in H.263 or MPEG4, when compressing and encoding a video, digital image data is processed in units of macro blocks. In case of encoding, It includes motion estimation and compensation, discrete cosine transform (DCT), quantization of DCT coefficients, and variable length coding (VLC). In the case of decoding, inverse quantization, inverse DCT, It includes motion compensation and estimation process.

In the system that transmits / receives digital video signals, error detection according to the channel environment is to be taken into account, and the system that needs to correct damaged blocks in case of such error occurrence uses error detection and concealment techniques to improve the quality of restored (received) images. Doing.

For example, when an error is detected in a system that transmits and receives a digital video signal in units of blocks, an error is regarded as an error when the data deviates from the syntax information by using encoded syntax information. Alternatively, the image quality of the received image can be improved by applying correction using a spatial concealment technique.

The detection of the error using the syntax information as described above may be performed, for example, when a motion vector deviates from a specific range, a variable length coding (VLC) table, or a discrete cosine transform during processing of a digital video signal in a block unit. The case where the DCT) coefficient is out of a specific range, and the number of DCT coefficients in one block is out of the specific number.

In the case of error concealment, there are a method using spatial information as shown in Fig. 1A and a method using temporal information as shown in Fig. 1B.

The above two methods are performed under the assumption that the correlation between adjacent blocks of the video signal is high and the correlation between frames (blocks) in the adjacent time zone is high in a system for processing digital video signals on a block basis.

In case (a) of FIG. 1, an error occurs and a damaged block is interpolated using values of adjacent intact (assuming intact) blocks adjacent to each other. In the same case, an error occurs and the damaged block is obtained by filling in the block of the corresponding position of the previous frame (or field). In the case of FIG. 1B, the motion vector MV is used. If there is an error in the motion vector, motion vector information of the same position of the previous frame is obtained.

However, both of these methods replace the errored block with the surrounding or previous value (or interpolation through proper operation), and these filled values are, in strict sense, the original blocks of the errored block itself. This information is irrelevant. Of course, there may be some degree of similarity based on the correlation, but replacing the error part with the surrounding or other value beforehand is a big difference from the image to be restored (the original block information before the error occurred). This may result in severe image quality deterioration in some cases.

In addition, although the error is detected only by the syntax of the given information, even in such a case, in many cases, the error may not be detected despite the occurrence of the error. This is because an error has occurred, but it can be determined that an error has not occurred if the error is released from the case determined as the above-described error.

According to the present invention, in a system for compressing, encoding, and decoding digital image data by block unit, the detection and concealment of an error can be performed on the actual original image information without using spatial correlation or temporal correlation as in the prior art. We propose a video encoding and decoding method and apparatus therefor.

In the present invention, in the encoding process, the digital image data is divided into multi resolutions to configure n (n ≥ 2) multi-resolution images for one original image, and each of the multi-resolution image data is encoded into one bit stream. In the decoding process, an image of each resolution is separated and decoded from the bit stream of the multi-resolution image data, and an error is generated according to a result of detecting an error by comparing the image information of each resolution with each other. The present invention proposes a video encoding and decoding method and apparatus for improving the quality of a received image by performing error correction using different resolution image data of the actual image having no error.

1 is a diagram illustrating an error concealment method using spatial and temporal information in an image compression coding technique.

2 is an embodiment block diagram of a multi-resolution encoding and decoding apparatus of the present invention.

3 is a block diagram showing an internal configuration of a video encoder according to the present invention.

4 is a block diagram showing an internal configuration of a video decoder according to the present invention.

5 is a block diagram showing an internal configuration of an error detection unit of the present invention;

6 is a block diagram showing an internal configuration of an error concealment unit of the present invention;

The present invention provides a method of converting a video signal into a video signal having a different resolution and encoding the same into a multi-resolution video, constructing and transmitting the encoded video signal for each resolution into one bit stream, and Separating and decoding a video signal into video signals for each resolution, comparing the decoded video signals for each resolution, and detecting an error, and converting the video signal at the position where the error is detected to another resolution video signal without error. Restoring and replacing the output; Multi-resolution based image encoding and decoding method characterized in that consisting of.

The present invention also provides a resolution conversion means for separating a video signal into video signals having different resolutions and converting the video signal into a multi-resolution video, encoding means for encoding the separated and converted video signals for each resolution, and for each of the encoded resolutions. Multiplexing means for constructing and outputting a video signal into one bit stream; A multi-resolution based image encoding apparatus comprising a.

The present invention also provides demultiplexing means for separating the multi-resolution video signals transmitted in one bit stream into video signals for each resolution, decoding means for decoding the separated video signals for each resolution, and the decoded angles. Error detection means for detecting an error by comparing the video signals for each resolution, and error concealment means for correcting the video signal at the position where the error is detected using another resolution video signal without error and outputting a reconstructed video signal; A multi-resolution based video decoding apparatus comprising a.

In the present invention, the video signal of the multi-resolution structure is characterized in that the three video signals of different resolution.

In the present invention, in the multi-resolution video encoding, a video signal having a relatively low resolution among the multi-resolution video signals is generated using a video signal reconstructed from a coded video signal having a relatively high resolution. .

Also, in the present invention, the decoded video signals for each resolution are compared block by block, the PSNR (Peak Signal to Noise Ratio) is calculated between the blocks to be compared, and the PSNR is compared with a predetermined threshold to replace an error block. Map information is generated and the block in which the error is detected is replaced with a corresponding block value of an image signal having a different resolution using the map information.

Also, in the present invention, the smoothing process for the blocks is performed before calculating the PSNRs between the compared blocks.

In the present invention, the decoded video signal for each resolution is further increased or decreased in resolution, and the error block of the video signal is replaced with a block of a corresponding position of the resolution-converted video signal. .

In the present invention, when the error block of the video signal is replaced with a block of the corresponding position of the resolution-converted video signal, the video signal having a higher resolution than that of the video signal is replaced with a block of the corresponding position of the video signal converted to a lower resolution. Characterized in that.

Also, in the present invention, when the error block of the image having the highest resolution is replaced with the block of the corresponding position of the resolution-converted video signal, the next position of the corresponding position of the video signal obtained by inverting the higher resolution video to higher resolution. It is characterized by replacing with a block.

With reference to the accompanying drawings, the operation of the multi-resolution based image encoding and decoding method and the apparatus of the present invention made as described above will be described.

2 is a block diagram showing an embodiment of a multi-resolution based image encoding and decoding apparatus according to the present invention. In this embodiment, an encoding and decoding apparatus for an image signal having three different resolutions is shown.

In the present invention, g-picture (ground-picture)> f-picture (first-picture)> s-picture (second-picture), and encoders encoding the respective video, respectively, g encoder, f encoder, and s encoder The decoder which decodes each video is distinguished by naming g decoder, f decoder and s decoder, respectively.

First, the encoder; G encoder 201 for encoding an input video signal, a first wavelet converter 202 for converting the g video signal reconstructed by the g encoder into a lower resolution video signal, and the converted low resolution video signal. An f encoder 203 for encoding, a second wavelet converter 204 for converting the f video signal reconstructed by the f encoder into a lower resolution video signal, and an s encoder for encoding the converted low resolution video signal 205 and a multiplexer configured to output a compression-coded video signal (g-code stream, f-code stream, s-code stream) of different resolutions output from the encoders 201, 203, and 205 into one bit stream. 206.

And, the decoding unit; A demultiplexer 207 for receiving the compressed coded multi-resolution video signal bit stream and separating it into video signals for each resolution, and g for decoding the g, f, and s video signals for each resolution separated by the demultiplexer 207, respectively. The decoder 208, the f decoder 209 and the s decoder 210 are compared with an error detection unit 211 for detecting an error by comparing the video signals decoded by the respective decoders, and an error detection result of the error detection unit is input. And an error concealment unit 212 that performs an error concealment process by using the video signals decoded by the respective decoders.

The multiresolution-based image encoding and decoding process will now be described with reference to the image encoder and decoder of the present invention configured as described above.

The g encoder 201 receives motion picture digital data (digital sequence), performs motion estimation and compensation in macroblock units, performs DCT and quantizes the result, and then performs variable length coding (VLC) to encode the video. The encoded bit stream (g-code stream) is input to the multiplexer 206, and performs inverse quantization with IDCT, which is the same process as that performed by the decoder, for motion estimation and compensation. It is used as a reference image for estimation and compensation and is also input to the first wavelet transform unit 202.

The first wavelet transform unit 202 converts the input g image into a low resolution image by performing a 1 level forward wavelet transform, wherein the wavelet transform is half-resolution. A corresponding image (LL-frame) is generated, and the low resolution image is input to the f encoder 203.

That is, if the resolution of the current original image is m × n, the f image is lowered to a resolution of m / 2 × n / 2.

Like the g encoder, the f encoder 203 compresses and encodes a low resolution video to output a bit stream, which is input to the multiplexer 206. In addition, the f encoder 203 generates a reconstructed image (f image), which is converted into an s image having the lowest resolution by the second wavelet converter 204 and input to the s encoder 205. That is, the resolution of the f-image is reduced by half again by the second wavelet converter 204, and the s-image is eventually reduced to m / 4 × n / 4 than the original image.

Similarly to the f encoder and the g encoder, the s encoder 205 compresses and encodes a low resolution s image to output a bit stream, which is input to the multiplexer 206.

The encoders 201, 203, and 205 are typical video encoders and their internal structures are shown in FIG.

The encoder 301 encodes the input video signal, which is then decoded (dequantized, IDCT, motion compensation process) through the decoder 302 to restore the motion estimation and compensation. The reconstructed image is stored in the frame memory 304 through the delay unit 303, and the reconstructed image stored in the frame memory 304 is provided to the encoder 301 for motion estimation and compensation. It is used as a reference image.

As described above, the bit streams of the multi-resolution video signals encoded by the g encoder, the f encoder, and the s encoder, respectively, are multiplexed by the multiplexer 206 and constitute one bit stream, and the bit stream is transmitted through a predetermined channel. . Here, it may be effective to channel-code each bit stream with a different protection level.

The bit stream transmitted as described above is received by the decoder and decoded as follows.

First, the demultiplexer 207 is a video signal bit stream for each resolution, that is, a g-code stream, a f-code stream, and a s-code stream from the received bit stream of the multi-resolution video signal. The s-code stream is separated and input to the s decoder 208, f decoder 209 and s decoder 210, respectively.

The g decoder 208 outputs the decoded g image by performing the above-described decoding process on the bit stream of the input g image, and the f decoder 209 performs the decoding process on the bit stream of the input f image. The decoded f image is output, and the s decoder 210 performs a decoding process on the bit stream of the input s image to output the decoded s image.

4 shows an internal structure of the decoders 208, 209 and 210, which also applies to the decoder 302 of the encoder.

The decoder 400 is a typical decoder. The decoder 400 delays the reconstructed image (here, the reconstructed image output from the error concealment unit 212) in the delay unit 401 and stores the reconstructed image in the frame memory 402 so that the decoder 403 When performing the decoding process (decoding process such as inverse quantization, IDCT, motion estimation and compensation), it is used for motion estimation and compensation for the next image frame.

As described above, each decoder 208, 209, 210 decodes and outputs a video signal having a different resolution, and the output g, f, and s video signals for each resolution are output by an error detector 211 and an error concealment unit ( 212).

The error detection unit 211 compares the input video signals for each resolution with each other to determine an error block, and outputs the determination result as map information (gf-map, gs-map, fs-map) to the error concealment unit 212. to provide.

The internal structure of the error detector 211 is shown in FIG. 5, and the error detector 211 basically compares the g image, the f image, and the s image to determine whether there is an error, and what part of the image having the error is. (Block) Check for errors.

An internal configuration and operation of the error detector 211 will be described with reference to FIG. 5.

First, the error detector 211 may include a first wavelet transform unit 501 for converting a g image signal into a lower resolution image, and a low resolution image and an f image output from the first wavelet transform unit 501. A first comparison unit 502 for comparison, a second wavelet converter 503 for converting the image converted by the first wavelet converter 501 into a lower resolution image, and the second wavelet A second comparison unit 504 for comparing the low resolution image and the s image output from the conversion unit 503, a third wavelet converter 505 for converting the f image to a lower resolution image, and the It consists of a third comparison unit 506 for comparing the low-resolution image and the s image output from the three-wavelet converter 505, the operation is as follows.

The g image decoded by the g decoder is input to the first wavelet converter 501 to be converted into a lower resolution image (f-class image), and the first comparator 502 is a first wavelet converter. Comparing the output of the 501 and the f image decoded by the f decoder, the comparison process in the first comparison unit 502 is made in units of blocks, the comparison result is the gf map information (gf-map [i, j]) (where i, j is an index indicating a block position) and sent to the error concealment unit 212.

In addition, since it is difficult to see whether each block is correctly matched, a smooth signal operation of a 2 × 2 window is performed, and then a peak signal to noise ratio (PSNR) between the comparison blocks is calculated, and the calculated By comparing the PSNR value with a specific threshold, when the calculated PSNR value is lower than the threshold, the corresponding blocks to be compared with each other are determined as 'unmatched', and when higher than the threshold, it is determined as 'matched'.

Meanwhile, the output of the first wavelet converter 501 is converted into a lower resolution image (s-class image) through the second wavelet converter 503, and together with the s image decoded by the s decoder. The second comparator 504 is input to the second comparator 504, and the second comparator 504 compares the output of the second wavelet converter 503 with the s image. The comparison result is made in units of blocks as in the comparison unit 504, and the comparison result is stored as gs map information (gs-map [i, j]) (where i, j is an index indicating a block position) and an error concealment unit ( 212).

That is, similarly to the first comparison unit 502, the second comparison unit 504 performs block-wise comparison and planarization operations, and performs matching and inconsistency determination according to the PSNR.

In addition, the third wavelet converter 505 converts the f image decoded by the f decoder into a lower resolution image (s-class image) and inputs the same to the third comparison unit 506. 506 compares the wavelet transformed image with the s image decoded by the s decoder. This process, like other comparison units, also performs block-by-block comparison and planarization, and performs coincidence and mismatch determination according to PSNR. , fs map information (fs-map [i, j]) is sent to the error concealment unit 212.

As described above, the error concealment unit 212 uses an error block according to the gf-map [i, j], gs-map [i, j], and fs-map [i, j] information output from the error detection unit 211. In this case, the map information may be as follows.

(a). gf-map [i, j] = matches, gs-map [i, j] = matches, fs-map [i, j] = matches; In this case, even if there is no error in the block [i, j] or very small, it is negligible.

(b). gf-map [i, j] = matches, gs-map [i, j] = matches, fs-map [i, j] = matches; In this case, it is a problem of error detection (verification) or a negligible error in the block [i, j].

(c). gf-map [i, j] = matches, gs-map [i, j] = matches, fs-map [i, j] = matches; In this case, it is a problem of error detection (verification) or a negligible error in the block [i, j].

(d). gf-map [i, j] = matches, gs-map [i, j] = matches, fs-map [i, j] = matches; In this case, there is an error in the block [i, j] of the s image, and this block is required to be replaced by a block of another image. Therefore, in this case, the error concealment unit 212 converts the f image to a wavelet transform to produce a lower resolution image (which will be a s-class image) and replaces it with the block in which the problem occurs.

(e). gf-map [i, j] = unmatched, gs-map [i, j] = unmatched, fs-map [i, j] = matched; In this case, it is a problem of error detection (verification) or a negligible error in the block [i, j].

(f). gf-map [i, j] = unmatched, gs-map [i, j] = matched, fs-map [i, j] = unmatched; In this case, there is an error in the block [i, j] of the f picture, and this block is required to be replaced by a block of another picture. Therefore, in this case, the error concealment unit 212 converts the g image into a wavelet transform to produce a lower resolution image (which will be a f-class image) and replaces it with the block in which the problem occurs.

(g). gf-map [i, j] = unmatched, gs-map [i, j] = unmatched, fs-map [i, j] = matched; In this case, there is an error in the block [i, j] of the g picture, and this block is required to be replaced by a block of another picture. Therefore, in this case, the error concealment unit 212 converts the f image to inverse wavelet to produce a higher resolution image (which will be a class-G image) and replaces it with the block in which the problem occurs.

(h). gf-map [i, j] = inconsistency, gs-map [i, j] = inconsistency, fs-map [i, j] = inconsistency; In this case, the present invention is unavoidable. Therefore, other analysis means is used.

Looking at the map information, error concealment occurs when there is an error in one image among g image, f image, s image, and three images, which is the case of (d), (f), and (g). . Among the other cases, (b), (c), and (e) are difficult to occur because one of three images has an error in one of the three map information (gf-map, gs-map, fs-map) logically affects two. In the case of (h), an error occurs in two images at the same time. In this case, processing is performed using other known analysis means, or the one with the least error or unconditionally selecting one of three images. You can follow the other methods.

FIG. 6 is an embodiment showing the internal structure of the error concealment unit 212 for the case where error concealment is required in each case described above.

Looking at the configuration, the first image replacement unit 601 for outputting the error-corrected g image, the first wavelet converter 602 for wavelet transforming the g image, and the wavelet for inverse wavelet transforming the f image Outputting the inverse transform unit 603, a second image replacing unit 604 for outputting the corrected f-image, a second wavelet transform unit 605 for wavelet transforming the f-image, and the corrected s-image The third image replacement unit 606 and a controller 607 for receiving the map information and controlling the image replacement units in each case.

Here, the controller 607 selectively controls the image replacement unit to perform the operation (block replacement operation) for each of the eight cases described above for gf-map, gs-map, and fs-map, and receives the corresponding image under the control. The replacement unit replaces the error block with another image block having no error at the corresponding position, thereby outputting an image in which the error is corrected. The corrected g image, the corrected f image, and the corrected s image output from each image replacing unit are reconstructed image information for motion compensation to the g decoder 208, the f decoder 209, and the s decoder 210, respectively. Give feedback. And the g picture will be displayed as the finally decoded picture signal.

If there is an error in the g image, the first image replacing unit 601 inverts the image to the high resolution image by the wavelet inverse transform unit 603 (the resolution is increased to the grade g by inversely converting the f image). Is replaced with the block in which the problem occurs, and the second image replacement unit 604 converts the image converted into a low resolution image by the wavelet converter 602 when the f image has an error (wavelet transformed to g image). The resolution is lowered to f class), and the third image replacement unit 606 converts the image to a low resolution image by the wavelet converter 605 when there is an error in the s image. The resulting image (the wavelet-converted f-image and lowered to s level) is replaced with the problem block.

In this way, an error-corrected g, f, s image is output, and finally, the g-corrected g image is displayed.

However, since the g image is corrected from the f image having a lower resolution than the self image, the corrected block is low resolution and the uncorrected block is high resolution, so this part may be visually unobtrusive.

Therefore, the motion vector (MV) of the f-image can be used as an alternative to the low-resolution block obtained from the f-image (except for bringing a block from the f-image or bringing a motion vector, or both, to the corresponding block of the f-image). It should be assumed that it is not corrupted by this channel error).

At this time, the motion vector of the g image may be different from the motion vector of the f image, but considering that the f image is a low resolution image of the g image, the two images can be expected to be almost similar. The motion compensated block can thus be another candidate as a correction block.

That is, a low resolution block taken from the two candidate f-pictures as the correction block described above, and a block compensated from the motion vector of the f-picture, and a better image quality can be selected from the two blocks.

As a criterion for selecting this, SAD (Sum of Absolute Difference) between the boundary pixel of the current block and the boundary fixation of the adjacent block may be used.

That is, when the SAD of the low resolution block obtained from the f image is SADl and the SAD of the motion compensated block using the motion vector of the f image is SADm,

If SADl + threshold <SADm, a low-resolution block taken from the f-image is selected as the correction block, and vice versa, a motion-compensated block from the motion vector is selected as the correction block and replaced with the corresponding error block in the g-image.

The threshold is given here, which is generally the value expected and determined by SADl to be less than SADm.

According to the present invention, when the compressed video is encoded and transmitted, the input video is converted into a multi-resolution video having different resolutions, and the multi-resolution video is transmitted as one bit stream, and the decoder separates and decodes the multi-resolution video. Decode the error-corrected video signal by comparing each multi-resolution video block by block to find an error block and replacing the error block with an image block of a different resolution corresponding to the corresponding position in the error concealer. Display.

Therefore, according to the present invention, it is possible to expect an improvement in image quality of a reconstructed image. In particular, by effectively detecting and concealing an error generated according to a channel environment, an error detection and concealment technique with improved performance can be provided.

Claims (13)

Converting a video signal into a plurality of video signals having different resolutions and encoding the multi-resolution video; constructing and transmitting the encoded video signals for each resolution into one bit stream; and receiving the received multi-resolution video. Separating and decoding the signal into video signals for each resolution, comparing the decoded video signals for each resolution, and detecting an error, and converting the video signal at the position where the error is detected to another resolution video signal without error. Restoring and outputting the restored data; Multi-resolution based image encoding and decoding method characterized in that consisting of. The multi-resolution based video encoding and decoding method according to claim 1, wherein the multi-resolution video signal is three video signals having different resolutions. The method of claim 1, wherein in the multi-resolution video encoding, a video signal having a relatively low resolution among the multi-resolution video signals is generated by using a video signal obtained by reconstructing a coded video signal having a relatively high resolution. Multi-resolution based image encoding and decoding method. The method of claim 1, wherein the decoded video signals for each resolution are compared block by block, PSNRs are calculated between blocks to be compared, and PSNRs are compared with a predetermined threshold to generate map information for error block replacement. And converting the block in which the error is detected by using the information into a corresponding block value of a video signal having a different resolution and outputting the same. 5. The method of claim 4, wherein a smoothing process is performed prior to calculating the PSNRs between the blocks to be compared. 2. The method of claim 1, further comprising converting a video signal for each resolution to increase or decrease the resolution when concealing the error, and replacing an error block of the video signal with a block of a corresponding position of the resolution-converted video signal. A multi-resolution based video encoding and decoding method. The video signal according to claim 6, wherein when the error block of the video signal is replaced with a block of the corresponding position of the resolution-converted video signal, the video signal having a higher resolution than that of the video signal is converted into a lower position of the block of the video signal. Multi-resolution based image encoding and decoding method characterized in that for replacing. 7. The block according to claim 6, wherein when the error block of the highest resolution image is replaced with the block of the corresponding position of the resolution-converted video signal, the block of the corresponding position of the video signal in which the next higher resolution image is inversely converted to a higher resolution. Multi-resolution based image encoding and decoding method characterized in that for replacing. The method according to claim 6, wherein when concealing using a motion vector of a corresponding position of a lower resolution image is advantageous than concealing using a block of a corresponding position of a lower resolution image, the position of the corresponding block of the previous frame is replaced using a motion vector. Multi-resolution based image encoding and decoding method. A resolution converting means for separating a video signal into video signals having different resolutions and converting the video signal into a multi-resolution video, encoding means for encoding the separated and converted video signals for each resolution, and one encoded video signal for each resolution. Multiplexing means for constructing and outputting a stream of bits, demultiplexing means for separating the multi-resolution video signals received from the multiplexing means into video signals for each resolution, and decoding means for decoding the separated video signals for each resolution; Error detection means for detecting an error by comparing the decoded video signals for each resolution, and an error for correcting the video signal at the position where the error is detected using another resolution video signal without error and outputting a restored video signal. Concealing means; Multi-resolution based image encoding and decoding apparatus comprising a. delete 12. The apparatus of claim 10, wherein the error detecting means comprises: a resolution converting means for converting a resolution of a video signal decoded for each resolution, and comparing the resolution-converted video signals with video signals of each decoded resolution on a block-by-block basis; And a comparison means for outputting an error detection result. 11. The apparatus of claim 10, wherein the error concealment means comprises: resolution converting means for converting a resolution for a decoded video signal for each resolution, and an error block at a position designated by the error detecting means for the decoded video signal. And an error block replacement means for replacing the block of the corresponding position of the resolution-converted video signal.