JP2011061527A - Apparatus and method for conversion of image code, and image decoding apparatus and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To convert a structure of a code stream obtained by coding an image. <P>SOLUTION: An image code conversion apparatus has: an additional data analyzer that analyzes a structure of a first data row consisting of code data obtained by coding a plurality of divisional images obtained by dividing an image independently, and additional data; a code data re-arranging part that re-arranges the code data of the plurality of divisional images coded independently in an alignment of code data when coding an image as a divisional image of a dividing number smaller than that of the first data row, based on a structure of the first data row analyzed by the additional data analyzer; and an additional data updating part that generates additional data to be added to the code data rearranged by the code data re-arranging part, and configures a second data row from the rearranged code data and the generated additional data. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to an image code conversion device, an image code conversion method, an image decoding device, and an image decoding method for converting a coding method of a code stream.

  Conventionally, as an example of an encoding algorithm for compressing still image data, a JPEG (Joint Photographic Experts Group) encoding method is currently widely used mainly on the Internet and digital cameras. On the other hand, the JPEG2000 encoding method was enacted as the next still image encoding algorithm against the background of further performance improvement and addition of functions.

  The JPEG2000 encoding system is composed of part 1 that defines the basic system and part 2 that defines the extended system. In JPEG2000 coding, both methods use arithmetic coding using context modeling in order from bit planes when the image is converted into wavelet transform coefficients by wavelet transform and the absolute values of the transform coefficients are expressed in natural binary numbers. Entropy coding is performed. As a result, JPEG2000 encoding not only has excellent compression performance, but also provides various functions that are not available in conventional JPEG encoding, including an expression method at the time of reproduction, by setting combining encoding parameters.

  For example, in the JPEG2000 encoding algorithm (Non-patent Document 1), the input image is divided into a plurality of rectangular small blocks called tiles, in addition to encoding the input image as a single image, and each tile is independent. It is specified that the encoding process may be performed.

  The advantage of this tile division processing is that the hardware scale of the encoding / decoding processor used for compression / decompression can be reduced, and parallel processing can be facilitated to improve the operation speed. is there. The encoding method using such tile division processing (hereinafter referred to as tile division method) is preferably used when, for example, a video is encoded and streamed in real time in order to broadcast a sports game live. Is mentioned.

  On the other hand, the point that tile division processing is not advantageous is that when the encoding rate is lowered, when applying tile division processing, a slight difference in image quality at the tile boundary is perceived as tile distortion (for example, discontinuity) This is the point that the smoothness is lost. In applications where such tile division processing is inappropriate, the use of tile division is restricted particularly in applications where image quality is important, thereby avoiding the occurrence of tile distortion. For example, in the DCI (Digital Cinema Initiatives) standard, which is an industry standard for digital cinema that employs JPEG2000 encoding, tile division processing is prohibited. Therefore, the digital cinema screening device compliant with the DCI standard only needs to be able to support a coding method based on single tile processing (hereinafter referred to as a single tile method), and therefore can support decoding and reproduction of a code stream encoded using the tile division method. It is not necessarily configured as such. As described above, in the digital cinema, movie content data is encoded and stored offline, and the pre-encoded content data is decoded while being read and displayed in theaters. Sex is not required.

ITU-T recommendation T.800 / ISO / IEC standard 15444-1

  A decoding processing device that only supports decoding of a code stream encoded by a single tile method, such as a conventional digital cinema screening device, has a problem that a code stream encoded by a tile division method cannot be decoded. .

  Therefore, when a digital cinema screening device that supports only a single tile format is used as a decoding processing device and live broadcast video content is to be streamed in real-time in a public viewing format in a theater, the tile division encoding is used. There has been a problem that distribution cannot be performed using a processing device.

  Also, it is costly to make the decoding processing device of the single tile method compatible with decoding of the tile division method or to make the encoding processing device of the tile division method compatible with the encoding of the single tile method. On the other hand, there is a problem that the existing decoding processing device and the encoding processing device cannot be continuously used without being replaced without being replaced.

  The present invention has been made to solve the above-described problems. A code stream (first data) encoded by a tile division type encoding processing apparatus without processing the encoded data itself. An object of the present invention is to obtain an image code conversion device and an image code conversion method for converting an image divided into a code stream (second data string) obtained by encoding an image divided by a smaller number of tile divisions than the first data string. To do.

  The present invention also includes the image code conversion apparatus and the image code conversion method, and includes fewer code streams (first data strings) encoded by the tile division type encoding processing apparatus than the first data strings. An object is to obtain an image decoding apparatus and an image decoding method for converting an image divided by the number of tile divisions into an encoded code stream (second data string) and decoding the converted code stream.

  The image code conversion apparatus according to the present invention includes an additional data analysis unit that analyzes a structure of a first data string composed of code data and additional data obtained by independently encoding a plurality of divided images obtained by dividing an image, Based on the structure of the first data sequence analyzed by the additional data analysis unit, the code data of the plurality of separately encoded images is divided into divided images having a smaller number of divisions than the first data sequence. Code data rearrangement section to be rearranged, and additional data to be added to the code data rearranged by the code data rearrangement section, the rearranged code data and generated And an additional data update unit that configures the second data string from the additional data.

  The image decoding apparatus according to the present invention further includes an image code conversion apparatus that converts the first data string to form a second data string, and the second data update unit of the image code conversion apparatus configures the second data string. An image decoding processing unit that decodes the code data of the data sequence into orthogonal transform coefficients, and an inverse orthogonal transform unit that performs inverse wavelet transform on the orthogonal transform coefficients decoded by the image decoding processing unit.

  According to the image code conversion device and the image code conversion method according to the present invention, the code stream encoded by the tile division type encoding processing device is converted into a code stream that can be decoded by the single tile type decoding processing device. can do.

  Further, according to the image decoding device and the image decoding method according to the present invention, the image code conversion device and the image code conversion method provided in the preceding stage are configured to convert the code stream encoded by the tile division type encoding processing device into a single code stream. The one-tile decoding device converts the code stream into a decodable code stream, and the single-tile decoding device provided in the subsequent stage can decode the converted code stream to obtain a decoded image.

  As described above, according to the image code conversion device and the image code conversion method according to the present invention, a digital cinema screening device that supports only a single tile method is used as a decoding processing device, and a public viewing format is used in a theater. Even when a live video content is screened in real time, it can be distributed by using the existing tile division type encoding processing device as it is by connecting via the image code conversion device according to the present invention. It becomes like this.

  Further, the present invention can be applied to a single tile decoding apparatus that does not support tile division decoding or a tile division encoding apparatus that does not support single tile encoding. By connecting through the image code conversion apparatus according to the above, it becomes possible to continuously utilize the existing decoding processing apparatus and encoding processing apparatus without replacing them.

  In addition, according to the image code conversion device and the image code conversion method according to the present invention, the code stream encoded by the tile division method encoding processing device is decoded by the tile division method configured with a smaller number of tile divisions. The processing apparatus can convert the code stream into a decodable code stream.

  Further, according to the image decoding device and the image decoding method according to the present invention, the image code conversion device and the image code conversion method provided in the preceding stage are configured so that the code stream encoded by the tile division type encoding processing device is more The tile division decoding processing device configured with a small number of tile divisions converts the code stream into a decodable code stream, and the tile division decoding processing device provided in the subsequent stage decodes and decodes the converted code stream. An image can be obtained.

It is a block block diagram which shows an example of the encoding processing apparatus which concerns on Embodiment 1 of this invention. It is explanatory drawing which shows an example of the code stream structure which the encoding processing apparatus which concerns on Embodiment 1 of this invention outputs. It is a block block diagram which shows an example of the decoding processing apparatus which concerns on Embodiment 1 of this invention. It is explanatory drawing which shows an example of the image which the tile division part of the encoding processing apparatus of the tile division system which concerns on Embodiment 1 of this invention divided. It is explanatory drawing which shows an example of the wavelet transformation coefficient which the wavelet transformation part of the encoding processing apparatus of the tile division system which concerns on Embodiment 1 of this invention wavelet-transformed. It is explanatory drawing which shows an example of the extraction order of the code data which the entropy encoding part of the encoding processing apparatus of the tile division | segmentation system which concerns on Embodiment 1 of this invention encoded. It is a block diagram which shows an example of the code stream which the data format process part of the encoding processing apparatus of the tile division system which concerns on Embodiment 1 of this invention outputs. It is explanatory drawing which shows an example of the wavelet transformation coefficient which the wavelet transformation part of the encoding processing apparatus of the single tile system which concerns on Embodiment 1 of this invention wavelet transformed. It is explanatory drawing which shows an example of the extraction order of the code data which the entropy encoding part of the encoding processing apparatus of the single tile system which concerns on Embodiment 1 of this invention encoded. It is a block diagram which shows an example of the code stream which the encoding processing apparatus data format process part of the single tile system which concerns on Embodiment 1 of this invention outputs. It is a block block diagram which shows an example of the image code conversion apparatus which concerns on Embodiment 1 of this invention. It is a flowchart figure which shows an example of the process sequence of the image code conversion apparatus which concerns on Embodiment 1 of this invention. It is a conceptual diagram which shows an example of the rearrangement which the image code converter which concerns on Embodiment 1 of this invention converts. It is a block diagram which shows an example of the code stream after the conversion which the image code converter which concerns on Embodiment 1 of this invention outputs. It is explanatory drawing which shows an example of the pixel reference of the wavelet transformation / inverse transformation of the tile boundary before and behind the code conversion of the image code conversion apparatus which concerns on Embodiment 1 of this invention. It is explanatory drawing which shows an example of a response | compatibility with the image which the tile division part of the encoding processing apparatus of the tile division system which concerns on Embodiment 2 of this invention tile-divides, and a precinct. It is a conceptual diagram which shows an example of the tile division | segmentation with which the input code stream of the tile division system which the image code conversion apparatus which concerns on Embodiment 3 of this invention handles corresponds. It is a conceptual diagram which shows the input code stream structure of FIG. 17 which the image code converter which concerns on Embodiment 3 of this invention handles. It is a conceptual diagram which shows an example of the tile division | segmentation with which the output code stream of the tile division system which the image code conversion apparatus which concerns on Embodiment 3 of this invention handles corresponds. It is a conceptual diagram which shows the output code stream structure of FIG. 19 which the image code converter which concerns on Embodiment 3 of this invention handles.

  The present invention is arranged between an encoding processing device that encodes an image by a tile division method suitable for real-time streaming such as live broadcast and a decoding processing device that supports only a single tile method such as a digital cinema screening device. Further, the present invention relates to an image code conversion device and an image code conversion method for converting a code stream based on a tile division method encoded by the encoding processing device into a code stream having the same structure as a single tile method that can be decoded by the decoding processing device. It is. Hereinafter, an image code conversion device and an image code conversion method according to the present invention will be described according to each embodiment.

Embodiment 1 FIG.
Hereinafter, the code conversion device and the image code conversion method according to the first embodiment of the present invention will be described.

  In the following description, for example, in order to use the JPEG2000 encoding / decoding processing procedure, an outline of the JPEG2000 processing procedure will be described first.

  FIG. 1 is a block diagram illustrating an example of a JPEG 2000 encoding processing apparatus 10. A basic encoding process procedure will be described with reference to FIG.

  The tile dividing unit 101 divides input image data into a plurality of rectangular small blocks (hereinafter referred to as tiles). In subsequent processing units, processing is performed independently for each tile. However, it is also possible to process the entire image as a single tile without dividing the tile.

  The wavelet transform unit 102 performs two-dimensional discrete wavelet transform (DWT: Discrete Wavelet Transform) on each tile divided by the tile partitioning unit 101 to generate an LL component (horizontal low frequency component, vertical low frequency component), and HL component (horizontal This is divided into four subbands: a high-frequency component in the direction, a low-frequency component in the vertical direction), a LH component (low-frequency component in the horizontal direction, a high-frequency component in the vertical direction), and a HH component (a high-frequency component in the horizontal direction and a high-frequency component in the vertical direction). When the decomposition level is 2 or more, only the LL component of each level is subjected to two-dimensional discrete wavelet transform, and the process of decomposing into four subbands is repeated.

  The quantization unit 103 quantizes the wavelet transform coefficient decomposed into subbands for each subband. Here, when a reversible wavelet transform is applied in the wavelet transform unit 102 and the quantization step is set to 1 in the quantization unit 103, it becomes possible to perform lossless compression (lossless compression). It can be regarded as not applicable.

  The entropy encoding unit 104 performs entropy encoding processing on the quantized wavelet transform coefficients. At this time, the entropy encoding unit 104 further divides the quantized wavelet transform coefficient into units of precincts and code blocks, and performs entropy encoding processing for each code block. Note that the precinct and the code block are small rectangular blocks each having a size that is a power of 2 both vertically and horizontally, and each precinct includes one or more code blocks.

  The data format processing unit 105 packetizes entropy-encoded code data in units of precinct (wavelet transform coefficient division), resolution, component, and layer (hierarchy in the bit depth direction), and parameter information relating to encoding called a marker Is added to form a code stream. At this time, the arrangement order of the packets in the code stream can be determined for each tile from the layer, resolution, precinct, and component elements, and a combination of the priority order of these elements that defines the expression method during playback. Is called a progression order. Hereinafter, for example, a progression order in which the resolution R, the layer L, the component C, and the precinct P are packetized in the priority order is represented as RLCP.

  FIG. 2 is an explanatory diagram showing an example of a JPEG2000 code stream structure. A main header including information related to the entire encoded image is added to the head of the entire code stream. The packetized tile code data (tile stream data) is collected into tile parts, and a tile part header is added to the head of each tile part. An end marker indicating the end is added to the end of the entire code stream.

  FIG. 3 is a block diagram illustrating an example of a JPEG 2000 decoding processing apparatus 20. A basic decoding process procedure will be described with reference to FIG. In the decoding process, a reproduction image is obtained from the code stream by restoring each process of the blocks shown in the coding block diagram of FIG. 1 in reverse order.

  The data format processing unit 201 analyzes the header from the code stream and extracts code data. The entropy decoding unit 202 entropy decodes the extracted code data into wavelet transform coefficients. The inverse quantization unit 203 inversely quantizes the entropy-decoded wavelet transform coefficients. The wavelet inverse transform unit 204 performs two-dimensional discrete wavelet inverse transform (IDWT: Inverse Discrete Wavelet Transform) on the inversely quantized wavelet transform coefficients. The tile combining unit 205 combines the images of the tiles that have been subjected to wavelet inverse transformation into one image.

  Here, if the encoding processing device applies the lossless wavelet transform and generates the quantized lossless compression code stream of the quantization step 1, the decoding processing device performs quantization in the inverse quantization unit 203. By setting the step to 1 and applying the reversible wavelet inverse transform in the wavelet inverse transform unit 204, lossless decoding can be performed. Note that, in the quantization step 1, it can be considered that the inverse quantization process is not substantially applied.

  The JPEG 2000 encoding processing device 10 in FIG. 1 and the JPEG 2000 decoding processing device 20 in FIG. 3 have been illustrated and described as a tile division encoding processing device and decoding processing device, but only a single tile method is handled. 1, an apparatus configuration (not shown) in which the tile dividing unit 101 of the encoding processing apparatus 10 of FIG. 1 and the tile combining unit 205 of the decoding processing apparatus 20 of FIG. 2 are not mounted. It can also be taken.

  Generation of a code stream to be converted that is input to the image code conversion apparatus and the image code conversion method according to Embodiment 1 of the present invention will be described. This code stream to be converted is encoded and generated by setting the following encoding parameter for an encoding processing apparatus that encodes an image by the tile division method.

[Example of coding parameter setting for tile division method]
・ Original image size: 2 ^{(n + 2)} × 2 ^{(m + 2)}
-Image area offset: (0, 0)
・ Tile offset: (0,0)
・ Tile size: 2 ^{(n + 1)} x2 ^{(m + 1)}
・ Wavelet transform decomposition level: 1
・ Quantization step: 1
-Precinct size: No setting (maximum size)
・ Progression order: RLCP

First, FIG. 4 is an explanatory diagram showing an example of an image that is tile-divided by the tile dividing unit of the tile-division coding processing apparatus according to Embodiment 1 of the present invention. Here, under the encoding parameter setting described above, the tile dividing unit 101 offsets the input original image of size 2 ^{(n + 2)} × 2 ^{(m + 2)} in the vertical and horizontal directions, and the tile size 2 ^{(N + 1)} × 2 ^{(m + 1)} tile images T ₀ to T ₃ are divided into four. Here, the offset represents a position at which actual image data arrangement is started or a position at which tile division is started with respect to the origin (0, 0) of the coordinate system indicating each pixel position of the image data. . In this example, it is assumed that both the image area and the tile division start from the origin (0, 0).

Next, FIG. 5 is an explanatory diagram showing an example of a wavelet transform coefficient that has been wavelet transformed by the wavelet transform unit of the tile division type encoding processing apparatus according to Embodiment 1 of the present invention. The wavelet transform unit 102 performs decomposition level 1 wavelet transform processing on the tile-divided image, and divides the image into four subbands of LL component, HL component, LH component, and HH component. At this time, the size of each subband is 2 ⁿ × 2 ^m . Here, the LL component, the HL component, the LH component, and the HH component that are the four subbands of the tile image T _k are expressed as LL _k , HL _k , LH _k , and HH _k (k = 0 to 3).

  Since the quantization unit 103 has a quantization step of 1, the quantization process is substantially omitted here. Note that the value may not be 1 depending on the setting of the quantization step size.

  FIG. 6 is an explanatory diagram showing an example of the extraction order of the code data encoded by the entropy encoding unit of the tile division type encoding processing apparatus according to Embodiment 1 of the present invention. Here, in each tile image, the extraction order is shown in units of the LL component of the subband and the other three subbands HL component, LH component, and HH component. Since the entropy encoding unit 104 sets the precinct size to the maximum size (the same size as the subband), the entropy encoding unit 104 encodes the wavelet transform coefficient of each subband in a numbered unit.

  FIG. 7 is a block diagram showing an example of a code stream output by the data format processing unit of the tile division type encoding processing apparatus according to Embodiment 1 of the present invention. The data format processing unit 105 takes out the encoded code data in the numbered order of FIG. 6 and arranges them as packets according to the progression order RLCP for each precinct, and adds a main header, tile part header, and end marker. Outputs a structured codestream. In FIG. 7, the code data of the HL component, the LH component, and the HH component with the same numbers as those in FIG. 6 are shown together. Here, the HL component, the LH component, and the HH component are shown in this order. However, the order is not limited to this order as long as it is a predetermined order.

  The converted code stream output by the image code conversion apparatus and the image code conversion method according to Embodiment 1 of the present invention will be described. Here, similarly to the encoding parameter setting of the encoding processing device of the tile division method described above, a single tile decoding device that decodes the converted code stream generates a code stream that can be decoded. This will be described from the viewpoint of setting the encoding parameter for the one-tile encoding processing apparatus.

[Single tile encoding parameter setting example]
・ Original image size: 2 ^{(n + 2)} × 2 ^{(m + 2)}
-Image area offset: (0, 0)
・ Tile offset: (0,0)
・ Wavelet transform decomposition level: 1
・ Quantization step: 1
Precinct size: 2 ⁿ × 2 ^m (LL component), 2 ^{(n + 1)} × 2 ^{(m + 1)} (other than LL component)
・ Progression order: RLCP

First, FIG. 8 is an explanatory diagram showing an example of a wavelet transform coefficient that has been wavelet transformed by the wavelet transform unit of the single tile coding processing apparatus according to the first embodiment of the present invention. The wavelet transform unit 102 applies the decomposition level 1 wavelet transform process without dividing the original image having a vertical and horizontal size of 2 ^{(n + 2)} × 2 ^{(m + 2)} into tiles, so that the LL component, the HL component, the LH component, and the HH component. Into four subbands. By this wavelet transform processing, the size of each subband is 2 ^{(n + 1)} × 2 ^{(m + 1)} .

  Since the quantization unit 103 has a quantization step of 1, the quantization process is substantially omitted here. Note that the value may not be 1 depending on the setting of the quantization step size.

Next, FIG. 9 is an explanatory diagram showing an example of the extraction order of code data encoded by the entropy encoding unit of the single tile encoding processing apparatus according to Embodiment 1 of the present invention. Here, in each subband, the extraction order in units of precincts is shown. The entropy encoding unit 104 is a unit in which the precinct size is set to 2 ⁿ × 2 ^{m for} the LL component and 2 ^{(n + 1)} × 2 ^{(m + 1)} other than the LL component, so that the precinct wavelet transform coefficients of each subband are numbered. Encode with

  FIG. 10 is a block diagram showing an example of a code stream output by the data format processing unit of the single tile encoding processing apparatus according to Embodiment 1 of the present invention. The data format processing unit 105 takes out the encoded code data in the numbered order of FIG. 9 and arranges them as packets according to the progression order RLCP for each precinct, and adds a main header, tile part header, and end marker. Outputs a structured codestream. In FIG. 10, the code data of the HL component, the LH component, and the HH component with the same numbers as those in FIG. 9 are shown together. Here, the HL component, the LH component, and the HH component are shown in this order. However, the order is not limited to this order as long as it is a predetermined order.

  The code stream having the configuration shown in FIG. 10 encoded by setting the encoding parameters shown here can be decoded by a single tile decoding processing apparatus such as a digital cinema screening apparatus. .

  As described above, the code conversion apparatus according to Embodiment 1 of the present invention converts the code stream shown in FIG. 7 of the encoding processing apparatus that performs encoding by setting the encoding parameter of the tile division method to the single tile method. The decoding is performed by setting a coding parameter and converting the code stream shown in FIG. Hereinafter, the conversion process into the code stream of the code conversion apparatus will be described in detail.

  The 2 × 2 tile code stream of the progression order RLCP in the tile division method includes an LL component, LL, following each tile part header, as shown in the code stream structure of FIG. 7 and the arrangement of precinct divided code data of FIG. Code data packets of components other than are configured to continue for each tile in the order of upper left, upper right, lower left, and lower right.

  On the other hand, the single tile code stream of the progression order RLCP in the single tile system follows the single tile part header as shown in the code stream structure of FIG. 10 and the arrangement of the precinct divided code data of FIG. Signs in the order of LL components in the upper left, upper right, lower left, and lower right regions, components other than LL in the upper left region, components other than LL in the upper right region, components other than LL in the lower left region, and components other than LL in the lower right region Consists of data packets.

  FIG. 11 is a block configuration diagram showing an example of an image code conversion apparatus according to Embodiment 1 of the present invention. FIG. 12 is a flowchart showing an example of the processing procedure of the image code conversion apparatus according to Embodiment 1 of the present invention. The processing unit in the block configuration diagram and the processing operation in the flowchart will be described in association with each other.

  First, the additional data analysis unit 301 of the image code conversion apparatus 300 analyzes the header information included in the additional data of the input tile division method code stream (step S401). Next, based on the analysis result of the header information, the code data rearrangement unit 302 rearranges the code data decomposed by applying to the code stream structure of FIG. 10 into an order corresponding to the single tile method (step S402). The additional data update unit 303 generates the additional data corresponding to the single tile method and adds it to the code data rearranged in the order corresponding to the single tile method, thereby updating the code stream to the single tile method. (Step S403).

  That is, the image code conversion apparatus 300 according to Embodiment 1 of the present invention is configured by the additional data analysis unit 301 including code data obtained by independently encoding a plurality of divided images obtained by dividing an image and additional data including a header. Of the tile division method code stream (first data string), and the code data rearrangement unit 302 analyzes the tile division method code stream (first data string) analyzed by the additional data analysis unit 301. Based on the structure, the code data of a plurality of divided images independently encoded are rearranged along with the code data when the image is encoded as a single image, and the additional data update unit 303 re-arranges the code data Generated additional data to be added to the code data rearranged by the unit, and encoded the image as a single image from the rearranged code data and the generated additional data It constitutes a code stream of a single tiling (second data string) with the encoded data structure.

  The processing procedure of each processing unit in the block configuration diagram of FIG. 11 will be described in more detail. The additional data analysis unit 301 analyzes the input 2 × 2 tile code stream of the progression order RLCP shown in FIG. 7 based on the main header and tile part header included in the additional data, and performs the precinct division shown in FIG. So that it can be extracted in units of the encoded data.

  FIG. 13 is a conceptual diagram showing an example of rearrangement converted by the image code conversion apparatus according to Embodiment 1 of the present invention. The code data rearrangement unit 302 corresponds to the single tile method after code conversion shown in the right row in the code data arrangement of FIG. 6 shown corresponding to the tile division method before code conversion shown in the left row. Rearrangement to rearrange the code data arrangement of FIG. 9 shown in FIG. Note that the numbers assigned to the right code data arrangement are different from the numbers given in FIG. 9 because they are indicated by the numbers assigned to the corresponding left code data arrangement.

  Based on the concept of such rearrangement, the code data rearrangement unit 302 converts the code data of the tile division scheme shown in FIG. 7 into the code stream configuration of the single tile scheme of FIG. Perform a conversion that sorts. Here, only the order of the code data of the coding unit of the image is rearranged, and the coding unit of the image is not re-encoded.

  The additional data update unit 303 generates a code stream that can be decoded by a single tile decoding device by re-adding the tile division header included in the additional data of the code stream to the single tile header.

  FIG. 14 is a block diagram showing an example of a converted code stream output by the image code conversion apparatus according to Embodiment 1 of the present invention. The structure of the code stream corresponding to the numbering of the code data arrangement after rearrangement on the right side of FIG. 13 is shown.

  The code stream converted in this way has a completely identical structure when compared with a code stream obtained by encoding an image using a single tile method, but the code data is only used for wavelet transform processing at tile boundaries. There are differences.

  FIG. 15 is an explanatory diagram illustrating an example of pixel reference for tile boundary wavelet transform / inverse transform before and after code conversion of the image code conversion apparatus according to the first embodiment of the present invention. In the wavelet transform from the original image to the wavelet transform coefficient in the tile splitting type coding processing apparatus at the time of the coding processing shown in the upper stage before the code conversion, the tile is used to guarantee that the tile can be encoded and decoded independently. Pixel reference of adjacent tiles beyond the boundary is not performed. On the other hand, since the original tile is regarded as a precinct in the wavelet inverse transformation from the wavelet transform coefficient to the reproduced image in the decoding processing apparatus of the single tile method at the time of decoding processing shown in the lower stage after the code conversion, the tile boundary In other words, the pixel reference of the adjacent tile beyond the precinct boundary is performed.

  For this reason, the code stream converted by the image code conversion device and the image code conversion method according to Embodiment 1 of the present invention may use a lossless compression (lossless compression) mode for encoding before conversion. Although it cannot be completely decoded into the original image, the region that is not decoded into the original image is limited to a portion where pixel reference is made across the tile boundary by inverse wavelet transformation. If there is no sharp image quality change at the tile boundary, the influence on the image quality at the tile boundary is extremely small.

  In the image code conversion device and the image code conversion method according to Embodiment 1 of the present invention, a tile-divided code stream (first data string) divided into 4 × 2 × 2 tiles is received. Although an example of conversion to a code stream (second data string) of a single tile method has been described, the conversion is not limited to conversion from four tile division to a single tile without tile division, and the number of divisions is not four tiles. However, the code stream can be converted in accordance with the reference order of the predetermined subband, tile, precinct, and code block.

  Further, the image code conversion apparatus 300 of FIG. 11 according to Embodiment 1 of the present invention is provided in the front stage and a decoding processing unit corresponding to the single tile system is provided in the rear stage, and the received tile division system code stream (first (Data sequence) may be converted into a single tile code stream (second data sequence), and the converted single tile code stream may be decoded into an image. Here, the decoding processing unit arranged in the subsequent stage may be, for example, the single tile decoding processing device 200 of FIG. 3 described above. At this time, since tile combining is not performed in the single tile method, a configuration without the tile combining unit 205 may be used. However, if the configuration has the tile combining unit 205, the output of the wavelet inverse transform unit 204 may be directly passed. .

  Further, the code data rearrangement unit 302 and the additional data update unit 303 of the decoding processing apparatus 200 of the single tile method of FIG. 3 and the image code of FIG. The data format processing unit 201 of the conversion apparatus 300 cooperates to control the entropy decoding unit 202 to appropriately extract code data and additional data from the tile division code stream in the single tile processing order. An image decoding device may be realized as one processing unit.

  As described above, according to the image code conversion device and the image code conversion method according to Embodiment 1 of the present invention, the encoding that can be handled by the tile division coding processing device and the single tile decoding processing device. A code stream that can be decoded by a single tile decoding processor by appropriately setting parameters and rearranging the encoded data without processing the code data itself, by encoding the code stream encoded by the tile division encoding processor. Can be converted to

  In other words, when live-streamed video content is screened in the theater in real time, a single-tile digital cinema screening device can be used even if a tile-division encoding processing device that is advantageous in terms of processing speed and hardware scale is used. It can be used as it is as a decoding processing device. In this way, if the encoding processing device and the decoding processing device are connected via the image code conversion device of the present invention, for example, a tile division decoding function is added to the single tile decoding processing device, or tile division is performed. A single tile encoding function added to the encoding processing apparatus, and at least one of the encoding processing apparatus and the decoding processing apparatus is compatible with both the single tile method and the tile division method Alternatively, since it is not necessary to replace the decoding processing device, the tile division coding processing device and the single tile decoding processing device can be continuously used at low cost as they are.

  Further, according to the image code conversion device and the image code conversion method according to Embodiment 1 of the present invention, when there is a steep image quality change at the tile boundary in the tile division method, the tile boundary becomes conspicuous from the image quality difference. Even for an image, by converting it to a single tile method and decoding it, there is an effect of relieving the image quality difference and making the tile boundary inconspicuous by inverse wavelet conversion with pixel reference beyond the original tile boundary.

Embodiment 2. FIG.
In the description of the image code conversion device and the image code conversion method according to Embodiment 1 of the present invention described above, the number of vertical and horizontal pixels is 2 for each of the vertical and horizontal sizes of the rectangular tile image divided by the tile dividing unit. It is assumed that it is a power of. In the image code conversion device and the image code conversion method according to the second embodiment of the present invention, the boundary between adjacent tile images, that is, the combination boundary when combining tile images as a single image, An example corresponding to the case where the number of vertical and horizontal pixels is not a power of 2 by matching with the boundary will be described.

  In order to generate a code stream to be converted that is input to an image code conversion apparatus and an image code conversion method according to Embodiment 2 of the present invention, an encoding processing apparatus that encodes an image using a tile division method The next encoding parameter is set.

[Example of coding parameter setting for tile division method]
-Original image size: 384 x 384
-Image area offset: (0, 0)
・ Tile offset: (0,0)
・ Tile size: 192 x 192
・ Wavelet transform decomposition level: 1
・ Quantization step: 1
・ Code block size: 32 × 32
・ Progression order: RLCP

  The generation of the code stream of the tile division method is the same as when the vertical and horizontal sizes of the image are powers of 2, as described in FIGS. 4 to 7 in the first embodiment. The size of a subband obtained by dividing an image with a tile size of 192 × 192 and performing one-level decomposition by wavelet transform is 96 × 96.

  When the vertical and horizontal size of the image is a power of 2 as in the first embodiment described above, the code stream is obtained by matching this subband region with the region of precinct division when encoding with a single tile. However, in JPEG2000 encoding, there is a prescription that the vertical and horizontal size of the precinct is a power of 2, and therefore a size of 96 × 96 cannot be designated as the precinct size.

  Therefore, by designating a precinct size larger than 96 × 96 and offsetting the coordinate origin of precinct division, the subband region is divided into 96 × 96 regions and packetized.

  FIG. 16 is an explanatory diagram showing an example of a correspondence between an image and a precinct that has been tile-divided by the tile division unit of the tile-division coding apparatus according to Embodiment 2 of the present invention. The 192 × 192 LL component subband region is divided into 128 × 128 precincts indicated by broken lines, and the tile offset is set to (32, 32). be able to.

  A single tile decoding apparatus that decodes the converted code stream output from the image code conversion apparatus and the image code conversion method according to Embodiment 2 of the present invention generates a single code stream that can be decoded. The encoding parameter setting for the tile type encoding processing apparatus is as follows.

[Single tile encoding parameter setting example]
-Original image size: 384 x 384
Image area offset: (64, 64)
・ Tile offset: (64, 64)
(Thus, the offset of tiles other than the LL component is (64, 64), and the offset of tiles of the LL component is (32, 32)).
・ Wavelet transform decomposition level: 1
・ Quantization step: 1
Precinct size: 128 × 128 (LL component), 256 × 256 (other than LL component)
・ Code block size: 32 × 32
・ Progression order: RLCP

  The code stream encoded by setting the encoding parameters shown here can be decoded by a single tile decoding processing apparatus such as a digital cinema screening apparatus.

  Even if the image size is not a power of 2 by setting the encoding parameter to match the combination boundary when combining tile images as a single image with the boundary of the processing block at the time of encoding. Since the divided areas of the encoded data of the tile division code stream and the single tile code stream can be made the same, as in the case where the image size in the first embodiment described above is a power of 2, Code stream conversion is possible.

  In the image code conversion device and the image code conversion method according to the second embodiment of the present invention, the case where the number of vertical and horizontal pixels of the tile image is not a power of 2 has been described. Even when the number of pixels is not a power, it can be explained by setting an offset in the direction.

  As described above, according to the image code conversion device and the image code conversion method according to Embodiment 2 of the present invention, the effects in the image code conversion device and the image code conversion method according to Embodiment 1 of the present invention are the same. As a result, a code stream based on the tile division method can be converted into a code stream having the same structure as that of the single tile method, and can be decoded by a decoding processing apparatus using the single tile method.

Embodiment 3 FIG.
The image code conversion apparatus and the image code conversion method according to Embodiment 1 of the present invention described above have been described for converting from a tile-division code stream to a single tile code stream. In the third embodiment of the present invention, input of a code stream of a tile division method composed of a plurality of tiles is not changed, but adjacent tiles are collected by the method described in the first embodiment of the present invention. Thus, for example, an example of converting a 4 × 4 tile tile division code stream into a 2 × 2 tile division code stream will be described.

  First, an image code conversion device according to Embodiment 3 of the present invention is the image code conversion device described in FIG. 11 of Embodiment 1 of the present invention, in which the additional data analysis unit 301 divides the image into a plurality of divisions. The code data rearrangement unit 302 analyzes the structure of a code stream (first data string) of a tile division scheme composed of code data obtained by independently encoding an image and additional data including a header. Based on the structure of the code stream (first data string) of the tile division method analyzed by 301, the code data of a plurality of divided images independently encoded is obtained by dividing the image into a smaller number of tile divisions than the first data string. Are rearranged in the order of the code data when encoded as a divided image divided in step, and the additional data update unit 303 adds the code data rearranged by the code data rearrangement unit. Tile having a code data structure when the additional data is generated and the image is encoded as a divided image divided by a smaller number of tile divisions than the first data string from the rearranged code data and the generated additional data. A code stream (second data string) of a division method is configured.

  FIG. 17 is a conceptual diagram showing an example of tile division corresponding to the input code stream of the tile division method handled by the image code conversion apparatus according to Embodiment 3 of the present invention. In this example, the image is divided into 4 × 4 tiles (16 tiles).

  FIG. 18 is a conceptual diagram showing the input code stream structure of FIG. 17 handled by the image code conversion apparatus according to Embodiment 3 of the present invention.

FIG. 19 is a conceptual diagram showing an example of tile division corresponding to the output code stream of the tile division method handled by the image code conversion apparatus according to Embodiment 3 of the present invention. In this example, the 4 × 4 tile in FIG. 17 is reconfigured to a 2 × 2 tile. That is, the tile T ′ ₀ in FIG. 19 is reconstructed from the _four tiles T ₀ , T ₁ , T ₄ , and T _{5 in FIG} . Similarly, tile T ′ ₁ is from tiles T ₂ , T ₃ , T ₆ , T ₇ , tile T ′ ₂ is from tiles T ₈ , T ₉ , T ₁₂ , T ₁₃ , and tile T ′ ₃ is tiles Reconstructed from T ₁₀ , T ₁₁ , T ₁₄ , and T ₁₅ .

20 is a conceptual diagram showing the output code stream structure of FIG. 19 handled by the image code conversion apparatus according to Embodiment 3 of the present invention. In the figure, the first tile part header and tile stream data corresponding to the tile T ′ ₀ are based on the conversion described with reference to FIG. 13 from the _four tiles T ₀ , T ₁ , T ₄ , T ₅ in FIG. The tile part header and the tile stream data in FIG. The other three tiles T ′ ₁ , T ′ ₂ , T ′ ₃ are similarly reconstructed.

  In addition, since the original tile boundary shown by the broken line in FIG. 19 regards the original tile as a precinct as described above with reference to FIG. 15, the precinct boundary is not treated as a tile boundary and is adjacent to the precinct boundary. The pixel reference of the tile will be performed. On the other hand, the pixel reference of the adjacent tile beyond the tile boundary is not performed in the portion treated as the tile boundary as usual.

  As described above, according to the image code conversion device and the image code conversion method according to Embodiment 3 of the present invention, the transmission side tile division method encoding processing device and the reception side tile division method decoding processing device are provided. Reception that can process only the number of tiles smaller than the number of tiles processed by the encoding processing device of the tile division method on the transmission side by appropriately setting compatible encoding parameters and rearranging without processing the encoded data itself It can be converted into a decodable codestream even by the side tile division decoding device.

  Further, according to the image code conversion device and the image code conversion method according to the third embodiment of the present invention, the image code conversion device according to the first embodiment of the present invention is used for the image quality of the boundary portion where the original tiles are joined. The same effect as described in the image code conversion method can be obtained.

  As described above, in the image code conversion device and the image code conversion method according to the embodiments of the present invention, the image is described as being configured by M × N tiles (M, N> 1). The same can be said for the case of M × 1 tile and 1 × N tile.

  Further, in the image code conversion apparatus and the image code conversion method according to the embodiment of the present invention, for example, a vertical 4 × horizontal format can be used without converting a tile division code stream into a single tile code stream at a time. As a process of converting the code stream of the tile division method divided into 16 tiles into the code stream of the single tile method, it is regarded as a process of integrating the 4 tiles of 2 × 2 horizontally into 1 tile, Is converted into a tile-divided code stream that is divided into 4 vertical 2 × 2 horizontal tiles, and the tile division number is gradually changed to a single-tile code stream that is divided into vertical and horizontal halves. You may apply as conversion which decreases. Further, the conversion of integrating two tiles into one tile for each direction only in the vertical direction or only in the horizontal direction may be repeated stepwise. That is, the image code conversion apparatus and the image code conversion method according to the embodiment of the present invention are not only for converting a tile-division code stream at a time into a single-tile code stream, but also in the above-mentioned intermediate stage. It is apparent that the generated tile division code stream can be converted into an integrated tile division code stream. Depending on the number of tile divisions, the vertical and horizontal division numbers may be reduced to 1 / N instead of ½, and the vertical and horizontal division numbers may be reduced at different ratios.

  In the image code conversion apparatus and the image code conversion method according to the embodiment of the present invention, the encoded data of the tile image encoded by dividing the tile is described as being stored in a single code stream. The encoded data of the image may be encoded data of a plurality of images included in a plurality of code streams completely independent for each tile.

  Further, in the image code conversion apparatus and the image code conversion method according to the embodiment of the present invention, the case where JPEG2000 encoding is applied has been described as an example, but a JPEG2000 encoding method using similar image division and encoding units is described. However, other encoding methods may be applied. Moreover, although the case where the wavelet transformation is applied to the image has been described, the present invention is not limited to the wavelet transformation, and other orthogonal transformations may be applied.

DESCRIPTION OF SYMBOLS 10 Encoding processing apparatus 20 Decoding processing apparatus 30 Image code conversion apparatus 101 Tile division part 102 Wavelet transformation part 103 Quantization part 104 Entropy encoding part 105 Data format processing part 201 Data format processing part 202 Entropy decoding part 201 Inverse quantization part 204 Wavelet inverse transform unit 205 Tile combination unit 301 Additional data analysis unit 302 Encoded data rearrangement unit 303 Additional data update unit

Claims (11)

An additional data analysis unit that analyzes the structure of a first data string composed of code data and additional data obtained by independently encoding a plurality of divided images obtained by dividing an image;
Based on the structure of the first data sequence analyzed by the additional data analysis unit, the code data of the plurality of independently encoded divided images is converted into the image having a smaller number of divisions than the first data sequence. A code data rearrangement unit that rearranges the code data when encoded as a divided image;
The code data rearrangement unit generates additional data to be added to the rearranged code data, and the image includes the rearranged code data and the additional data update unit constituting the second data string from the generated additional data Code conversion device.   2. The image code according to claim 1, wherein the sequence of code data rearranged by the code data rearrangement unit is a sequence of code data when the image is encoded as a divided image having a division number of 1. Conversion device.   The image code conversion apparatus according to claim 1, wherein the plurality of divided images are tile images obtained by dividing the image into rectangular tiles.   4. The image code conversion apparatus according to claim 3, wherein the tile image is an image having a pixel number that is a power of 2 in both vertical and horizontal sizes. The tile image is encoded into code data in units of encoding processing blocks,
4. The image code conversion apparatus according to claim 3, wherein the encoding processing block is arranged with reference to a boundary between the tile image and an adjacent tile image.   3. The image code conversion according to claim 1, wherein the first data string is encoded with a wavelet transform coefficient obtained by wavelet transforming the plurality of divided images independently for each of the divided images. 4. apparatus. The image code conversion device according to claim 6, comprising the second data string by converting the first data string,
An image decoding processing unit for decoding the code data of the second data string formed by the additional data updating unit into a wavelet transform coefficient;
An image decoding apparatus comprising: an inverse wavelet transform unit that performs inverse wavelet transform on the wavelet transform coefficients decoded by the image decoding processing unit. An additional data analysis step of analyzing a structure of a first data string composed of code data and additional data obtained by independently encoding a plurality of divided images obtained by dividing an image;
Based on the structure of the first data sequence analyzed by the additional data analysis step, the code data of the plurality of independently encoded divided images is converted into the image having a smaller number of divisions than the first data sequence. A code data rearrangement step of rearranging the code data when encoded as a divided image; and
The image data including the additional data to be added to the code data rearranged in the code data rearrangement step, and the additional data update step of forming the second data string from the rearranged code data and the generated additional data Conversion method.   9. The image according to claim 8, wherein the sequence of code data rearranged in the code data rearrangement step is a sequence of code data when the image is encoded as a divided image having a division number of 1. Code conversion method. An additional data analysis step of analyzing the structure of a first data string composed of code data and additional data obtained by encoding wavelet transform coefficients obtained by independently wavelet transforming each of the divided images obtained by dividing the image;
Based on the structure of the first data sequence analyzed by the additional data analysis step, the code data of the plurality of independently encoded divided images is converted into the image having a smaller number of divisions than the first data sequence. A code data rearrangement step of rearranging the code data when encoded as a divided image; and
An additional data update step of generating additional data to be added to the code data rearranged by the code data rearrangement step, and constituting a second data string from the rearranged code data and the generated additional data;
An image decoding process step of decoding the code data of the second data string formed by the additional data update step into a wavelet transform coefficient;
An image decoding method including an inverse wavelet transform step of performing inverse wavelet transform on the wavelet transform coefficient decoded by the image decoding processing step.   11. The image according to claim 10, wherein the sequence of code data rearranged in the code data rearrangement step is a sequence of code data when the image is encoded as a divided image having a division number of one. Decryption method.

Images