JP2006042371A - Image recording and reproducing apparatus and image reproducing apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To improve coding efficiency while suppressing entire throughput in moving image coding and to realize appropriate image reproduction corresponding to a search speed when searching for an image. <P>SOLUTION: There are included: a block dividing unit 1 for dividing an input image 100 into a plurality of blocks; an error detecting unit 12 for detecting an error between a block image at the same position of a preceding frame stored/held in a frame memory 14 and a current coding target block; a coding mode control unit 2 for selectively controlling whether or not the coding target block is to be coded, in accordance with the detected error; and a circuit unit for wavelet transform coding for performing wavelet transform coding for each block image, the circuit unit for wavelet transform coding being comprised of a wavelet transform unit 4, a scanning unit 5, a quantization unit 6 and entropy coding unit 13. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a moving image encoding apparatus and method, a decoding apparatus and method, and an image recording / reproducing apparatus that can be used in a system for efficiently transmitting or storing images. The present invention relates to a moving image encoding apparatus and method, a decoding apparatus and method, and an image recording / reproducing apparatus that use block-based wavelet transform.

  As a conventional typical image compression method, there is a JPEG (Joint Photographic Coding Experts Group) method standardized by ISO (International Organization for Standardization). This JPEG system is a system that mainly compresses and encodes still images using DCT (Discrete Cosine Transform). When relatively high bits are assigned, a good encoded / decoded image is obtained. It is known to provide However, in this method, when the number of encoded bits is reduced to some extent, block distortion peculiar to DCT becomes remarkable, and deterioration becomes conspicuous subjectively.

  Apart from this, recently, research on a method of dividing an image into a plurality of bands by a filter that combines a high-pass filter and a low-pass filter called a filter bank, and performing coding for each band has become active. ing. Among them, wavelet coding is regarded as a promising new technology to replace DCT because there is no defect that block distortion becomes remarkable due to high compression, which is a problem in DCT.

  Currently, many products such as electronic still cameras and video movies adopt the JPEG method, MPEG (Moving Picture image coding Exparts Group) method, or the so-called DV (Digital Video) method for compression coding. All the conversion methods use DCT as the conversion method. In the future, it is speculated that the products based on the wavelet transform will appear on the market, but studies for improving the efficiency of the coding method are being actively conducted by each research institution. In MPEG, encoding efficiency is improved by performing motion compensation prediction with a previous frame in block units called macroblocks to be encoded.

  Patent document 1 is known as a prior art.

Japanese Patent Laid-Open No. 5-284368

  By the way, as a technique for image compression / decompression required for electronic still cameras, video movies, etc., wavelet transform, which has higher compression efficiency than DCT, is used, and motion compensation prediction processing that occupies most of the processing in MPEG It is desired to obtain a good image while omitting.

  In view of such a situation, the present invention is provided with a frame memory means especially in the compression / decompression of moving images, so that the encoding efficiency can be improved while suppressing the entire processing amount. Focusing on the fact that one of the excellent features that can support multi-resolution, the video compression recording / decompression device equipped with the image compression / decompression device of the present invention matches the search speed when searching for images. Thus, by selectively decoding a predetermined band component, a moving picture coding apparatus and method, decoding apparatus and method, and image recording capable of performing a search that could not be realized by a conventional video recorder or the like An object is to provide a playback device.

  In order to solve the above-described problems, a moving image encoding apparatus and method according to the present invention divides an input image into a plurality of blocks, and block images at the same position of the previous frame stored and held in the image memory. And an error with the current encoding target block, select whether to encode the encoding target block according to the detected error, and perform wavelet transform encoding for each block image It is characterized by that.

  Here, the image memory includes storing and holding an image obtained by performing wavelet inverse transform decoding on the encoded output obtained by the wavelet transform encoding.

  Further, the moving picture decoding apparatus and method according to the present invention inputs or reads an encoded bitstream, performs inverse wavelet transform decoding, outputs a decoded image, extracts a target area from the obtained decoded image, and obtains the obtained decoding The image is stored and held in the image memory, and the decoded image of the previous frame stored in advance in the image memory is transmitted according to the supplied encoding control signal, or the image decoded by the wavelet inverse transform decoding means is transmitted. It is characterized by selecting whether to send.

  Here, when performing the wavelet inverse transform decoding, the quantized transform coefficient is inversely quantized and returned to the transform coefficient, the transform coefficient is reverse-scanned (reverse scanning), and the object to be decoded is obtained from the obtained transform coefficient. Extraction of the transform coefficient to be obtained, and inverse wavelet transform of the obtained decoding target coefficient.

  Furthermore, the image recording / reproducing apparatus according to the present invention comprises means for converting an input image into a plurality of bands by means of filter bank means comprising a plurality of low-pass filters and high-pass filters, An image compression apparatus having means for recording an encoded bit stream on a recording medium by means for scanning, quantizing, and entropy encoding the transform coefficient every time, and reading the encoded bit stream from the recording medium and entropy decoding , Inverse quantization, inverse scanning, and inverse transform by the filter bank means, and an image decompression device having means for outputting a decoded image, and recorded on the recording medium in accordance with the search speed. And a search means for reading out the corresponding band component from the encoded bit stream.

  According to the present invention, an input image is divided into a plurality of blocks, and an error between the block image at the same position of the previous frame stored and held in the image memory and the current encoding target block is detected and detected. By selecting whether to encode the target block according to the error, and performing wavelet transform coding for each block image, the conventional motion-compensated prediction means with very heavy processing is used. However, since a simple prediction means in units of blocks is used, the processing amount can be very small. Here, since the background portion hardly moves in a normal moving image, the present invention works effectively in such a portion, and there is an effect that the compression efficiency is increased.

  In addition, since the motion compensation means can be greatly simplified, the motion compensation means is conventionally realized by dedicated hardware (LSI), but can be realized by software in the present invention.

  Also, using block-based coding means, and for the pixels where the filter protrudes from the block when wavelet transform coding is used, the block boundary distortion is not noticeable, so smooth connection surface decoding is always performed. There is also an effect that an image can be obtained.

  According to the image recording / reproducing apparatus of the present invention, the filter bank means comprising a plurality of low-pass filters and high-pass filters converts the input image into a plurality of bands. An image compression apparatus comprising: means for scanning, quantizing, and entropy encoding the transform coefficient for each band; and recording the encoded bitstream on the recording medium; and reading the encoded bitstream from the recording medium and entropy An image decompression device having means for decoding, inverse quantization, inverse scanning, and inverse transformation by the filter bank means, and outputting a decoded image, and recording it on the recording medium in accordance with the search speed The encoded bit stream recorded on the recording medium has search means for reading out the corresponding band component from the encoded bit stream. Read out, when performing a search of the decoded image can be read only optimum band components tailored to search speed can be decoded to obtain a consistently high-quality decoded image without frame dropping.

  In addition to the size of the original image as well as the resolution of the band component, the decoded image can also have the resolution of the band component, so that the variation of the decoded image increases, so that it can be used properly according to the application.

  Therefore, according to the present invention, both high image quality under high compression and partial block decoding are realized, and further, the wavelet transform unit and the wavelet inverse transform unit can be matched, so that highly efficient coding can be achieved. And high-quality decoding.

  Furthermore, in the case of an interlaced image, it is separated into an odd-numbered field image and an even-numbered field image and encoded separately, so that even when a subject with large motion exists in the image, a high coding efficiency is always obtained. Can be maintained.

  DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments of an image moving image encoding apparatus and method and a decoding apparatus and method according to the present invention will be described below with reference to the drawings.

FIG. 1 shows a configuration example of an image moving image coding apparatus according to the first embodiment of the present invention.
The moving image coding apparatus for an image shown in FIG. 1 has the same position of a block dividing unit 1 that divides an input image 100 into a plurality of blocks and a previous frame stored and held in a frame memory 14 that is an image memory. An error detection unit 12 for detecting an error between the block image in FIG. 5 and the current encoding target block, and outputting the current encoding target block without encoding according to the detected error or wavelet transform encoding A coding mode control unit 2 that selectively controls the coding mode to be applied, a coding target region extraction unit 3 that extracts a coding target region according to a control signal of the coding mode, and a block image A wavelet transform unit 4 for performing wavelet transform, a transform coefficient scanning unit 5 for scanning wavelet transform coefficients, and an amount for quantizing the coefficients after scanning The unit 6 is configured by including an entropy coding unit 13 for outputting an encoded bit stream by entropy coding the quantized coefficients. As will be described later, the wavelet transform unit 4 includes means for performing a convolution operation by symmetrically expanding pixels inside a block within a range covered by filtering outside the block.

  In the moving picture coding apparatus shown in FIG. 1, an input image 100 is first input to the block division unit 1, divided into a plurality of block images 101 by the block division unit 1, and sent to the error detection unit 12. In the error detection unit 12, the block image 115 at the same position as the block image 101 read from the decoded image of the previous frame stored and held in the frame memory 14, and the block image 101 from the block division unit 1 , The error 102 is detected, and the detected error 102 is sent to the encoding mode control unit 2. The encoding mode control unit 2 determines one of two encoding modes depending on the magnitude of the error 102.

  That is, when the error 102 is larger than the predetermined threshold (error 102> threshold), it is determined that the content of the block image has changed greatly. Output a stream. On the other hand, when the error 102 is equal to or less than a predetermined threshold (error 102 ≦ threshold), the block image has hardly changed, so that the wavelet transform coding is skipped and the block image 115 in the frame memory 14 is decoded. What is necessary is just to hold | maintain as an image as it is. Therefore, the operation will be described in detail below for the former case. The threshold value may be determined in advance before the encoding means. Also, the encoding identification information 114 is sent from the encoding mode control unit.

  The encoding target area extraction unit 3 receives the lock image 101 from the block division unit and the control information 103 from the encoding mode control unit 2. The encoding target area extraction unit 3 extracts a block image 101 that is a target of the wavelet transform in the subsequent stage. The region from which the block image 101 is extracted at this time varies depending on the wavelet transform filtering (convolution operation) method described in the following embodiment. The information on the filtering method of the wavelet transform is the control information 103 from the encoding mode control unit 2.

  The block image 104 to be encoded extracted by the encoding target region extracting unit 3 is sent to the wavelet transform unit 4, and wavelet transform filtering (convolution operation) is performed by the wavelet transform unit 4. The transform coefficient 105 obtained as a result is input to the scanning unit 5, where the wavelet transform coefficient is scanned (scanning). For example, here, it is assumed that wavelet transform coefficients are scanned from left to right (horizontal direction) and from top to bottom (vertical direction).

  The scanned transform coefficient 106 is quantized by the quantizing unit 6 and a quantized coefficient 107 is output. Here, as the quantization means, for example, normally used scalar quantization may be used as shown in Equation 1 below.

Q = x / Δ (Formula 1)
In Equation 1, x is a wavelet transform coefficient value, and Δ is a quantization index value.

  The quantized coefficient 107 from the quantization unit 6 obtained by the scalar quantization or the like is entropy-coded in the entropy coding unit 13 and an encoded bit stream 116 is output. In addition, as means used in the entropy coding unit 13, there are arithmetic coding means in addition to variable length coding means, and research results of these entropy coding means have been reported by each research institution. They can be used.

  In the local decoding (local decoding) loop which is the other path, the quantized coefficient 107 is dequantized again and the transform coefficient 108 is output. Further, through the inverse scanning unit 8, the original two-dimensional array conversion coefficient 109 is obtained. Next, a coefficient to be decoded is extracted from the same transform coefficient 109, and the extracted transform coefficient 110 is subjected to inverse wavelet transform to output a decoded image 111. The specific operation of the decoding target coefficient extraction unit 9 will be described in an embodiment described later. Finally, the region image 112 extracted from the decoded image 111 is written into the frame memory 14.

  The above is the basic configuration and operation of the moving picture coding apparatus according to the first embodiment.

  Here, the general wavelet transform coding means is configured to include the wavelet transform unit 4, the scanning unit 5, the quantization unit 6, and the entropy coding unit 13 of FIG. The wavelet inverse transform decoding means includes the inverse quantization unit 7, the inverse scanning unit 8, and the wavelet inverse transform unit 10 shown in FIG.

  Next, examples of specific configurations of the wavelet transform unit 2 and the wavelet inverse transform unit 10 in the first embodiment of the present invention will be described with reference to the drawings.

  First, FIG. 2 is given as a configuration of a normal wavelet transform unit. This shows a configuration example in which octave division, which is the most general wavelet transform among several methods, is performed over a plurality of levels. In the case of FIG. 2, the number of levels is 3 (level 1 to level 3), the image signal is divided into low and high frequencies, and only the low frequency components are divided hierarchically. Yes. In FIG. 2, wavelet transform for a one-dimensional signal (for example, a horizontal component of an image) is illustrated for convenience, but it can be dealt with a two-dimensional image signal by extending this to two dimensions.

Next, the operation will be described.
The input image signal 120 to the wavelet transform unit shown in FIG. 2 is obtained by first performing band division by the low-pass filter 21 (transfer function H ₀ (z)) and the high-pass filter 22 (transfer function H ₁ (z)). The low-frequency component and the high-frequency component are each thinned out by a half (level 1) by the corresponding down-samplers 23a and 23b. At this time, there are two outputs, an L component 121 and an H component 126. Here, L is Low and indicates a low frequency, and H is High and indicates a high frequency. The low-pass filter 21, the high-pass filter 22, and the two down samplers 23a and 23b shown in FIG.

  Of the signals thinned out by the down-samplers 23a and 23b, only the low-frequency component, that is, the signal from the down-sampler 23a, is further band-divided by the low-pass filter and high-pass filter of the level 2 circuit unit 202, The corresponding downsampler reduces the resolution by half each (level 2). As the circuit unit 202 including the level 2 low-pass filter, high-pass filter, and down-sampler, the same configuration as the circuit unit 201 including the level 1 low-pass filter 21, high-pass filter 22, and down-samplers 23a and 23b is used. .

  By performing such processing to a predetermined level, band components obtained by hierarchically dividing the low frequency component into bands are sequentially generated. The band components generated at level 2 are the LL component 122 and the LH component 125. FIG. 2 shows an example in which the band is divided up to level 3, and the output from the downsampler on the low pass filter side of the circuit unit 202 of level 2 is sent to the circuit unit 203 of level 3 having the same configuration as the circuit unit 201 described above. Have been supplied. As a result of the band division to level 3, the LLL component 123, the LLH component 124, the LH component 125, and the H component 126 are generated.

  Next, a specific configuration example of the wavelet inverse transform unit that performs the reverse operation to the wavelet transform unit shown in FIG. 2 will be described with reference to FIG. The wavelet inverse transform unit in FIG. 3 can be used as the wavelet inverse transform unit 10 in FIG.

  When the band components 123, 124, 125, and 126 that are the outputs of the wavelet transform unit 2 described in FIG. 2 are input to the wavelet inverse transform unit in FIG. 3, first, the LLL component 123 and the LLH component 124 are Each of the upsamplers 24a and 24b is upsampled to double the resolution. Subsequently, the low-frequency component is filtered by the low-pass filter 25 and the high-frequency component is filtered by the high-pass filter 26, and the adder 27 combines both band components. By the circuit unit 206 so far, the inverse conversion as the inverse process of the conversion in the level 3 circuit unit 203 of FIG. 2 is completed, and the LL component 127 which is the band component on the low frequency side of the level 2 is obtained. It is done. By repeating this process to level 1 thereafter, the final decoded image 129 after inverse transformation is output. That is, the level 2 circuit unit 207 and the level 1 circuit unit 208 have the same configuration as the level 3 circuit unit 206, and the output of the level 3 circuit unit 206 is lower than that of the level 2 circuit unit 207. And the output of the level 2 circuit unit 207 is sent as the low frequency side input of the level 1 circuit unit 208, respectively. The above is the basic configuration of a normal wavelet inverse transform unit.

  Here, FIG. 4 illustrates band components obtained as a result of band-dividing the two-dimensional image up to level 2. FIG. The notation of L and H in FIG. 4 is different from the notation of L and H in FIGS. 2 and 3 that deal with one-dimensional signals. That is, in FIG. 4, first, the four components LL, LH, HL, and HH are divided by level 1 band division (horizontal and vertical directions). Here, LL means that both horizontal and vertical components are L. LH means that the horizontal component is H and the vertical component is L. Next, the LL component is band-divided again to generate LLLL, LLHL, LLLH, and LLHH. In this way, in addition to dividing the low-frequency component hierarchically, the entire band is also divided equally.

  Next, a moving picture coding apparatus according to the second embodiment of the present invention will be described. The overall configuration of the second embodiment is the same as that of the moving picture encoding apparatus shown in FIG. 1, but is more specific to the encoding target area extracting unit 3 shown in FIG. Prior to the description of the encoding target region extraction unit and the wavelet transform operation related thereto, when performing wavelet transform for each block, filtering is performed on the surrounding pixels of the block by the tap length of the filter. Will be described. That is, filtering is performed while overlapping with adjacent blocks.

FIG. 5 shows a block _RT to be encoded and a range R _F covered by filtering when performing an overlap type block-based wavelet transform. In FIG. 5, a, b, c, d, e, f, h, i, j, k, l, and m all represent pixels. For example, when the pixel c is filtered in the horizontal direction, three pixels d, e, and f are read out from the block image on the right side, and a predetermined filter coefficient is convolved with these. Similarly, for example, when the pixel j is filtered in the vertical direction, three pixels k, l, and m are read out from the lower block image, and a predetermined filter coefficient is convolved with them.

  Therefore, in the encoding target area extraction unit in the second embodiment, the encoding target area is expanded by reading out pixels of adjacent blocks to the extent that the influence of the filtering by the wavelet transform unit extends around the block. To do. The above is the description of the operation of the wavelet transform in the encoding target region extracting unit 3 and the wavelet transform unit 4 according to the second embodiment.

Next, a third embodiment of the present invention will be described.
The overall configuration of the third embodiment is the same as that of the moving picture encoding apparatus shown in FIG. 1, but the encoding target area extracting unit 3 of FIG. 1 is more specific. In the third embodiment, as already described in the second embodiment, it is necessary to prepare pixels to be encoded around the block to the extent that the influence of filtering by the wavelet transform unit is affected. . In the second embodiment, the means for directly reading out the pixels of the adjacent block is used. However, in the third embodiment, the pixel values in the block are symmetrically extended so as to have a mirror image relationship at the block boundary. , Filtering. That is, in the third embodiment, no overlapping region is provided between images of adjacent blocks, and convolution is performed by extending the wavelet transform coefficient inside the block symmetrically within the range covered by filtering outside the block. I have to. This is shown in FIGS. 6 and 7, where FIG. 6 shows the process of performing wavelet division while subjecting the original image to symmetrical extension convolution, and FIG. 7 shows a specific example around the block. Yes.

  First, FIG. 6 is a diagram for explaining the concept of wavelet coding by symmetric convolution. The original image shown in (A) of FIG. 6 is divided into block images as shown in (B), and then, outside the block, until the filter processing indicated by the broken line (C) reaches each block image. Perform symmetric expansion of pixels in the region.

FIG. 7 specifically illustrates this target extension. The horizontal c, b, a pixel columns in the region _RT of the block to be encoded or decoded are separated from each other by the block boundary. It can be seen that, in a symmetrical order of a, b, c, the filtering range is extended to R _F. Similarly, in the vertical direction, the pixel columns of f, e, and d in the block region _RT are symmetrically extended to the range R _{F in} which d, e, and f are filtered, with respect to the block boundary. . It is known that wavelet transform coefficients are generated only by the same number as the number of pixels in the block image if such mirror image relational expansion is performed. That is, there is an advantage that there is no redundancy.

  Subsequently, each symmetrically expanded block in FIG. 6C is subjected to WT (wavelet transform). As a result, as already described in FIG. 4, for example, it is divided into four band components (see FIG. 6D). The hatched portion in FIG. 6D is the above-described low-frequency LL component. Further, the block of the low-frequency component (LL) in the hatched portion is similarly subjected to symmetrical expansion and subjected to wavelet transform (WT) as shown in FIG. Thereafter, the same operation is repeated up to a predetermined number of wavelet divisions. The above is the description of the operation of the wavelet transform with symmetric expansion for each block in the encoding target region extracting unit and the wavelet transform unit of the third embodiment.

Next, a fourth embodiment of the present invention will be described.
This fourth embodiment shows a specific example of the encoding target area extracting means and the wavelet transform means linked thereto in the moving picture encoding apparatus having the basic configuration shown in FIG. 1 described above. In the fourth embodiment, as the wavelet transform unit, a method of performing a convolution operation with all pixels within the range covered by filtering outside the block as 0 is used. In this case as well, it is not necessary to use pixels of adjacent blocks on the original screen. Next, the operation will be described.

FIG. 8 is a diagram illustrating a range covered by filtering outside the block. In FIG. 8, the outside of the block to be encoded or decoded _RT is c, b, a, 0, 0, 0 by setting all the ranges R _F covered by horizontal filtering to 0. By setting all the ranges R _F covered by vertical filtering to 0, f, e, d, 0, 0, 0 are obtained. A convolution operation is performed on the coefficients of the wavelet filter.

Next, a fifth embodiment of the present invention will be described.
In this fifth embodiment, as a wavelet transform means, the pixels within the range covered by filtering outside the block are symmetrically expanded so as to have a point-symmetric relationship with the pixel values on the block boundary. The encoding target pixel is expanded. FIG. 9 illustrates this, and shows two examples (a) and (b) in FIG. In the figure, eight sample points X [0], X [1], X [2], X [3], X [4], X [5], X [6], X [7] are blocks The inner pixels (one-dimensional direction only) are shown.

  On the other hand, the pixel at the sample point indicated by the dotted line is calculated using the pixel value X [0] or X [7] as a point-symmetric reference value. In the case of FIG. 9A, the point Pa indicating the pixel value at the sample position of the sample point X [0] on the block boundary is set as a point symmetrical reference point. For example, the sample point x [1] outside the block is used. Is obtained by calculating equidistant points on an extension line extending from the sample point X [1] having a point-symmetric positional relationship within the block so as to pass through the reference point Pa. That is, with the reference point Pa as the center, the sample point x [1] is located symmetrically with respect to the sample point X [1]. Similarly, with the reference point Pa as the center, the sample point x [2] is in a point-symmetrical position with respect to the sample point X [2], and the sample point x [3] is in a point-symmetrical position with respect to the sample point X [3]. is there. The same applies when the pixel value at the sample point of X [7] at the other boundary of the block is used as a point symmetry point.

  Next, the case of FIG. 9B will be described. The difference from FIG. 9A is that the reference point Pb, which is the center of point symmetry, is shifted from the sample position by a distance corresponding to a half sample. That is, a point Pb indicating a pixel value equal to the sample point X [0] at a position obtained by shifting the sample position of the sample point X [0] on the block boundary by a half sample distance outside the block It is said. Accordingly, the sample point x [0] at a position one sample away from the sample point X [0] outside the block is located symmetrically with respect to the sample point X [0] around the reference point Pb. It is the same value as the sample point X [0]. As a result, the values inside the block boundary (for example, X [0]) and the external point (for example, x [0]) are the same, and the connection at the block boundary can be made smooth.

  As described above, by performing pixel expansion using the point symmetry relationship by any one of the methods shown in FIGS. 9A and 9B, the pixel values in the block can be blocked without using the pixel values of the adjacent blocks. It is possible to perform a convolution operation by expanding to a range where filtering by wavelet transform is performed externally.

Next, a sixth embodiment of the present invention will be described.
In the sixth embodiment, in particular, the conversion coefficient extraction unit 9 to be decoded, the wavelet inverse conversion unit 10 linked thereto, and the decoded image region extraction of the basic configuration of the moving image encoding apparatus shown in FIG. 1 described above. Part 11 is embodied. In FIG. 1, the quantization coefficient 107 is inversely quantized by the inverse quantization unit 7 to obtain a transform coefficient 108. The transform coefficient 108 is reverse-scanned, and from the transform coefficients 109 rearranged again in the form of the two-dimensional image array, the decoding target coefficient extraction unit 9 extracts only those to be decoded.

  The extracted coefficient 110 is then subjected to inverse wavelet transform, and the decoded image 111 is output from the inverse wavelet transform unit 10. In addition, a region image to be decoded is extracted from the decoded image 111 and written as 112 in the frame memory 14. The above is the description of the operation of the local decoding means inside the encoding apparatus according to the sixth embodiment.

Next, a seventh embodiment of the present invention will be described.
The seventh embodiment corresponds to a modification of the sixth embodiment. In the seventh embodiment, wavelet inverse transform filtering is performed outside the block as the decoding target coefficient extraction unit. At this time, the conversion coefficient of the adjacent block is read out to the range covered by the filter. That is, this corresponds to the case where a, b, c, d, e, and f are regarded as conversion coefficients in FIG. Then, the wavelet inverse transform is performed by performing a convolution operation on the read conversion coefficient outside the block and the conversion coefficient of the block. Then, only a region to be decoded may be extracted from the decoded image restored by the inverse wavelet transform in the subsequent decoded image region extraction unit.

Next, an eighth embodiment of the present invention will be described.
The eighth embodiment corresponds to a modification of the sixth embodiment. In the eighth embodiment, wavelet inverse transform filtering is performed outside the block as the decoding target coefficient extraction unit. In this case, the coefficients are symmetrically extended so as to have a mirror image relation at the block boundary up to the range covered by the filter. In the decoding target coefficient extraction unit, the wavelet inverse transformation unit performs a convolution operation of wavelet inverse transformation including the expanded transformation coefficient.

  FIG. 10 illustrates this series of operations, and shows a process of performing wavelet inverse transform by symmetrically expanding low-frequency and high-frequency transform coefficient components so as to have mirror image relationships.

  In FIG. 10A, the wavelet transform coefficients of the four band components LL, LH, HL, and HH are symmetrically expanded as shown in FIG. 10B, and each is subjected to wavelet inverse transform (IWT). As a result, each block image is decoded and output as shown in FIG. For the symmetric extension, the method of the third embodiment already described in FIG. 7 may be used. However, in the description of the wavelet transform described in the third embodiment, a, b, c,... In FIG. 7 indicate pixels. However, in the current wavelet inverse transform, these are wavelet transform coefficients. Note that it means.

  An advantage of using this symmetrically extended wavelet inverse transform is that it can perform the inverse transform and decoding completely independently of the neighboring surrounding blocks. In the case of a relatively high bit rate (low compression rate), almost no deterioration of the boundary portion between blocks can be detected.

Next, a ninth embodiment of the present invention will be described.
In the ninth embodiment, when the wavelet transform filtering is performed outside the block as the decoding target coefficient extraction unit, all the wavelet transform coefficients within the range covered by the filtering outside the block are set to 0, and the convolution operation is performed. To do. Next, the operation will be described.

  FIG. 11 is a diagram for specifically explaining this. For example, while the wavelet transform coefficients of the four band components LL, LH, HL, and HH shown in FIG. 11A are left as they are, the range in which filtering is performed outside the block as shown in FIG. All wavelet transform coefficients in the (dotted line region) are set to 0, and each block is subjected to inverse wavelet transform (IWT). Thus, each block image is decoded and output. The convolution operation of the wavelet inverse transformation at this time is called a linear convolution.

  The linear convolution means is the same as that described with reference to FIG. As an advantage of this linear convolution, there is an excellent feature that a block boundary portion is not detected even under a low bit rate (high compression). This is due to the fact that by making the coefficient of the boundary between adjacent blocks uniform with zero, the effect of linear interpolation by wavelet filtering is produced at the boundary and smooth connection is established.

  Further, even with this wavelet inverse transform means, it is not necessary to read the overlapped wavelet transform coefficients of adjacent blocks, so that it is possible to perform inverse transform and decoding completely independently of adjacent neighboring blocks.

  Therefore, as the wavelet inverse transform means, the wavelet transform coefficients within the range covered by the filtering outside the block are all set to 0 so that the convolution operation is performed, which realizes high image quality and partial block decoding under high compression. It is an excellent combination of both.

Next, a tenth embodiment of the present invention will be described.
The tenth embodiment corresponds to a modification of the sixth embodiment, and in the tenth embodiment, the wavelet inverse transform is filtered outside the block as the decoding target coefficient extraction unit. In this case, the transform coefficient outside the block is expanded so that it is point-symmetric across the block boundary up to the range covered by the filter, and the wavelet inverse transform is performed including the obtained transform coefficient and the transform coefficient inside the block. It is. The specific expansion method is as shown in FIG. 9, but the pixel value is expanded in FIG. 9, but the conversion coefficient value is expanded in the present embodiment. The subsequent operation is the same as that in the eighth embodiment.

Next, an eleventh embodiment of the present invention will be described.
This embodiment is an embodiment of an image wavelet decoding apparatus and method to which an encoded bitstream obtained by performing wavelet encoding of an image as described above is supplied.

  FIG. 12 is a block diagram showing a schematic configuration of the moving picture decoding apparatus according to the eleventh embodiment, corresponding to the moving picture encoding apparatus shown in FIG.

  In FIG. 11, an image wavelet decoding apparatus is obtained by an entropy decoding unit 16 that inputs or reads an encoded bitstream and entropy decodes, and an inverse quantization unit 7 that inversely quantizes the obtained quantization coefficient. The transform coefficient inverse scanning unit 8 that reverse-scans the transformed coefficients and returns them to the order of the original coefficients, the decoding target coefficient extraction unit 9 that extracts the decoding target coefficients of the coefficients after the inverse scanning, and the obtained decoding target coefficients A wavelet inverse transform unit 10 that generates a block image by inverse wavelet transform, a decoded image region extraction unit 11 that extracts a decoded image region, and further includes a decoding mode control unit 15 and a frame memory 14. Is done. Next, the operation will be described.

  In FIG. 11, the decoding mode control unit 15 that has received the encoding mode information 104 skips decoding and outputs the block image stored and held in the frame memory 14 as the output 119 as it is, or performs decoding. Select. When the decoding is skipped, the decoding mode control unit 15 transmits a control signal 118 for outputting the block image stored and held as it is to the frame memory 14. When performing the latter decoding, a control signal is issued from the decoding mode control unit 15 to the entropy decoding unit 16, and the subsequent operations are performed.

  That is, when performing the above decoding, first, the entropy decoding unit 16 entropy decodes the encoded bit stream 116 to reproduce the quantization coefficient 113. The quantization coefficient 113 is sent to the inverse quantization unit 7 and inversely quantized to become a transform coefficient 108. The transform coefficient 108 is sent to the inverse scanning unit 8 and subjected to scanning in the reverse direction to become the original two-dimensional array wavelet transform coefficient 109. A coefficient to be decoded is extracted from the transform coefficient 109 by the decoding target coefficient extraction unit 9, and the extracted conversion coefficient 110 is sent to the wavelet inverse transform unit 10, subjected to inverse wavelet transform, and a decoded image 111 is obtained. can get. An area image 112 extracted by the decoded image area extraction unit 11 from the decoded image 111 is written into the frame memory 14. Needless to say, the entropy decoding means must be integrated with the means used in the entropy encoding means in the encoding apparatus.

Next, a twelfth embodiment of the present invention will be described.
The twelfth embodiment embodies the decoding target coefficient extraction unit 9 in the eleventh embodiment, and transform coefficients up to a range affected by filtering outside the block at the time of wavelet inverse transform. Is obtained by applying the method of any of the second to fifth embodiments described above to the conversion coefficient instead of the pixel. That is, transform coefficients outside the block are obtained by reading from adjacent blocks, expanding symmetrically with respect to the block boundary, setting it as 0, or expanding symmetrically with respect to the block boundary as the center, and wavelet inverse Conversion is performed. This needs to be linked to the operation of the encoded image area extracting unit 3 of the moving image encoding apparatus of FIG. Further, the decoded image area extraction unit 11 which is a component of the moving picture decoding apparatus needs to have means for extracting the target block area from the decoded image restored by the wavelet inverse transformation means.

Next, a thirteenth embodiment of the present invention will be described.
In the thirteenth embodiment, all the encoding target blocks are encoded by wavelet transform encoding once in a plurality of frames in the moving picture encoding apparatus of FIG. That is, as described above, the moving image encoding apparatus in FIG. 1 encodes an image in units of blocks, and stores and holds it in the frame memory when there are few changes in the pixels in the block, such as movement of the subject. The block image located at the same position in the previous encoded frame is used as a decoded image as it is, and on the other hand, when the change is large, the subsequent wavelet transform encoding means is performed.

  However, if the wavelet transform coding means is applied to all blocks for every predetermined number of frames, error accumulation is eliminated and a high-quality decoded image can always be provided. Therefore, in the thirteenth embodiment, all the encoding target blocks are subjected to wavelet transform coding for each predetermined number of frames by a predetermined means or in the form of external input.

Next, a fourteenth embodiment of the present invention will be described.
The fourteenth embodiment includes means for converting an input image into a plurality of bands by means of filter bank means comprising a plurality of low-pass filters and high-pass filters, and conversion for each of the bands. An image compression apparatus having means for recording a coded bitstream on a recording medium by means for scanning, quantizing, and entropy coding coefficients, and reading the coded bitstream from the recording medium for entropy decoding, An image compression / decompression apparatus comprising quantization, reverse scanning, and an image decompression apparatus having means for outputting a decoded image by performing inverse transformation by the filter bank means, in accordance with the search speed. Search means for reading out the corresponding band component from the encoded bit stream recorded on the recording medium is provided.

  FIG. 13 is a block diagram showing an image recording / reproducing apparatus as the fourteenth embodiment. The apparatus shown in FIG. 13 includes an external input control unit 38, a CCD unit 39, an AD conversion unit 30, and an encoding unit. 31, a decoding unit 37, a recording medium control unit 34, a writing unit 32, a reading unit 33, a recording medium 40, and a display unit 35. Next, the operation will be described.

  The encoding unit 31 is a part that performs the wavelet transform encoding unit already described in the embodiment, and includes an image wavelet band dividing unit, a transform coefficient scanning unit, a quantization unit, and an entropy encoding unit. . On the other hand, the decoding unit 27 includes entropy decoding means, inverse quantization means, inverse scanning means, and wavelet band synthesis means.

  As shown in FIG. 13, the image input units and the recording units on the recording medium are connected to the encoding unit 31 and the decoding unit 37. Normally, the image is input as an optical signal 131 to a CCD unit 39 such as a camera, and the CCD signal 132 is AD converted by the AD conversion unit 30 and converted into a digital image signal 133. The digital image signal 133 may be a single signal in addition to the case where it is output as separate signals of R, G, B, for example, in accordance with the characteristics of the CCD.

  The digital image signal 133 is subjected to wavelet transform encoding processing by the encoding unit 31 to generate an encoded bit stream 134. The encoded bit stream 134 is sent to the decoding unit 37, and immediately after being decoded by the same unit 37, the decoded image 135 is output via the display unit 35 and the display image 136 is output. On the other hand, the encoded bit stream 134 is sent to the writing unit 32, is sent to the recording medium control unit 34 as a recording encoded bit stream 141 together with the write control signal, and is multiplexed with the control information to become a recording signal. It is written on the recording medium 40. Between the recording medium 40 and the recording medium control unit 34, recording and reproduction of the signal 139 according to the recording format of the recording medium 40 is performed. Here, as the recording medium 40, an optical disk such as a CD-ROM, CD-R, DVD-ROM, etc. can be used in addition to a magnetic tape.

  On the other hand, a case where a high-speed search is performed by reading out an encoded bit stream of an image recorded on the recording medium 40, for example, will be described below. The external signal 137 holds, for example, search speed information. This information is decoded by the external input control unit 38 and the control signal 138 is transmitted to the recording medium control unit 34. From the recording medium 40 to the recording medium control unit 34, the drive system of the recording medium 40 operates in accordance with the search speed, and an encoded bit stream signal 139 according to the recording format of the recording medium 40 is read and recorded. An encoded bit stream 140 (including additional information and header information) is sent to the reading unit 33 via the medium control unit 34. The pure encoded bit stream 142 (without additional information or header information) output from the reading unit 33 is decoded by the decoding unit 37, and the decoded image 135 is transmitted from the same unit as described above. The above is the operation at the time of search.

  Next, reading of an encoded bit stream at the time of search will be described. In the present embodiment, the coded bit stream recorded on the recording medium 40 is used in order to take advantage of the wavelet coding means that hierarchically divides the image band as the coding means in the coding section. Therefore, a means for selecting and reading out the one corresponding to the predetermined band component in accordance with the search speed is used. That is, the corresponding band component is read from the encoded bit stream recorded on the recording medium 40 in accordance with the search speed. Specifically, for example, when the search speed is high, the low band component Only the encoded bit stream corresponding to the band of the low frequency component is read and the band image of the low frequency component is decoded. Conversely, when the search speed is low, the encoded bit stream corresponding to the band of the low frequency band and the high frequency component is decoded. And the band image is decoded. More detailed configurations and operations will be described in the following embodiments.

Next, a fifteenth embodiment of the present invention is described.
The apparatus of the fifteenth embodiment decodes a low-frequency component band image when the above-described search speed is high, and decodes a low-frequency and high-frequency component band image when the search speed is low. The overall configuration shown in FIG. 13 can be used.

  Usually, since it is necessary to reproduce many image frames in a predetermined time at the time of search, it is difficult in terms of hardware to read and decode all the encoded bit streams. Therefore, ingenuity is usually taken such as skipping frames at certain intervals or decoding only a part of the image.

  However, the fifteenth embodiment implements a search means that decodes all frames and does not impair the overall image impression. That is, as already described with reference to FIG. 4, using wavelet encoding means, it is possible to easily obtain a decoded image from only a low frequency component (for example, the LLLL component). In this case or when the reading speed of the encoded bit stream cannot be sufficiently secured, only the encoded bit stream corresponding to the low-frequency component band is read and the decoded image can be generated by the decoding unit 37 in FIG. I can do it.

  On the other hand, in the case of low-speed search or when the reading speed of the encoded bit stream can be sufficiently secured, the encoded bit stream corresponding to the band of the middle / high frequency component is read in addition to the low frequency component, and the decoding unit 37 A decoded image can be generated. By the above means, decoding using one or a plurality of optimum bands can be realized according to the search speed or according to the read speed of the encoded bit stream.

  In this case, the decoded image may be decoded with the size of the predetermined band component in the wavelet transform described above. For example, as described above, when only the encoded bit stream corresponding to the LLLL component is read and decoded, the size of the image included in the band component of the LLLL is, as is apparent from FIG. The horizontal and vertical are 1/4. Therefore, the display image 136 displays the original 1/4 decoded image (1/4 in length ratio and 1/16 in area ratio).

  On the other hand, as another method, the decoded image may be generated as a decoded image having the same size as the original image and using a predetermined band component. In this case, since the decoded image using only the LLLL component does not have a high frequency component, it has a slightly blurred image quality with reduced resolution. However, since most of the energy of the image is concentrated in the low-frequency component, the rough features of the image can be sufficiently grasped.

Next, a sixteenth embodiment of the present invention will be described.
The fourteenth and fifteenth embodiments are encoding / decoding devices including an image input unit and a display unit, whereas the sixteenth embodiment is a recording / recording unit as shown in FIG. This is a configuration of an apparatus including only a part for reading and decoding the medium 40. In other words, it is a device suitable for a receiver or a portable player without an encoder. Next, the configuration and operation will be described.

  The apparatus shown in FIG. 14 includes a recording medium 40, a control unit 45, an encoded bitstream reading unit 46, and a decoding unit 37. In accordance with the external input information 137, the encoded bitstream reading unit 46 reads the encoded bitstream 148 including additional information and header information from the recording medium 40 according to the control information 145 and 147 from the control unit 45, and A pure encoded bit stream 149 that does not include header information or the like is output. The pure encoded bit stream 149 is decoded by the decoding unit 37, and a decoded image 130 is generated and output. Since the operation of the decoding unit 37 is the same as that described in the fourteenth and fifteenth embodiments, description thereof is omitted.

Next, a seventeenth embodiment of the present invention will be described.
The seventeenth embodiment is a moving image encoding apparatus when an input image is an interlaced image.

The TV broadcast that we are currently familiar with is an interlaced image using interlaced scanning, which consists of odd and even fields. FIG. 15 illustrates this. In the figure, lines La ₁ , La ₂ ,... Represent odd fields, and hatched lines Lb ₁ , Lb ₂ ,.

  In an interlaced image, odd fields and even fields are alternately arranged for each line, as in the left frame structure of FIG. 15, and the efficiency may be reduced if these are encoded as frame images. is there. This is shown in FIG. As shown in FIG. 16, in the case of a frame image of an interlaced image, if the movement of the subject is large, there is a time difference between the odd field and the even field, so the left frame structure in FIG. In this way, there appears to be a shift between the odd field and even field images. This leads to a reduction in efficiency when performing subsequent encoding. On the other hand, by dividing the structure into the field structure on the right side in FIG.

  Next, the actual operation will be described with reference to FIG. The interlaced image 100 of the input image is first divided into an odd field image 155 and an even field image 156 by the field dividing unit 56. The odd field image 155 and the even field image 156 are encoded by the encoding unit 57, and an encoded bit stream 157 and encoding mode information 158 are transmitted.

  Here, for the specific operation in the encoding unit 57, the encoding configuration means described in the first embodiment may be used as it is. That is, as described with reference to FIG. 1, it is only necessary to include each part from the block dividing unit 1 to the entropy encoding unit 13. With the above configuration, it is possible to realize a moving image encoding means with high encoding efficiency without reducing the efficiency even with interlaced images.

Next, a seventeenth embodiment of the present invention will be described.
The seventeenth embodiment corresponds to a decoding device corresponding to the encoding device of the seventeenth embodiment. Next, the operation will be described with reference to FIG. The decoding unit 58 that has received the encoded bitstream 157 and the encoding mode information 158 outputs a decoded image 151 of an odd field and a decoded image 152 of an even field using decoding means. Then, the field synthesis unit 59 synthesizes the decoded image 151 of the odd field and the decoded image 152 of the even field to synthesize the composite image 153 of both fields, that is, the frame shown in FIG. 15 or FIG. Return to the structure image.

  As a specific configuration of the decoding unit 58, the decoding circuit configuration described in the eleventh embodiment may be used as it is. That is, each component from the entropy decoding unit 16 to the frame memory 14 in FIG. 12 is used as decoding means.

  With the above configuration, the encoded bit stream of the interlaced image can be decoded for each field, a decoded image can be generated for each field, and the field image can be combined with the frame image for generation.

  Specific application examples of the embodiment of the present invention include an electronic camera, a portable / mobile image transmission / reception terminal (PDA), a printer, a satellite image, a compression / decompression unit for medical images, or a software module thereof, a game, Examples include a texture compression / decompression apparatus used in three-dimensional CG or a software module thereof.

  According to the embodiment of the present invention, an input image is divided into a plurality of blocks, and an error between the block image at the same position of the previous frame stored and held in the image memory and the current encoding target block is calculated. Detects and selects whether or not to encode the block to be encoded according to the detected error, and performs wavelet transform coding for each block image. Since simple prediction means in units of blocks is used without using the prediction means, the amount of processing is very small. Here, since the background portion hardly moves in a normal moving image, the present invention works effectively in such a portion, and there is an effect that the compression efficiency is increased.

  In addition, since the motion compensation means can be greatly simplified, the motion compensation means is conventionally realized by dedicated hardware (LSI), but can be realized by software in the present invention.

  Also, using block-based coding means, and for the pixels where the filter protrudes from the block when wavelet transform coding is used, the block boundary distortion is not noticeable, so smooth connection surface decoding is always performed. There is also an effect that an image can be obtained.

  Further, according to the image recording / reproducing apparatus of the embodiment of the present invention, means for converting an input image into a plurality of bands by a filter bank unit comprising a plurality of low-pass filters and a high-pass filter; An image compression apparatus comprising: means for scanning, quantizing, and entropy-encoding the transform coefficient for each band; and means for recording the encoded bitstream on the recording medium; and the encoded bitstream from the recording medium. Read out, entropy decoding, inverse quantization, inverse scanning, further performing inverse transformation by the filter bank means, and an image decompression device provided with means for outputting a decoded image, and according to the search speed, Since there is a search means for reading out the corresponding band component from the encoded bit stream recorded on the recording medium, the encoded bit recorded on the recording medium is provided. It reads the stream, when performing a search of the decoded image can be read only optimum band components tailored to search speed can be decoded to obtain a consistently high-quality decoded image without frame dropping.

  In addition to the size of the original image as well as the resolution of the band component, the decoded image can also have the resolution of the band component, so that the variation of the decoded image increases, so that it can be used properly according to the application.

  Therefore, according to the embodiment of the present invention, both high image quality under high compression and partial block decoding are achieved, and the wavelet transform unit and the wavelet inverse transform unit can be matched. Efficient encoding and high-quality decoding are possible.

  Furthermore, in the case of an interlaced image, it is separated into an odd-numbered field image and an even-numbered field image and encoded separately, so that even when a subject with large motion exists in the image, a high coding efficiency is always obtained. Can be maintained.

  It should be noted that the present invention is not limited to the above-described embodiments, and various modifications can be made without departing from the scope of the present invention.

It is a block diagram which shows schematic structure of the moving image encoder used for description of this invention. It is a block diagram which shows schematic structure (until level 3) of a normal wavelet transformation part. It is a block diagram which shows schematic structure (until level 3) of a normal wavelet inverse transformation part. It is a figure for demonstrating the zone division (division level = 2) of a two-dimensional image. It is a figure for demonstrating the convolution calculation in the case of overlap type wavelet encoding. It is a figure for demonstrating the concept of the wavelet encoding which performs the symmetrical convolution operation of a pixel. It is a figure for demonstrating the concept of a symmetrical convolution operation. It is a figure for demonstrating the concept of a linear convolution calculation. It is a figure for demonstrating the concept of a point symmetrical convolution operation. It is a figure for demonstrating the concept of the wavelet decoding which performs the symmetrical convolution operation of a wavelet transform coefficient. It is a figure for demonstrating the concept of the wavelet decoding which performs the linear convolution operation of a wavelet transform coefficient. It is a block diagram which shows schematic structure of the moving image decoding apparatus with which it uses for description of this invention. It is a block diagram which shows schematic structure of the image recording / reproducing apparatus with which it uses for description of this invention. It is a block diagram which shows schematic structure of the image decoding apparatus with which it uses for description of this invention. It is a figure for demonstrating the frame image and field image in the case of an interlaced image. It is a figure for demonstrating the frame image and field image in the case of an interlaced image. It is a block diagram which shows schematic structure of the image coding apparatus for coding an interlaced image. It is a block diagram which shows schematic structure of the image decoding apparatus for decoding an interlaced image.

Explanation of symbols

  1 block division unit, 2 encoding mode control unit, 4 wavelet transform unit, 5 scanning unit, 46 quantization unit, 7 inverse quantization unit, 8 inverse scanning unit, 9 decoding target coefficient extraction unit, 10 wavelet inverse transform unit, DESCRIPTION OF SYMBOLS 11 Decoding image area extraction part, 12 Error detection part, 13 Entropy encoding part, 14 Frame memory, 15, Decoding mode control part, 16 Entropy decoding part, 21 Low pass filter for analysis, 22 High pass filter for analysis, 23a, 23b 2 1 / down sampler, 24a, 24b 2x upsampler, 25 synthesis low-pass filter, 26 synthesis high-pass filter, 31, 57 encoding unit, 32 writing unit, 33 reading unit, 34 recording medium control unit, 35 display unit, 37, 57 Decoding unit, 38 external input / control unit, 39 CCD unit, 40 recording medium, 45 control unit, 46 encoded bitstream reading unit, 56 field division unit, 59 field synthesis unit

Claims (34)

Block dividing means for dividing the input image into a plurality of blocks;
Error detection means for detecting an error between the block image at the same position of the previous frame stored and held in the image memory and the current encoding target block;
Encoding mode control means for selectively controlling whether or not to perform encoding of the encoding target block according to the detected error;
And a wavelet transform coding means for performing wavelet transform coding for each block image.   Wavelet inverse transform decoding means for performing wavelet inverse transform decoding on the encoded output obtained by the wavelet transform coding means, and storing and holding the image from the wavelet inverse transform decoding means in the image memory The moving picture coding apparatus according to claim 1, wherein:   The wavelet transform coding means includes a wavelet transform means for performing wavelet transform for each block image, a scanning means for scanning wavelet transform coefficients, a quantizing means for quantizing the coefficients after scanning, and a quantized coefficient. The moving picture coding apparatus according to claim 1, further comprising entropy coding means for entropy coding and outputting a coded bit stream.   An encoding target region extraction unit is provided in front of the wavelet transform unit, which determines a current encoding target block region in accordance with information from the encoding mode control unit and extracts pixels of the encoding target region. The moving picture coding apparatus according to claim 3.   The encoding target area extracting unit expands the encoding target area by reading out pixels of an adjacent block to the extent that the influence of filtering by the wavelet transform unit is around the block. Item 5. The moving image encoding device according to Item 4.   The encoding target area extracting unit symmetrically expands the pixel values inside the block around the block so that the influence of filtering by the wavelet transform unit is affected by the block boundary. 5. The moving picture encoding apparatus according to claim 4, wherein the encoding target area is enlarged.   The encoding target region extracting unit expands the encoding target region by inserting a zero value around the block up to a range affected by filtering by the wavelet transform unit. 4. The moving image encoding apparatus according to 4.   The encoding target area extracting unit symmetrically expands the pixel values in the block around the block so as to have a point-symmetrical relationship at the block boundary to the extent that the filtering by the wavelet transform unit is affected. 5. The moving picture encoding apparatus according to claim 4, wherein the encoding target area is enlarged.   Wavelet inverse transform decoding means for performing wavelet inverse transform decoding on the quantized wavelet transform coefficients from the wavelet transform coding means, and storing and holding the image from the wavelet inverse transform decoding means in the image memory In addition, the wavelet inverse transform decoding means includes an inverse quantization means for inversely quantizing the quantized transform coefficient to return it to a transform coefficient, an inverse scanning means for inversely scanning the transform coefficient, and an obtained transform coefficient. 2. The moving picture coding apparatus according to claim 1, further comprising wavelet inverse transform means for performing inverse wavelet transform.   The decoding object area extracting means for extracting a conversion coefficient to be decoded is provided in the preceding stage of the wavelet inverse transforming means, and the decoded image area extracting means is provided in the subsequent stage of the inverse wavelet inverse transforming means. 9. The moving image encoding device according to 9.   The decoding target coefficient extraction unit expands the coefficient region of the decoding target by reading out the transform coefficient of the adjacent block to the extent that the influence of the filtering by the wavelet inverse transformation unit extends around the block. The moving image encoding apparatus according to claim 10.   The decoding target coefficient extraction unit symmetrically expands the transform coefficient inside the block around the block so that the influence of the filtering by the wavelet inverse transform unit is affected so as to have a mirror image relation at the block boundary. The moving picture encoding apparatus according to claim 10, wherein the coefficient area to be decoded is expanded.   The decoding target coefficient extraction unit expands the coefficient region to be decoded by inserting a zero value around the block up to a range affected by the filtering by the wavelet inverse transformation unit. Item 11. The moving image encoding device according to Item 10.   The decoding target coefficient extraction means symmetrically expands the transform coefficients inside the block around the block so as to have a point-symmetrical relationship at the block boundary to the extent that the filtering by the wavelet inverse transform means is affected. 11. The moving picture encoding apparatus according to claim 10, wherein the coefficient area to be decoded is expanded.   11. The moving picture coding apparatus according to claim 10, wherein the decoded image area extracting unit extracts a target block area from the decoded image restored by the wavelet inverse transformation unit.   2. A moving picture encoding apparatus according to claim 1, wherein means for dividing the interlaced image as the input image into an odd field and an even field is provided before the means for dividing the input image into a plurality of blocks. . Dividing the input image into a plurality of blocks;
Detecting an error between the block image at the same position of the previous frame stored and held in the image memory and the current encoding target block;
Selecting whether or not to perform encoding of the encoding target block according to the detected error; and
And a step of performing wavelet transform coding for each block image.   A step of performing wavelet inverse transform decoding on the encoded output obtained by the step of performing wavelet transform coding, and the image memory stores the image obtained by the step of performing wavelet inverse transform decoding The video encoding method according to claim 17, wherein the video encoding method is held. Wavelet inverse transform decoding means for inputting or reading an encoded bitstream, inverse wavelet transform decoding and outputting a decoded image;
A decoded image region extracting unit that extracts a target region from the decoded image obtained by the wavelet inverse transform decoding unit;
An image memory for storing and holding the obtained decoded image;
Decoding mode control means for selecting whether to send the decoded image of the previous frame stored in advance in the image memory or to send the image decoded by the wavelet inverse transform decoding means in accordance with the supplied encoding control signal A moving picture decoding apparatus comprising:   The wavelet inverse transform decoding means includes: an inverse quantization means for inversely quantizing the quantized transform coefficient to return it to a transform coefficient; an inverse scanning means for inversely scanning the transform coefficient; and an object to be decoded from the obtained transform coefficient. 20. The moving picture decoding apparatus according to claim 19, further comprising: a decoding target coefficient extracting unit that extracts a transform coefficient, and a wavelet inverse converting unit that performs inverse wavelet transform on the obtained decoding target coefficient. .   The decoding target coefficient extraction unit expands the coefficient region of the decoding target by reading out the transform coefficient of the adjacent block to the extent that the influence of the filtering by the wavelet inverse transformation unit extends around the block. The moving picture decoding apparatus according to claim 20.   The decoding target coefficient extraction unit symmetrically expands the transform coefficient inside the block around the block so that the influence of the filtering by the wavelet inverse transform unit is affected so as to have a mirror image relation at the block boundary. 21. The moving picture decoding apparatus according to claim 20, wherein a coefficient area to be decoded is enlarged.   The decoding target coefficient extraction unit expands the coefficient region to be decoded by inserting a zero value around the block up to a range affected by the filtering by the wavelet inverse transformation unit. Item 20. The video decoding device according to Item 20.   The decoding target coefficient extraction means symmetrically expands the transform coefficients inside the block around the block so as to have a point-symmetrical relationship at the block boundary to the extent that the filtering by the wavelet inverse transform means is affected. 21. The moving picture decoding apparatus according to claim 20, wherein the coefficient area to be decoded is enlarged.   21. The moving picture decoding apparatus according to claim 20, wherein the decoded image area extracting unit extracts a target block area from the decoded image restored by the wavelet inverse transformation unit.   The coded bitstream includes an odd field image coded bitstream and an even field image coded bitstream, and these odd field image coded bitstream, even field image coded bitstream, and 20. The moving picture decoding apparatus according to claim 19, wherein the wavelet inverse transform decoding means respectively decodes and synthesizes an odd field decoded image and an even field decoded image. Inputting or reading an encoded bitstream, inverse wavelet transform decoding and outputting a decoded image;
Extracting a target area from the decoded image obtained by the inverse wavelet transform decoding;
Storing and holding the obtained decoded image in an image memory;
Selecting whether to send the decoded image of the previous frame stored in advance in the image memory or the image decoded by the wavelet inverse transform decoding means in accordance with the supplied encoding control signal. A moving picture decoding method characterized by the above.   The wavelet inverse transform decoding is performed by dequantizing the quantized transform coefficient to return it to the transform coefficient, an inverse scanning process for scanning the transform coefficient in reverse, and decoding from the obtained transform coefficient. 28. The moving picture decoding method according to claim 27, further comprising: a decoding target coefficient extracting step for extracting a target transform coefficient, and a wavelet inverse transform step for performing inverse wavelet transform on the obtained decoding target coefficient. Means for transforming an input image into a plurality of bands by means of a filter bank composed of a plurality of low-pass filters and high-pass filters, and scanning, quantization, and entropy coding of transform coefficients for each of the bands Means for recording an encoded bitstream on a recording medium by means for
An image decompression device comprising means for reading an encoded bit stream from the recording medium, performing entropy decoding, inverse quantization, inverse scanning, and performing inverse transformation by the filter bank means, and outputting a decoded image. And
An image recording / reproducing apparatus comprising: search means for reading out a corresponding band component from an encoded bit stream recorded on the recording medium in accordance with a search speed.   When the search speed by the search means is high, only the encoded bit stream corresponding to the low-frequency component band is read, and the low-frequency component band image is decoded by the image decompression device, and the search speed is low. 30. The image recording / reproducing apparatus according to claim 29, wherein the image recording / reproducing apparatus decodes the band image by reading out the encoded bit stream corresponding to the band of the low frequency band and the high frequency band.   31. The image recording / reproducing apparatus according to claim 30, wherein the decoded band image is decoded with a resolution of a predetermined band component.   31. The image recording / reproducing apparatus according to claim 30, wherein the decoded band image has been decoded using a predetermined band component at a resolution of the original image.   The search means reads out a predetermined code word from a recording medium on which an encoded bit stream is recorded, by a search speed from a control means or control from an external input, and outputs a decoded image. Item 29. The image recording / reproducing apparatus according to Item 29. The encoded bit stream is read from the recording medium on which the encoded bit stream obtained by scanning, quantizing, and entropy encoding the transform coefficient for each of a plurality of bands is read, entropy decoding, dequantization, Means for performing reverse scanning, further performing reverse transformation by the filter bank means, and outputting a decoded image;
An image reproducing apparatus comprising: search means for reading out a corresponding band component from an encoded bit stream recorded on the recording medium in accordance with a search speed.

Images