JP2008283701A - Data processor and data processing method, and recording medium - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To embed second data without increasing an amount of encoded data generated by encoding first data and to accurately decode the first and second data. <P>SOLUTION: An encoder 11 generates the encoded data in which the second data has been embedded by shifting the correspondence relationship between quantized code obtained by quantizing the first data and a quantization error by a manipulated variable corresponding to the second data. A decoder 12 decodes temporary encoded data by shifting the correspondence relationship between a quantized code of the buried encoded data and a quantization error by a specified manipulated variable, and decodes temporary decoded data from the temporary encoded data. It is determined whether the temporary encoded data are correct, and the first and second data are decoded based upon the judgment result. The present invention is applicable to a case where the second data are embedded without increasing the amount of encoded data generated by encoding the first data, etc. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a data processing device, a data processing method, and a recording medium, and in particular, for example, so that information can be embedded in an image without degrading the image quality of a decoded image and without increasing the amount of data. The present invention relates to a data processing apparatus, a data processing method, and a recording medium.

  As a technique for embedding information without increasing the amount of data, there is, for example, a method of converting the least significant bit or the lower two bits of digital audio data into information to be embedded. This method uses the fact that the lower bits of the digital audio data do not significantly affect the sound quality, and simply replaces the lower bits with information to be embedded. Therefore, at the time of playback, the information is embedded. The digital audio data is output as it is without returning its lower bits. In other words, it is difficult to restore the low-order bits embedded with information, and the low-order bits do not affect the sound quality so much, so the digital audio data is output with the information embedded. Is done.

  However, with the above-described method, data different from the original data is output. Therefore, when the data is audio data, the sound quality is affected, and when the data is video data, the image quality is affected.

  The present invention has been made in view of such circumstances. For example, information can be embedded in an image without degrading the image quality and without increasing the data amount. It is.

  The data processing apparatus or the recording medium according to the first aspect of the present invention includes an encoding unit that quantizes the first data and outputs a quantization code and a quantization error obtained as a result as encoded data. The embedded encoding in which the second data is embedded by shifting the correspondence between each value of the quantization code as the encoded data and the quantization error by an operation amount corresponding to the second data A data processing apparatus including an embedding unit for generating data, or a recording medium on which a program for causing a computer to function as a data processing apparatus is recorded.

  In the first data processing method of the present invention, the data processing apparatus quantizes the first data, and outputs a quantization code and a quantization error obtained as a result as encoded data. A step of generating embedded encoded data in which the second data is embedded by shifting the correspondence between each value of the quantization code and the quantization error by an operation amount corresponding to the second data. It is a data processing method including.

  In the first aspect of the present invention, the first data is quantized, and the resulting quantization code and quantization error are output as encoded data. Furthermore, the embedded code in which the second data is embedded by shifting the correspondence between each value of the quantization code as the encoded data and the quantization error by an operation amount corresponding to the second data Data is generated.

  In the data processing apparatus or the recording medium according to the second aspect of the present invention, the second data is embedded in the encoded data including the quantization code and the quantization error obtained by quantizing the first data. Temporary encoded data decoding means for decoding the temporary encoded data by shifting the correspondence between each value of the quantization code and the quantization error in the embedded encoded data by a specific operation amount; Temporary decoded data decoding means for decoding temporary decoded data, the quantization code and the quantization error, and the temporary decoded data decoding means decoded from the temporary encoded data according to a predetermined code book A determination unit that determines whether or not the temporary encoded data is correct by comparing a quantization code and a quantization error obtained by re-encoding the temporary decoded data; The temporary encoded data decoding means changes the manipulated variable according to the determination result by the determination means and decodes the temporary encoded data again, and the determination means determines that the temporary encoded data is correct. A data processing device that outputs data corresponding to the manipulated variable as the second data, and outputs temporary decoded data from the temporary decoded data decoding means as true decoded data, or a data processing device A recording medium on which a program for causing a computer to function is recorded.

  In the data processing method according to the second aspect of the present invention, the data processing device embeds the second data in the encoded data including the quantization code and the quantization error obtained by quantizing the first data. A temporary encoded data decoding step of decoding the temporary encoded data by shifting the correspondence between each value of the quantization code and the quantization error in the embedded encoded data by a specific operation amount; In accordance with a predetermined code book, the temporary decoded data decoding step for decoding the temporary decoded data from the temporary encoded data, the quantization code and the quantization error, and the temporary decoded data decoding step Determination of whether or not the temporary encoded data is correct by comparing the quantization code obtained by re-encoding the temporary decoded data with a quantization error In the temporary encoded data decoding step, the operation amount is changed according to the determination result in the determination step, and the temporary encoded data is decoded again. In the determination step, the temporary encoding data is decoded. A data processing method for outputting data corresponding to the manipulated variable as the second data and outputting the provisional decoded data decoded in the provisional decoded data decoding step as true decoded data when it is determined that the data is correct; is there.

  In the second aspect of the present invention, in the embedded encoded data in which the second data is embedded in the encoded data including the quantization code obtained by quantizing the first data and the quantization error, Temporarily encoded data is decoded by shifting the correspondence between each value of the quantization code and the quantization error by a specific operation amount. Further, the temporary decoded data is decoded from the temporary encoded data according to a predetermined code book, and the quantization code, the quantization error, and the temporary decoded data are encoded again. By comparing the quantization code and the quantization error, it is determined whether or not the temporary encoded data is correct. In this case, the manipulated variable is changed according to the determination result as to whether or not the temporary encoded data is correct, the temporary encoded data is decoded again, and the temporary encoded data is determined to be correct. At this time, data corresponding to the manipulated variable is output as the second data, and provisional decoded data is output as true decoded data.

  According to the first and second aspects of the present invention, the second data can be embedded without increasing the data amount of the encoded data obtained by encoding the first data. The first and second data can be accurately decoded from the data in which the second data is embedded.

  FIG. 1 shows an embedded compression / decoding system to which the present invention is applied (a system means a logical collection of a plurality of devices, regardless of whether or not each component device is in the same casing). 2 shows a configuration example of an embodiment.

  The embedded compression / decoding system includes an encoding device 1 and a decoding device 2, and the encoding device encodes, for example, an image as an encoding target, and the decoding device 2 displays the encoding result. The original image is decoded.

  In other words, the encoding apparatus 1 includes an embedded compression encoder 11, and an image to be encoded and information embedded in the image (hereinafter referred to as additional information as appropriate) are supplied thereto. It is like that. The embedded compression encoder 11 compresses and encodes the image (digital image data) according to a predetermined encoding rule, and destroys the encoding rule based on the additional information (digital data). The additional information is embedded, and embedded encoded data is obtained and output. The embedded encoded data output from the embedded compression encoder 11 is recorded on the recording medium 3 formed of, for example, a semiconductor memory, a magneto-optical disk, a magnetic disk, an optical disk, a magnetic tape, a phase change disk, or the like. The data is transmitted via a transmission medium 4 including a terrestrial wave, a satellite line, a CATV (Cable Television) network, the Internet, a public line, and the like, and is provided to the decoding device 2.

  The decoding device 2 includes a decoder 12 in which embedded encoded data provided via the recording medium 3 or the transmission medium 4 is received. Then, the decoder 12 restores the embedded encoded data to encoded data encoded according to a predetermined encoding rule, thereby decoding the additional information embedded therein, and further encoding Decode the data into the original image. The decoded image is supplied to and displayed on a monitor (not shown), for example.

  As additional information, for example, text data related to the original image, audio data, a reduced image obtained by reducing the image, and data irrelevant to the original image can be used. That is, any data (including programs) can be used as the additional information.

  Further, as the additional information, it is possible to use a part of an image as an encoding target supplied to the embedded compression encoder 11. That is, it is possible to supply a part of an image as additional information to the embedded compression encoder 11 and supply the remaining image as an encoding target.

  Next, FIG. 2 shows a configuration example of the embedded compression encoder 11 of FIG.

  The frame memory 21 stores the image data supplied to the embedded compression encoder 11 in units of frames, for example. The additional information memory 22 stores additional information supplied to the embedded compression encoder 11.

  The variable length coding unit 23 performs, for example, Huffman coding on the image data stored in the frame memory 21 as variable length coding, and supplies the resulting coded data to the coding rule destruction unit 25. . In addition, the variable length coding unit 23 creates a Huffman table as described later at the time of Huffman coding, and supplies the Huffman table to the conversion table creation unit 24. Furthermore, the variable length encoding unit 23 supplies HUXman table related information as information necessary for obtaining the Huffman table output to the conversion table creating unit 24 to the MUX (multiplexer) 26.

  The conversion table creation unit 24 creates a conversion table for converting codes in the Huffman table supplied from the variable length coding unit 23 based on the additional information stored in the additional information memory 22. That is, the Huffman table describes the correspondence between values to be Huffman encoded (here, pixel values of an image) and codes (encoded data) of each code length. 24 creates a conversion table for converting codes in such a Huffman table into codes based on additional information. The conversion table created by the conversion table creation unit 24 is supplied to the encoding rule destruction unit 25.

  The encoding rule destruction unit 25 embeds the additional information by destroying the encoding rule in the variable length encoding unit 23 based on the additional information. That is, the encoding rule destruction unit 25 converts (manipulates) the encoded data (code) output from the variable length encoding unit 23 according to the conversion table created by the conversion table creation unit 24 based on the additional information, and changes the variable. Encoded data encoded according to an encoding rule that destroys the encoding rule in the long encoding unit 23 is obtained. The encoded data encoded by the destroyed encoding rule is supplied from the encoding rule destruction unit 25 to the MUX 26 as embedded encoded data in which additional information is embedded in the original encoded data.

  The MUX 26 multiplexes the embedded encoded data from the encoding rule destruction unit 25 and the Huffman table related information from the variable length encoding unit 23 and outputs the multiplexed data. This multiplexed data is provided to the decoding device 2 via the recording medium 3 or the transmission medium 4 as described in FIG.

  Next, FIG. 3 shows a configuration example of the variable length encoding unit 23 of FIG.

  In the frame memory 21 of FIG. 2, each frame of the image data stored therein is sequentially set as the attention frame, for example, in time order, and the image data of the attention frame is read out. The image data of the frame of interest is supplied to the frequency table creation unit 31 and the encoding unit 34.

  The frequency table creation unit 31 creates a frequency table in which each pixel value is associated with the appearance frequency for the pixels constituting the target frame supplied thereto, and supplies the frequency table to the Huffman tree creation unit 32. Further, the frequency table is supplied from the frequency table creation unit 31 to the MUX 26 in FIG. 3 as Huffman table related information.

  Here, in the embodiment of FIG. 3, the frequency table is used as the Huffman table related information, but the Huffman table related information can obtain a Huffman table created by the Huffman table creating unit 33 described later. Therefore, as the Huffman table related information, for example, the Huffman table itself created by the Huffman table creation unit 33 or the like can be used as the Huffman table related information.

  The Huffman tree creation unit 32 creates a so-called Huffman tree based on the frequency table supplied from the frequency table creation unit 31 and supplies the so-called Huffman table creation unit 33. The Huffman table creation unit 33 creates a Huffman table based on the Huffman tree from the Huffman tree creation unit 32. That is, the Huffman table creation unit 33 is assigned to each pixel value in the frame of interest a code having a shorter code length (a code having a longer code length as the appearance frequency is lower) as the appearance frequency of the pixel value is higher. Create a Huffman table. This Huffman table is supplied to the encoding unit 34 and also to the conversion table creating unit 24 in FIG.

  The encoding unit 34 sequentially sets the pixel of the target frame supplied thereto as the target pixel in order of, for example, raster scan, and associates the pixel value of the target pixel in the Huffman table from the Huffman table creation unit 33. And is output as encoded data.

  In the variable-length encoding unit 23 configured as described above, the frequency table creation unit 31 creates a frequency table for the frame of interest and supplies it to the Huffman tree creation unit 32. In the Huffman tree creation unit 32, a Huffman tree is created based on the frequency table from the frequency table creation unit 31 and supplied to the Huffman table creation unit 33. The Huffman table creation unit 33 creates a Huffman table based on the Huffman tree from the Huffman tree creation unit 32 and supplies the Huffman table to the encoding unit 34. In the encoding unit 34, each pixel value of the frame of interest is converted into a code associated with the pixel value in the Huffman table and output as encoded data.

  Next, the variable length coding process in the variable length coding unit 23 will be further described with reference to FIG.

  Assume that the frequency table creation unit 31 creates a frequency table as shown in FIG. 4A, for example. Here, FIG. 4A shows that the pixel values “0”, “1”, “2”, “3”, and “4” are respectively 5, 4, 3, 2, and 1 times in the frame of interest. Indicates that it is appearing.

  With respect to the frequency table of FIG. 4A, in the Huffman tree creation unit 32, for example, as shown in FIGS. 4B to 4E, based on the appearance frequency of each pixel value, in a so-called bottom-up manner. A Huffman tree is created.

  That is, the Huffman tree creation unit 32 selects two pixels having the lowest appearance frequency from the pixel values in the frequency table, and configures one contact point. Further, the Huffman tree creation unit 32 assigns, for example, “0” of the bits “0” or “1” to the lower one of the two selected pixel values, and “ Assign 1 ". Then, the Huffman tree creation unit 32 assigns the added value of the appearance frequencies of the two selected pixel values as the appearance frequency of the contact points to the contact points constituted by the selected two pixel values.

  Accordingly, in the frequency table shown in FIG. 4A, as shown in FIG. 4B, the pixel value “4” having the appearance frequency 1 and the pixel value “3” having the appearance frequency 2 are selected. The contact # 1 is configured, and 3 (= 1 + 2) is assigned as the appearance frequency of the contact # 1. Further, among the selected two pixel values “4” and “3”, the bit value “0” is assigned to the pixel value “4” having the lower appearance frequency, and the pixel value “4” having the higher appearance frequency is assigned. Bit “1” is assigned to “3”.

  Here, in FIGS. 4B to 4E, numbers representing the appearance frequencies are enclosed in parentheses ().

  Note that when the frequency of appearance of the two selected pixel values is the same, the bit “0” or “1” may be assigned to any pixel value. However, it is necessary to determine in advance which pixel value the bit “0” or “1” is assigned to, for example, the bit “0” is assigned to the smaller pixel value.

  In the Huffman tree creation unit 32, the same processing is repeated until one contact converges.

  Therefore, from the state of FIG. 4B, as shown in FIG. 4C, the contact # 1 and the pixel value “2” having the appearance frequency of 3 are selected, and the contact # 1 and the pixel value “2” are selected. “2” is one contact # 2. Further, the appearance frequency of the contact # 2 is 6 (= 3 + 3), and bits “0” and “1” are assigned to the contact # 1 and the pixel value “2”, respectively.

  From the state of FIG. 4C, as shown in FIG. 4D, a pixel value “1” having an appearance frequency of 4 and a pixel value “0” having an appearance frequency of 5 are selected. “1” and “0” are one contact # 3. Further, the appearance frequency of the contact # 3 is 9 (= 4 + 5), and the bits “0” and “1” are assigned to the pixel values “1” and “0”, respectively.

  From the state of FIG. 4D, as shown in FIG. 4E, the contact # 2 with the appearance frequency 6 and the contact # 3 with the appearance frequency 9 are selected, and the contacts # 2 and # 3 are selected. One contact # 4. Further, the appearance frequency of the contact # 4 is 15 (= 6 + 9), and bits “0” and “1” are assigned to the contacts # 2 and # 3, respectively.

  In the state shown in FIG. 4E, since the contacts converge to one contact # 4, the Huffman tree is completed, and the Huffman tree creating unit 32 supplies the Huffman tree to the Huffman table creating unit 33. .

  The Huffman table creation unit 33 recognizes the code assigned to each pixel value by tracing the Huffman tree from the converged contact point in the direction of the pixel value.

  That is, for example, when the Huffman tree shown in FIG. 4E is traced from the contact # 4 in the direction of the pixel value “0” and the bits assigned to the respective contacts (or pixel values) are arranged, “ 0 "→" 0 "→" 0 ". Thereby, the Huffman table creation unit 33 recognizes that the code “000” is assigned to the pixel value “0”. Further, when the Huffman tree shown in FIG. 4E is traced from the contact # 4 in the direction of the pixel value “1” and the bits assigned to the respective contacts are arranged, “0” → “0” → It becomes "1". Thereby, the Huffman table creation unit 33 recognizes that the code “001” is assigned to the pixel value “1”.

  Hereinafter, similarly, the Huffman table 33 recognizes the code assigned to each pixel value, and creates a Huffman table representing the correspondence between the pixel value and the code. Therefore, a Huffman table as shown in FIG. 4F is created from the Huffman tree shown in FIG.

  In addition, the variable length coding unit 23 can perform variable length coding by other methods such as those described in Japanese Patent No. 2977570.

  In the Huffman table shown in FIG. 4 (E), the pixel values “0”, “1”, “2”, “3”, and “4” each having an appearance frequency of 5, 4, 3, 2, and 1 are shown. The codes “11”, “10”, “01”, “001”, and “000” are assigned. Therefore, basically, the higher the appearance frequency, the shorter the code length.

  Incidentally, in the Huffman table of FIG. 4E, the pixel values “0”, “1”, “2”, “3”, and “4” have different appearance frequencies but the pixel values “relatively high appearance frequency”. A 2-bit code is assigned to “0”, “1”, and “2”, and a 3-bit code is assigned to pixel values “3” and “4” that have relatively low appearance frequencies. .

  In this way, in the Huffman table, even if the appearance frequencies are different, codes having the same code length may be assigned, and the relationship between the appearance frequency of the pixel value and the code length of the code assigned to the pixel value Is generally as shown in FIG.

Now, in FIG. 5, if there are x pixel values to which an n-bit code is assigned (where x is an integer ^{equal to} or less than 2 ⁿ ), an n-bit code assignment pattern for the x pixel values is x ! There are only patterns (! Represents the factorial), but in the Huffman table, that x! Only one of the patterns is employed based on the rules for creating the Huffman tree described above.

  On the other hand, even if the n-bit code allocation pattern for x pixel values in the Huffman table is changed, the code amount does not increase. That is, in the Huffman table, even if another n-bit code is assigned to a certain pixel value to which an n-bit code is assigned, the assigned code length remains n bits, so the code amount does not increase.

  Furthermore, even if the n-bit code allocation pattern for the x pixel values in the Huffman table is changed, the code allocation pattern is restored to the original, for example, based on the appearance frequency of the x pixel values. Can do.

  From the above, even if the code allocation pattern for the pixel values to which the code of the same code length is assigned in the Huffman table is changed, that is, even if the coding rule of variable length coding is destroyed, the code amount increases. In addition, the changed assignment pattern can be restored.

  This is because, by changing the code allocation pattern for pixel values based on some information, it is possible to embed information without increasing the overall data amount, and to embed the embedded information in overhead. It means that it can be decrypted without.

  That is, FIG. 6 shows an example of a Huffman table created for pixel values (0 to 255) represented by 8 bits. In FIG. 6, in addition to the pixel value and the code (encoded data) assigned to the pixel value, the appearance frequency of each pixel value is also shown.

In FIG. 6, for example, when attention is paid to seven pixel values of pixel values “12” to “18”, a 9-bit code is assigned to each of these seven pixel values. How to assign a 9-bit code to a value is 7! Only patterns exist. Therefore, int [log ₂ 7!] Bit information can be embedded by changing the 9-bit code allocation pattern for these seven pixel values. Note that int [] means the maximum integer value less than or equal to the value in [].

  Here, FIG. 7A shows the appearance frequencies of the pixel values “12” to “18” in FIG. 6 in a graph, and FIG. 7B shows the pixel value “12” in FIG. "To" 18 "represent 9-bit codes assigned to each.

Now, it is assumed that the code assignment shown in FIG. 7B is changed as shown in FIG. 7C, for example, based on additional information of int [log ₂ 7!] Bits or less.

  That is, in FIG. 7B, for the pixel values “12” to “18”, the signs “110111111”, “110111010”, “110100001”, “110011001”, “11011000”, “011101011”, “010001010” "Is assigned, and in FIG. 7C, the code" 110111111 "assigned to the pixel value" 12 "is assigned to the pixel value" 15 "and the code assigned to the pixel value" 15 ". 110011001 "is assigned to the pixel value" 12 ", the code" 11011000 "assigned to the pixel value" 16 "is assigned to the face value" 17 ", and the code" 011101011 "assigned to the pixel value" 17 "is assigned to the pixel value" The code “010001010” assigned to the pixel value “18” is changed to the pixel value “16”. In FIG. 7, the code assignment to other pixel values is not changed.

  Here, also for pixel values to which codes having other code lengths are assigned, the code assignment pattern can be changed based on the additional information. FIG. 8 shows an example of embedded encoded data as encoded data in which the code allocation pattern in FIG. 6 is changed based on the additional information. FIG. 8 shows the decoded pixel values (embedded encoded data encoded by the variable length decoding of the embedded encoded data together with the embedded encoded data according to the Huffman table shown in FIG. 6. The pixel value decoded by the Huffman table used when obtaining data) is also shown.

  In FIG. 8, for example, when attention is paid to seven pixel values of pixel values “12” to “18” and the embedded encoded data is decoded in accordance with the Huffman table of FIG. 6, the decoding as shown in FIG. A pixel value is obtained. That is, if the pixel values obtained by decoding the embedded encoded data according to the Huffman table used at the time of encoding are referred to as embedded decoded pixel values, the codes “110011001”, “110111010”, “110100001”, “110111111”, “ 010001010 "," 11011000 "," 011101011 "are variable length decoded into embedded decoded pixel values" 15 "," 13 "," 14 "," 12 "," 18 "," 16 "," 17 ", respectively The

  The appearance frequencies of the embedded decoded pixel values “12” to “18” are counted as shown in FIG. 9B. That is, in the Huffman table of FIG. 6, since the codes assigned to the pixel values “12” to “18” are changed as shown in FIG. 7B and FIG. The appearance frequencies of “12”, “15”, “16” to “18” do not match those shown in FIG. 7A, specifically, pixel values before variable-length encoding. “15”, “12”, “18”, “16”, “17” respectively correspond to the appearance frequencies.

  However, the appearance frequencies of the original (before variable length coding) pixel values “12” to “18” and the appearance frequencies of the pixel values “12” to “18” obtained as a result of variable length decoding match. Therefore, as described above, it is strange that the appearance frequency of the pixel value is different between before the variable length coding and after the variable length decoding.

  Therefore, the appearance frequency of the original pixel value shown in FIG. 7A is compared with the appearance frequency of the embedded decoded pixel value shown in FIG. The allocation pattern of the code (variable length code) changed based on the information can be restored. In other words, the embedded encoded data can be returned to the encoded data encoded based on the Huffman table. Then, the embedded additional information can be decoded from the method of changing the code allocation pattern when the embedded encoded data is returned to the encoded data, and the original encoded data is variable-length decoded. Thus, the original pixel value can be decoded.

  Specifically, by comparing the appearance frequency of the original pixel value shown in FIG. 7A and the appearance frequency of the embedded decoded pixel value shown in FIG. 9B, the embedded decoded pixel value “15” is compared. ”,“ 13 ”,“ 14 ”,“ 12 ”,“ 17 ”,“ 18 ”,“ 16 ”, and the appearance frequencies of the original pixel values“ 12 ”to“ 18 ”respectively match. Can be detected.

  As a result, the embedded encoded data values “110011001”, “110111010”, “15”, “13”, “14”, “12”, “17”, “18”, “16” are obtained. “110100001”, “110111111”, “010001010”, “11011000”, “011101011” are assigned to the original pixel values “12” to “18” in the Huffman table of FIG. 6 before the additional information is embedded. It can be seen that the codes are “110111111”, “110111010”, “110100001”, “110011001”, “11011000”, “011101011”, “010001010”.

  Therefore, as shown in FIG. 9C, the appearance frequency of the original pixel value shown in FIG. 7A matches the appearance frequency of the embedded decoded pixel value shown in FIG. By converting the embedded encoded data, the encoded data encoded according to the Huffman table shown in FIG. 6 is restored, and added from the correspondence between the embedded encoded data and the recovered encoded data. Information can be decoded. Furthermore, by performing variable length decoding on the restored encoded data, the original pixel value can be decoded as shown in FIG.

  In the embedded compression encoder 11 of FIG. 2, as described above, the embedded encoding process capable of decoding without overhead is performed without increasing the data amount.

  Next, the embedded encoding process performed by the embedded compression encoder 11 of FIG. 2 will be further described with reference to the flowchart of FIG.

  In step S <b> 1, the variable length encoding unit 23 performs variable length encoding on the pixel value of each pixel of the target frame supplied from the frame memory 21, and outputs the encoded data obtained as a result to the encoding rule destruction unit 25. To do. Further, the variable length coding unit 23 (FIG. 3) outputs the frequency table created by the frequency table creation unit 31 to the MUX 26 as Huffman table related information in the variable length coding in step S1, and also Huffman. The Huffman table created by the table creation unit 33 is output to the conversion table creation unit 24.

  In step S2, the conversion table creation unit 24 changes the code allocation pattern of the same code length in the Huffman table from the variable length coding unit 23 based on the additional information supplied from the additional information memory 22, Conversion table creation processing for creating a conversion table in which the code before the change and the code after the change are associated is performed. This conversion table is supplied from the conversion table creation unit 24 to the encoding rule destruction unit 25.

  In step S3, the encoding rule destruction unit 25 performs encoded data conversion processing for converting the encoded data from the variable-length encoding unit 23 into embedded encoded data according to the conversion table from the conversion table creation unit 24. The embedded encoded data obtained as a result is supplied to the MUX 26.

  In step S4, the MUX 26 multiplexes and outputs the embedded encoded data of the frame of interest from the encoding rule destruction unit 25 and the Huffman table related information from the variable length encoding unit 23, and proceeds to step S5.

  In step S5, it is determined whether or not the frame memory 21 stores a frame next to the frame of interest. If it is determined in step S5 that the next frame is stored in the frame memory 21, the next frame is newly set as a frame of interest, the process returns to step S1, and the same processing is repeated thereafter. .

  If it is determined in step S5 that the next frame is not stored in the frame memory 21, the embedded encoding process is terminated.

  Next, the conversion table creation process performed by the conversion table creation unit 24 in FIG. 2 in step S2 in FIG. 10 will be further described with reference to the flowchart in FIG.

  In the conversion table creation process, first, in step S11, the conversion table creation unit 24 pays attention to a certain code length of codes in the Huffman table from the variable length coding unit 23, and the number of codes of the code length of interest. Recognize That is, the conversion table creation unit 24 recognizes the number of codes having the same code length as the target code length.

Further, in step S12, the conversion table creation unit 24 calculates the number of bits of additional information that can be embedded in the code of the target code length based on the number of codes of the target code length recognized in step S11. That is, assuming that the number of codes of the target code length is x, the conversion table creation unit 24 calculates y = int [log ₂ (x!)] To obtain the number of bits y of additional information that can be embedded.

  In step S13, the conversion table creation unit 24 reads the additional information for the number of bits y obtained in step S12 from the additional information memory 22, and proceeds to step S14. In step S14, the conversion table creation unit 24 creates a conversion table for the code of interest code length based on the additional information read from the additional information memory 22 in step S13. That is, the conversion table creation unit 24 changes the allocation pattern of the x codes of the target code length assigned to each pixel value in the Huffman table based on the additional information of y bits, and the code before the change And the code after the change are created as a conversion table for the code of the target code length.

  Thereafter, in step S15, the conversion table creation unit 24 determines whether the additional information is still stored in the additional information memory 22, and if it is determined that the additional information is not stored, the conversion table creation unit 24 generates the attention frame until then. The converted table is output to the encoding rule destruction unit 25 (FIG. 2), and the conversion table creation process is terminated.

  If it is determined in step S15 that the additional information is still stored in the additional information memory 22, the process proceeds to step S16, where the conversion table creation unit 24 converts all codes of codes in the Huffman table. It is determined whether the table has been created. If it is determined that the table has not been created, the process returns to step S11. In this case, in step S11, one of the code lengths for which no conversion table has been created is newly set as the target code length, and the same processing is repeated thereafter.

  On the other hand, if it is determined in step S16 that a conversion table has been created for all codes of codes in the Huffman table, the conversion table creation process ends.

  Next, the encoded data conversion processing performed by the encoding rule destruction unit 25 in FIG. 2 in step S3 in FIG. 10 will be further described with reference to the flowchart in FIG.

  As described above, the variable length coding unit 23 performs variable length coding on the pixel value of the target pixel with each pixel in the target frame as the target pixel in the raster scan order, and sequentially obtains the encoded data obtained as a result. Output. Now, assuming that the encoded data output by the variable length encoding unit 23 for the target pixel is referred to as target encoded data, the encoding rule destruction unit 25 firstly encodes the target encoded data in step S21. The length is recognized and the process proceeds to step S22.

  In step S22, the encoding rule destruction unit 25 converts the encoded data of interest in accordance with the conversion table for the code length of the encoded data of interest recognized in step S21, and MUX 26 (FIG. 2) as embedded encoded data. Output to.

  Then, the process proceeds to step S23, where the encoding rule destruction unit 25 determines whether there is next encoded data from the variable length encoding unit 23, and if it is determined that there is, the next encoded data is determined. Then, the process returns to step S21 as newly encoded data of interest, and the same processing is repeated thereafter.

  If it is determined in step S23 that there is no next encoded data, the encoded data conversion process ends.

  Next, in the embedded compression encoder 11 shown in FIG. 2, the encoded data obtained by the variable length encoding in the variable length encoding unit 23 is converted according to the conversion table to embed the additional information embedded therein. In addition, for example, the variable length encoding unit 23 creates a Huffman table in which the correspondence between the pixel value and the code is changed (hereinafter, referred to as a changed Huffman table as appropriate) By using the modified Huffman table, it is possible to perform variable length coding and embedding additional information at the same time, that is, variable length coding based on a broken coding rule.

  That is, FIG. 13 shows a configuration example of the embedded compression encoder 11 that performs such an embedded encoding process. In the figure, portions corresponding to those in FIG. 2 or FIG. 3 are denoted by the same reference numerals, and description thereof will be omitted below as appropriate. That is, the embedded compression encoder 11 of FIG. 13 is not provided with the encoding rule destruction unit 25. Instead, the variable length encoding unit 23 includes the encoding rule destruction unit 41. The configuration is the same as in FIG.

  The encoding rule destruction unit 41 is supplied with a Huffman table output from the Huffman table creation unit 33 and a conversion table output from the conversion table creation unit 24.

  The encoding rule destruction unit 41 changes the code (encoded data) associated with each pixel value in the Huffman table to the embedded encoded data associated with the code in the conversion table, thereby Then, the Huffman table from the Huffman table creation unit 33 is changed to a changed Huffman table that is a Huffman table in which pixel values and embedded encoded data are associated with each other. The changed Huffman table is supplied to the encoding unit 34, and the encoding unit 34 converts the pixel value according to the changed Huffman table.

  Therefore, in FIG. 13, the encoded unit 34 outputs embedded encoded data.

  Next, FIG. 14 shows a configuration example of the decoder 12 of FIG. 1 when the embedded compression encoder 11 is configured as shown in FIG. 2 or FIG.

  The DEMUX (demultiplexer) 51 separates the data provided from the embedded compression encoder 11 shown in FIG. 2 or FIG. 13 into embedded encoded data and Huffman table related information, and embeds each frame. The encoded data is supplied to the variable length decoding unit 52 and the encoding rule restoration unit 55, and the Huffman table related information is supplied to the variable length decoding unit 52, the inverse conversion table creation unit 54, and the variable length decoding unit 56. .

  The variable length decoding unit 52 sequentially sets the embedded encoded data output from the DEMUX 51 and the frame of the Huffman related information as the frame of interest, and the variable length encoding of FIG. 3 from the frequency table as the Huffman table related information for the frame of interest. As in the case of the unit 23, a Huffman table is created and supplied to the inverse conversion table creation unit 54.

  Here, the Huffman table created from the frequency table as the Huffman table related information is the same as that used for variable-length coding in the variable-length coding unit 23, and is hereinafter referred to as a normal Huffman table as appropriate. . In addition, the frequency table as the Huffman table related information used for obtaining the normal Huffman table is hereinafter referred to as a positive frequency table as appropriate.

  The variable length decoding unit 52 further performs variable length decoding of the embedded encoded data for the frame of interest according to the normal Huffman table, and generates a decoded value of the pixel value obtained as a result, that is, an embedded decoded pixel value, as a Huffman table. To the unit 53.

  The Huffman table generating unit 53 generates a Huffman table for variable-length encoding the embedded decoded pixel value for the target frame supplied from the variable-length decoding unit 52, and the Huffman table and the Huffman table are generated in the process of generating the Huffman table The frequency table to be processed is supplied to the inverse conversion table creation unit 54.

  Here, the Huffman table generated by the Huffman table generation unit 53 is a variable length encoding of embedded decoding pixel values (embedded encoded data obtained by variable length decoding using a normal Huffman table). The encoded data cannot be correctly decoded. Therefore, the Huffman table obtained from the embedded decoded pixel value is hereinafter referred to as an erroneous Huffman table as appropriate for the correct Huffman table. In addition, the frequency table used to obtain the erroneous Huffman table is hereinafter referred to as an erroneous frequency table as appropriate with respect to the correct frequency table.

  Based on the normal Huffman table from the variable length decoding unit 52, the normal frequency table as the Huffman table related information from the DEMUX 51, the erroneous Huffman table and the erroneous frequency table from the Huffman table generating unit 53, the inverse conversion table creation unit 54 An inverse conversion table for converting the embedded encoded data into the original encoded data is created. That is, the reverse conversion table creation unit 54 creates the same reverse conversion table as the conversion table created by the conversion table creation unit 24 of FIG. 2 (or FIG. 13). This inverse conversion table is supplied to the encoding rule restoration unit 55.

  The encoding rule restoration unit 55 encodes the embedded encoded data from the DEMUX 51 based on the normal Huffman table as an encoding rule based on the inverse conversion table supplied from the inverse conversion table creation unit 54. Restore to. Further, the encoding rule restoration unit 55 decodes the additional information embedded in the embedded encoded data based on the correspondence relationship between the embedded encoded data and the recovered encoded data, that is, the inverse conversion table. Then, the encoding rule restoration unit 55 outputs the restored encoded data to the variable length decoding unit 56, and outputs the decoded additional information as decoded additional information.

  The variable length decoding unit 56 creates a normal Huffman table from the Huffman related information supplied from the DEMUX 51, and variable length decodes the encoded data supplied from the encoding rule restoration unit 55 based on the normal Huffman table. The decoded pixel value obtained as a result is output.

  Next, FIG. 15 shows a configuration example of the variable length decoding unit 52 of FIG.

  The embedded encoded data output from the DEMUX 51 is supplied to the decoding unit 63. Similarly, the frequency table as the Huffman table related information output from the DEMUX 51 is supplied to the Huffman tree creating unit 61. It has become.

  Similar to the Huffman tree creation unit 32 in FIG. 3, the Huffman tree creation unit 61 creates a Huffman tree from the frequency table for the frame of interest and supplies it to the Huffman table creation unit 62.

  The Huffman table creation unit 62 creates a Huffman table based on the Huffman tree from the Huffman tree creation unit 61 and supplies it to the decoding unit 63, as with the Huffman table creation unit 33 in FIG.

  In accordance with the Huffman table from the Huffman table creation unit 62, the decoding unit 63 decodes the embedded encoded data supplied thereto into a pixel value (embedded decoded pixel value) and outputs it.

  14 is configured in the same manner as the variable length decoding unit 52 shown in FIG.

  Next, FIG. 16 shows a configuration example of the Huffman table generation unit 53 of FIG.

  In the Huffman table generation unit 53, the embedded decoded pixel value output from the variable length decoding unit 52 is supplied to the frequency table 71. The frequency table 71, the Huffman tree creation unit 72, or the Huffman table table creation unit 73 performs the same processing as in the frequency table creation unit 31, the Huffman tree creation unit 32, or the Huffman table creation unit 33 in FIG.

  As a result, the frequency table creation unit 71 creates an error frequency table (not a frequency table for the original pixel values), which is a frequency table for the embedded decoded pixel value of the frame of interest, and the inverse conversion table creation unit 54 (FIG. 14). In addition, the Huffman table creation unit 73 converts an embedded decoded pixel value of the target frame into a code having a code length corresponding to the appearance frequency, and an erroneous Huffman table (in the Huffman table for the original pixel value). Is created) and supplied to the inverse conversion table creation unit 54.

  Next, FIG. 17 shows a configuration example of the inverse conversion table creation unit 54 of FIG.

  As described above, a normal Huffman table, a normal frequency table, an erroneous Huffman table, and an erroneous frequency table are supplied to the reverse conversion table creation unit 54. Based on the table, as described with reference to FIGS. 6 to 9, the inverse relationship in which the correspondence relationship between the embedded encoded data and the encoded data encoded by the normal Huffman table is recognized and the corresponding relationship is described. A table is created.

  That is, the correct Huffman table and the incorrect Huffman table are supplied to the code association unit 82, and the correct frequency table and the incorrect frequency table are supplied to the comparison unit 81.

  Then, the comparison unit 81 compares the correct frequency table and the incorrect frequency table, and detects pixel values having the same frequency in the correct frequency table and the incorrect frequency table for each code of each code length. Furthermore, the comparison unit 81 creates a corresponding pixel value table in which pixel values having the same frequency are associated in the correct frequency table and the incorrect frequency table, and supplies the corresponding pixel value table to the code association unit 82.

  The code association unit 82 retrieves the pixel values associated in the corresponding pixel value table from the comparison unit 81 from the correct Huffman table and the erroneous Huffman table, and associates the codes associated with the retrieved pixel values. Is created to create an inverse conversion table. This inverse conversion table is supplied to the encoding rule restoration unit 55.

  Next, decoding processing by the decoder 12 in FIG. 14 will be described with reference to the flowchart in FIG.

  A DEMUX (demultiplexer) 51 separates data supplied thereto into embedded encoded data and Huffman table related information, and converts the embedded encoded data for each frame into a variable length decoding unit 52 and an encoding rule restoration And the Huffman table related information is supplied to the variable length decoding unit 52, the inverse conversion table creating unit 54, and the variable length decoding unit 56.

  In step S31, the variable length decoding unit 52 sequentially uses the embedded encoded data and the Huffman related information frame output from the DEMUX 51 as a frame of interest, and performs variable length decoding of the embedded encoded data for the frame of interest.

  That is, the variable length decoding unit 52 creates a Huffman table (correct Huffman table) from the frequency table (correct frequency table) as the Huffman table related information for the frame of interest, and supplies it to the inverse conversion table creation unit 54. Further, the variable length decoding unit 52 performs variable length decoding on the embedded encoded data for the frame of interest according to the normal Huffman table, and supplies the embedded decoded pixel value obtained as a result to the Huffman table generating unit 53.

  In step S32, the Huffman table generating unit 53 creates a frequency table (error frequency table) from the embedded decoded pixel values for the frame of interest supplied from the variable length decoding unit 52, and further, the error frequency table. From this, a false Huffman table is created. The error frequency table and the error Huffman table are supplied to the inverse conversion table creation unit 54.

  In step S33, the reverse conversion table creating unit 54 converts the reverse conversion table into the reverse Huffman table, the normal frequency table as the Huffman table related information, the erroneous Huffman table, and the erroneous frequency table as described in FIG. Created and supplied to the encoding rule restoration unit 55.

  In step S34, the encoding rule restoration unit 55 refers to the correspondence between the embedded encoded data and the encoded data for each code length in the inverse conversion table supplied from the inverse conversion table creating unit 54, thereby embedding. The additional information embedded in the encoded data is decoded, and the decoded additional information obtained as a result is output.

  Further, the encoding rule restoration unit 55 proceeds to step S35, refers to the inverse conversion table, converts the embedded encoded data from the DEMUX 51 into encoded data, and supplies the encoded data to the variable length decoding unit 56.

  The variable length decoding unit 56 creates a normal Huffman table from the Huffman related information supplied from the DEMUX 51, and variable length decodes the encoded data supplied from the encoding rule restoration unit 55 based on the normal Huffman table. The decoded pixel value obtained as a result is output.

  Thereafter, the process proceeds to step S37, where it is determined whether the embedded encoded data and the Huffman table related information for the next frame are output from the DEMUX 51. If it is determined that the output is performed, the next frame is The frame is newly set as the attention frame, and the process returns to step S31. Thereafter, the same processing is repeated.

  If it is determined in step S37 that the embedded encoded data and the Huffman table related information for the next frame have not been output from the DEMUX 51, the decoding process ends.

  As described above, as the Huffman table related information, the Huffman table itself can be used instead of the frequency table. However, in this case, the Huffman table as the Huffman table related information recognizes the appearance frequency of pixel values to which the code (encoded data) of the same code length is assigned by referring to the Huffman table. Must be created (with rules that can be recognized).

  In the above case, the Huffman table is created in units of frames. However, the Huffman table can be created in units of fields or in units of two frames or more. .

  Further, the Huffman table is created in advance using, for example, predetermined image data, and the embedded compression encoder 11 and the decoder 12 are caused to perform processing using the previously created Huffman table. It is also possible. However, in this case, when the Huffman table is created from the image data to be processed and the Huffman table does not match the Huffman table created from the predetermined image data, the decoder 12 converts the embedded encoded data into Since it becomes difficult to restore the encoded data, the additional information needs to be embedded only in the image data from which the same Huffman table as the Huffman table created from the predetermined image data is obtained.

  Next, FIG. 19 shows another configuration example of the embedded compression encoder 11 of FIG.

  The encoding unit 91 is supplied with image data as an encoding target. The encoding unit 91 encodes the image data according to a predetermined encoding rule, and the resulting encoding is performed. The data is output to the embedding unit 92.

  Additional information is supplied to the embedding unit 92, and the embedding unit 92 operates the encoded data supplied from the encoding unit 91 based on the additional information, thereby destroying the encoded data. By making the encoded data encoded according to the rule, the additional information is embedded in the encoded data.

  Next, the process (embedded encoding process) of the embedded compression encoder 11 of FIG. 19 will be described with reference to the flowchart of FIG.

  The encoding unit 91 is supplied with image data as an encoding target, for example, in units of frames. The encoding unit 91 sequentially sets each frame as a frame of interest, and in step S41, the image data of the frame of interest is converted. Encoding is performed according to a predetermined encoding rule. Then, the encoding unit 91 outputs the encoded data obtained as a result to the embedding unit 92.

  In step S42, the embedding unit 92 embeds the additional information by destroying the encoding rule in the encoding unit 91 based on the additional information. That is, the embedding unit 92 operates the encoded data supplied from the encoding unit 91 based on the additional information, and thereby embeds the additional information in the encoded data. Furthermore, the embedding unit 92 outputs embedded encoded data obtained by embedding additional information in the encoded data, and the process proceeds to step S43.

  In step S43, the encoding unit 91 determines whether there is a frame to be encoded next. If it is determined that there is a next frame, the next frame is newly set as a frame of interest, and the process returns to step S41. The same process is repeated.

  If it is determined in step S43 that there is no next frame to be encoded, the embedded encoding process is terminated.

  Next, FIG. 21 shows a configuration example of the encoding unit 91 of FIG.

  In the encoding unit 91 in FIG. 21, for example, the pixel values constituting the image data are expressed in RGB (Red, Green, Blue), and each pixel value is vector-quantized in the RGB color space. , A code representing a centroid vector (hereinafter referred to as VQ code as appropriate), an error of the pixel value represented by the centroid vector corresponding to the code with respect to the original pixel value (hereinafter referred to as VQ residual as appropriate), The code book used for vector quantization is output as encoded data.

  That is, the image data to be encoded is supplied to the frame memory 101, and the frame memory 101 sequentially stores the image data supplied thereto, for example, in units of frames.

  The code book creation unit 102 sequentially sets the frames of the image data stored in the frame memory 101 as a target frame, and the code book used for vector quantization in the color space from the pixel value of each pixel constituting the target frame. Is created by, for example, a so-called LBG algorithm. This code book is supplied to the vector quantization unit 103 and is output as (a part of) encoded data.

  The vector quantization unit 103 reads the target frame from the frame memory 101, and sequentially sets each pixel constituting the target frame as the target pixel in the raster scan order, for example. Then, the vector quantization unit 103 performs vector quantization on the pixel value of the pixel of interest using the code book from the code book creation unit 102, and the resulting VQ code and VQ residual are converted into encoded data (one). Part).

  In the coding unit 91 configured as described above, as shown in FIG. 22, the code book creating unit 102 creates a code book from the pixel values of each pixel constituting the frame of interest, and the vector quantization unit 103 To be supplied. The vector quantization unit 103 performs vector quantization on the pixel value of the target pixel using the code book from the code book creation unit 102.

  That is, the vector quantization unit 103 detects a centroid vector representing a point having the shortest distance from the point on the RGB space represented by the pixel value of the target pixel from the code book, and VQ representing the centroid vector. Output code. Further, the vector quantization unit 103 obtains a difference between the centroid vector represented by the VQ code and a vector corresponding to a point on the RGB space represented by the pixel value of the target pixel, and obtains the difference vector obtained as a result. , Output as a VQ residual.

  Then, the encoding unit 91 outputs the code book obtained as described above, and the VQ code and VQ residual for each pixel of the target frame as encoded data for the target frame.

  Next, FIG. 23 illustrates a configuration example of the embedding unit 92 in FIG. 19 when the encoding unit 91 is configured as illustrated in FIG.

  The VQ code, VQ residual, and code book of the frame of interest output from the encoding unit 91 are supplied and stored in the VQ code memory 111, the VQ residual memory 112, and the code book memory 113, respectively.

  The VQ code and code book respectively stored in the VQ code memory 111 and the code book memory 113 are read out from the code and supplied to the compression unit 115.

On the other hand, the line shift unit 114 sequentially sets each line (horizontal line) of the frame of interest, for example, from the top to the bottom as the line of interest, and the VQ residual stored in the VQ residual memory 112 for the line of interest. Is read. Further, the line shift unit 114 receives additional information of the number of bits represented by int [log ₂ x], where x is the number of pixels constituting the target line, and for example, as shown in FIG. By shifting the VQ residual of the target line to the right by the number of pixels corresponding to the additional information, the additional information is embedded in the target line. That is, if attention is paid to a certain pixel on the target line due to this line shift, basically, the VQ residual for the pixel is different from that obtained by vector quantization. The rule is destroyed. And additional information is embedded by destroying an encoding rule.

  Thereafter, the line shift unit 114 supplies the VQ residual of the target line in which the additional information is embedded to the compression unit 115.

  Here, when the VQ residual of the target line is shifted to the right, the VQ residual of the rightmost pixel of the target line is shifted to the leftmost pixel of the target line. Similarly, when the VQ residual of the target line is shifted to the left, the VQ residual of the leftmost pixel is shifted to the rightmost pixel.

  The compression unit 115 compresses the VQ code, VQ residual, and code book supplied thereto using, for example, spatial correlation, entropy bias, and the like, and outputs the compression result to the MUX 116. The MUX 116 multiplexes and outputs the compression results of the VQ code, the VQ residual, and the code book from the compression unit 115.

  Next, referring to the flowchart of FIG. 25, the embedded compression code of FIG. 19 when the encoding unit 91 is configured as shown in FIG. 21 and the embedding unit 92 is configured as shown in FIG. The process (embedded encoding process) of the encoder 11 will be described.

  In the encoding unit 91 (FIG. 21), a predetermined frame stored in the frame memory 101 is set as a target frame, and in step S51, each pixel of the target frame is subjected to vector quantization as described above. Encoded. The VQ code, VQ residual, and code book as encoded data obtained as a result of vector quantization of the frame of interest are supplied to the embedding unit 92.

  In the embedding unit 92 (FIG. 23), the VQ code, VQ residual, and code book for the frame of interest from the encoding unit 91 are supplied to the VQ code memory 111, the VQ residual memory 112, and the code book memory 113, respectively. Is remembered. In step S52, the line shift unit 114 determines whether all lines of the frame of interest in which the VQ residual is stored in the VQ residual memory 112 have been processed.

  If it is determined in step S52 that all the lines of the target frame for which the VQ residual is stored in the VQ residual memory 112 have not yet been processed, the line shift unit 114 performs processing on the target frame. The line located at the top of the unexisting lines is newly set as the attention line, and the process proceeds to step S53.

  In step S53, the line shift unit 114 right-shifts each pixel having the pixel value of the VQ residual of the target line by the number of pixels corresponding to the additional information, thereby embedding the additional information in the target line, It supplies to the compression part 115. Thereafter, the process returns to step S52, and the same processing is repeated thereafter.

  If it is determined in step S52 that all lines of the frame of interest in which the VQ residual is stored in the VQ residual memory 112 have been processed, the process proceeds to step S54, and the compression unit 115 stores the data in the VQ code memory 111. The VQ code for the noticed frame of interest is read out, and the code book for the noticed frame stored in the codebook memory 113 is read out. Then, the compression unit 115 compresses the VQ code and codebook, and further the VQ residual in which the additional information from the line shift unit 114 is embedded, and supplies the compressed VQ code to the MUX 116.

  In step S55, the MUX 116 multiplexes and outputs the VQ code, VQ residual, and code book from the compression unit 115, and the process proceeds to step S56.

  In step S56, it is determined whether or not there is a frame to be encoded next in the encoding unit 91. If it is determined that there is a frame to be encoded next, the frame to be encoded next is a new frame of interest. Returning to S51, the same processing is repeated thereafter.

  If it is determined in step S56 that there is no next frame to be encoded, the embedded encoding process is terminated.

  Next, FIG. 26 shows a configuration example of the decoder 12 in FIG. 1 when the embedded compression encoder 11 is configured as shown in FIG.

  The embedded encoded data output from the embedding unit 92 in FIG. 19 is supplied to the encoding rule restoration unit 121.

  The encoding rule restoration unit 121 restores the embedded encoded data to the encoded data encoded according to the encoding rule in the encoding unit 91 of FIG. 19, and thereby embedded in the embedded encoded data. Decode additional information.

  In other words, the encoding rule restoration unit 121 performs an operation based on the parameters supplied from the parameter controller 124 on the embedded encoded data, so that encoded data candidates (hereinafter, referred to as temporary encoded data as appropriate). Ask for. Further, the encoding rule restoration unit 121 selects additional information candidates embedded in the embedded encoded data (hereinafter referred to as temporary decoded additional information) based on an operation of recovering the embedded encoded data to temporary encoded data. Decrypt. The provisionally encoded data is supplied to the decoding unit 122 and the determination unit 123, and the provisional decoding additional information is supplied to the determination unit 123.

  The decoding unit 122 performs a decoding process based on the encoding rule in the encoding unit 91 in FIG. 19 on the provisionally encoded data from the encoding rule restoration unit 121, so that the original pixel value candidates (hereinafter, As appropriate, this is decoded). The provisional decoded pixel value is supplied to the determination unit 123.

  The determination unit 123 controls the parameter controller 124 to supply the encoding rule restoration unit 121 with one or more parameter values, and obtains one corresponding to each of the one or more parameter values. A correct temporary decoded pixel value (matching the original pixel value) is determined from the above temporary decoded pixel values. Further, the determination unit 123 obtains a correct decoded pixel value from the one or more provisional decoding additional information supplied from the encoding rule restoration unit 121 in association with each of the one or more provisional decoding pixel values. Is selected as correct decoding additional information, and the correct temporary decoded pixel value and temporary decoding additional information are output as the final decoding result (decoded pixel value and additional decoding information) of the pixel value and additional information, respectively. To do.

  The parameter controller 124 supplies predetermined parameters for operating the embedded encoded data to the encoding rule restoration unit 121 under the control of the determination unit 123.

  Next, processing (decoding processing) of the decoder 12 in FIG. 26 will be described with reference to the flowchart in FIG.

  First, the determination unit 123 sets parameters to be given to the encoding rule restoration unit 121 by controlling the parameter controller 124 in step S61. As a result, the parameter controller 126 supplies parameters according to the control of the determination unit 123 to the encoding rule restoration unit 121.

  In step S62, the encoding rule restoration unit 121 converts the embedded encoded data into temporary encoded data by performing an operation based on the parameters supplied from the parameter controller 124, and decodes the decoding unit 122 and the determination unit. 123. Further, the encoding rule restoration unit 121 decodes the additional information embedded in the embedded encoded data based on an operation for recovering the embedded encoded data to temporary encoded data, that is, based on the parameter from the parameter controller 124. The provisional decoding additional information is supplied to the determination unit 123.

  In step S63, the decoding unit 122 decodes the temporary encoded data from the encoding rule restoration unit 121 based on the encoding rule in the encoding unit 91 in FIG. 19, and the pixel value obtained as a result of the decoding is The provisional decoded pixel value is supplied to the determination unit 123.

  In step S64, the determination unit 123 determines whether or not the provisional decoded pixel value from the decoding unit 122 is a correct decoding result (matches the original pixel value), and determines that they do not match. If so, the process returns to step S61. In this case, the determination unit 123 sets a new value as a parameter to be output to the parameter controller 124 in step S61, and thereafter repeats the same processing.

  On the other hand, when it is determined in step S64 that the temporary decoded pixel value from the decoding unit 122 is a correct decoding result, the process proceeds to step S65, and the determination unit 123 sets the temporary decoded pixel value to the original pixel value. It outputs as a decoded pixel value which is a decoding result. Further, the determination unit 123 outputs the provisional decoding additional information when the decoded pixel value is obtained as decoding additional information that is a decoding result of the embedded additional information, and proceeds to step S66.

  In step S66, it is determined whether there is still embedded encoded data to be decoded. If it is determined that there is, the process returns to step S61, and the same processing is performed for the embedded encoded data to be decoded. Repeated.

  If it is determined in step S66 that there is no embedded encoded data to be decoded, the decoding process ends.

  Next, FIG. 28 illustrates a configuration example of the encoding rule restoration unit 121 in FIG. 26 when the embedding unit 92 in FIG. 19 is configured as illustrated in FIG.

  Data output from the MUX 116 in FIG. 23 is supplied to the DEMUX 131, and the DEMUX 131 separates the data supplied thereto into a compressed VQ code, a VQ residual, and a code book, and supplies them to the decompression unit 132. The decompression unit 132 decompresses the compressed VQ code, VQ residual, and code book from the DEMUX 131, and decompresses the decompressed VQ code, VQ residual, and code book into the VQ code memory 133 and the VQ residual memory 134. , To the code book memory 135, respectively.

  The VQ code memory 133, the VQ residual memory 134, and the code book memory 135 store the VQ code, VQ residual, and code book from the decompression unit 132, respectively, in units of frames.

  The line shift unit 136 sets each line of the frame stored in the VQ residual memory 134 as a target line sequentially from the top to the bottom, for example, and the VQ stored in the VQ residual memory 134 for the target line. Read the residual. Further, the line shift unit 136 receives an integer value in a range from 0 to x as a parameter from the parameter controller 124 (FIG. 26), where x is the number of pixels constituting the target line, and corresponds to the parameter. The VQ residual of the target line is shifted to the left by the number of pixels. Then, the line shift unit 136 outputs the VQ residual of each line after the shift as temporary encoded data together with the VQ code stored in the VQ code memory 133 and the code book stored in the code book memory 135. To do.

  Further, the line shift unit 136 outputs the parameter value from the parameter controller 124 as provisional decoding additional information.

  Next, FIG. 29 illustrates a configuration example of the decoding unit 122 in FIG. 26 when the encoding unit 91 in FIG. 19 is configured as illustrated in FIG.

  The decoding unit 122 decodes the pixel value by performing vector inverse quantization based on one frame of the VQ code, the VQ residual, and the code book as the encoded data supplied from the encoding rule restoration unit 121. .

  That is, the vector inverse quantization unit 141 is supplied with a VQ code and a code book, and the vector inverse quantization unit 141 detects a centroid vector corresponding to the VQ code from the code book, The centroid vector is supplied to the adding unit 142. In addition to the centroid vector supplied from the vector dequantization unit 141, the addition unit 142 is also supplied with a difference vector as a VQ residual, and the addition unit 142 includes the centroid vector and the difference vector. And add. Then, the adding unit 142 outputs a pixel value having R, G, and B values for each component of the vector obtained as a result of the addition as a provisional decoded pixel value.

  Next, FIG. 30 illustrates a configuration example of the determination unit 123 in FIG. 26 when the encoding rule restoration unit 121 and the decoding unit 122 are configured as illustrated in FIGS. 28 and 29, respectively.

  The memory 151 is supplied with the provisional decoding additional information from the encoding rule restoration unit 121 and the provisional decoding pixel value from the decoding unit 122. The memory 151 stores the provisional decoding additional information and the provisional decoding. The pixel value is temporarily stored, and the stored temporary decoding additional information and temporary decoding pixel value are read according to the control of the correctness determination unit 154, and are output as decoding additional information and decoded pixel value, respectively.

  The encoding unit 152 is supplied with the codebook of the temporary encoded data output from the encoding rule restoration unit 121 and the temporary decoded pixel value output from the decoding unit 122. 152 encodes the provisional decoded pixel value in the same manner as the encoding unit 91 in FIG. That is, the encoding unit 152 performs vector quantization on the provisional decoded pixel value using the code book from the encoding rule restoration unit 121, and supplies the VQ code and VQ residual obtained as a result to the comparison unit 153. Here, hereinafter, the VQ code and the VQ residual obtained by vector quantization of the temporary decoded pixel value will be referred to as a temporary VQ code and a temporary VQ residual, respectively.

  In addition to the temporary VQ code and the temporary VQ residual output from the encoding unit 152, the comparison unit 153 is supplied with the VQ code and the VQ residual in the temporary encoded data output from the decoding unit 122. ing. The comparison unit 153 compares the temporary VQ code with the VQ code of the temporarily encoded data, compares the temporary VQ residual with the VQ residual of the temporarily encoded data, and determines whether each comparison result is correct or incorrect. To the unit 154.

  The correctness determination unit 154 controls the parameter controller 124 to cause the parameter controller 124 to supply the number of bits to be shifted to the encoding rule restoration unit 121 as a parameter. Further, the correctness determination unit 154 is based on the comparison result between the temporary VQ code from the comparison unit 153 and the VQ code of the temporarily encoded data, and the comparison result between the temporary VQ residual and the VQ residual of the temporarily encoded data. Then, it is determined whether the temporary decoded pixel value is correct as a decoding result, and reading of the temporary decoded pixel value and the temporary decoding additional information from the memory 151 is controlled based on the determination result.

  Next, with reference to FIG. 31, a description will be given of the determination principle by which the correctness / incorrectness determination unit 154 in FIG. 30 determines whether the provisional decoded pixel value is correct as a decoding result.

  In the encoding unit 91 (FIG. 19), vector quantization is performed in the RGB space, so that the centroid vector used for the vector quantization is composed of three components: an R component, a B component, and a G component. . Now, the centroid vector composed of the R component, the B component, and the G component is expressed as (R, G, B), and in order to simplify the explanation, the R component and the B component of the centroid vector in the codebook Assuming that the G component and the G component are all expressed by multiples of 10, the R, B, and G components are, for example, pixel values of 102, 103, and 99 (hereinafter, pixel values (102, 103, 99) is vector-quantized into a VQ code corresponding to the centroid vector (100, 100, 100) since the distance from the centroid vector (100, 100, 100) is the shortest. Here, the VQ code corresponding to the centroid vector (100, 100, 100) is set to 0, for example.

  In this case, the VQ residual is (2, 3, -1) by subtracting the centroid vector (100, 100, 100) from the pixel value (102, 103, 99). Accordingly, the pixel value (102, 103, 99) is encoded into VQ code 0 and VQ residual (2, 3, -1).

  Then, when vector inverse quantization is performed on the VQ code 0 and the VQ residual (2, 3, −1) as encoded data obtained by the encoding unit 91 in this way, as shown in FIG. By adding the centroid vector (100, 100, 100) corresponding to the VQ code 0 and the VQ residual (2, 3, -1), the pixel value (102, 103, 99) is obtained. Therefore, the original pixel value is correctly decoded.

  Further, when the decoded pixel values (102, 103, 99) are vector quantized again, as shown in FIG. 31A, the VQ code 0 and the VQ residual (2, 3, −1) are also obtained. ) Is obtained.

  As described above, when the VQ code and the VQ residual as encoded data are decoded, a correct decoding result is obtained, and the decoding result is encoded again (here, vector quantization), the VQ obtained as a result of the encoding is obtained. The code and the VQ residual coincide with the VQ code and the VQ residual as encoded data, respectively.

  On the other hand, VQ residual (2, 3, -1) out of VQ code 0 and VQ residual (2, 3, -1) as encoded data obtained by encoding section 91 is as described above. When the additional information is embedded by performing the line shift, the VQ code 0 and the VQ residual of the pixel corresponding to the VQ residual (2, 3, −1) are the VQ obtained for the other pixels. A residual is assigned. Now, assuming that the VQ residuals of the other pixels are (10, 11, 12) as shown in FIG. 31B, for example, VQ code 0 and VQ residuals (10, 11, 12), by adding the centroid vector (100, 100, 100) corresponding to the VQ code 0 and the VQ residual (10, 11, 12) as shown in FIG. Therefore, the original pixel value (102, 103, 99) is not correctly decoded.

  Therefore, even if the pixel values (110, 111, 112) that are not correctly decoded are subjected to vector quantization again, as shown in FIG. 31B, VQ code 0 or VQ residual (2, 3) as encoded data is obtained. , -1), a VQ code or a VQ residual is not obtained.

  That is, in this case, since the R component, B component, and G component of the centroid vector in the codebook are all expressed by multiples of 10, the distance from the pixel value (110, 111, 112) The shortest centroid vector is (110, 110, 110). Therefore, if the VQ code representing the centroid vector (110, 110, 110) is, for example, 1, the pixel value (110, 111, 112) is the VQ code 1, and the VQ residual (0, 1, 2). (= (110, 111, 112) − (110, 110, 110)). In this case, both the VQ code and the VQ residual are VQ code 0 or VQ which is the original encoded data. Does not match the residual (10, 11, 12).

  From the above, when the number of pixels shifted in accordance with the parameters in the line shift unit 136 in FIG. 28 does not match the additional information, that is, the temporary encoded data is changed to the encoded data before the additional information is embedded. If they do not match, the VQ code and VQ residual obtained by re-encoding the provisional decoded pixel value obtained from such provisionally encoded data become the VQ code and VQ residual as provisionally encoded data. Therefore, it can be determined that the provisional decoded pixel values obtained by decoding the provisionally encoded data are not correct decoding results.

  On the other hand, in the line shift unit 136 of FIG. 28, when the number of pixels shifted according to the parameters matches the additional information, that is, the temporary encoded data is equal to the encoded data before the additional information is embedded. If so, the VQ code and VQ residual obtained by re-encoding the provisional decoded pixel value obtained from such provisionally encoded data become the VQ code and VQ residual as provisionally encoded data, Accordingly, it can be determined that the provisional decoded pixel value obtained by decoding the provisionally encoded data is a correct decoding result.

  Next, referring to the flowchart of FIG. 32, the decoder of FIG. 26 in the case where the encoding rule restoration unit 121, the decoding unit 122, and the determination unit 123 are configured as shown in FIGS. 28 to 30, respectively. The 12 decoding processes will be described.

  In the decoding process, the DEMUX 131 of the encoding rule restoration unit 121 separates the data supplied thereto into a compressed VQ code, VQ residual, and code book, and supplies the compressed data to the decompression unit 132. In step S71, the decompressing unit 132 decompresses the compressed VQ code, VQ residual, and code book from the DEMUX 131, and decompresses the decompressed VQ code, VQ residual, and code book into the VQ code memory 133, VQ. The residual memory 134 and the code book memory 135 are respectively supplied and stored.

  Thereafter, in step S72, the correctness determination unit 154 of the determination unit 123 (FIG. 30) controls the parameter controller 124 (FIG. 26) to set a parameter having a predetermined value, and the parameter is restored to the coding rule. To the unit 121.

  Here, the correctness determination unit 154, for example, for each line of each frame, every time the process of step S72 is performed, an integer value ranging from 0 to the number of pixels of one line is sequentially set as the parameter value. Set.

  When the encoding rule restoration unit 121 receives a parameter from the parameter controller 124, in step S73, the line shift unit 136 shifts the VQ residual of the target line to the left by the number of pixels corresponding to the parameter, and the left The shifted VQ residual is supplied to the decoding unit 122 as temporary encoded data together with the VQ code of the line of interest stored in the VQ code memory 133 and the code book stored in the code book memory 135. Further, the line shift unit 136 supplies the number of pixels shifted from the target line to the determination unit 123 as provisional decoding additional information.

  In step S74, the decoding unit 122 performs vector inverse quantization on the basis of the VQ code, the VQ residual, and the code book as provisional encoded data from the encoding rule restoration unit 121, so that the pixel of the target line The value is decoded, and a provisional decoded pixel value obtained as a result of the decoding is supplied to the determination unit 123.

  In the determination unit 123 (FIG. 30), the provisional decoding additional information output from the encoding rule restoration unit 121 and the provisional decoded pixel value output from the decoding unit 122 are stored in the memory 151. Further, in the determination unit 123, in step S75, the encoding unit 152 vector-quantizes the provisional decoded pixel value from the decoding unit 122, and supplies the resultant provisional VQ code and provisional VQ residual to the comparison unit 153. To do.

  The comparison unit 153 compares the temporary VQ code from the encoding unit 152 with the VQ code as the temporary encoded data for the target line, and the temporary VQ residual from the encoding unit 152 and the temporary encoded data. Are compared with the VQ residuals, and the respective comparison results are supplied to the correctness determination unit 154. In step S76, the correctness determination unit 154 determines that the temporary VQ code matches the VQ code as temporary encoded data for the target line, the temporary VQ residual, and the VQ residual as temporary encoded data. If it is determined whether or not one or both of them match, that is, if the value of the parameter set in the immediately preceding step S72 does not match the additional information, step S72 Returning to, a new value parameter is set, and the same processing is repeated thereafter.

  In step S76, it is determined that the temporary VQ code matches the VQ code as the temporary encoded data, and the temporary VQ residual matches the VQ residual as the temporary encoded data. In this case, that is, the parameter value set in the immediately preceding step S72 matches the additional information. Therefore, the encoded data is restored, and the original additional information is correctly decoded as provisional decoding additional information. In this case, the process proceeds to step S77, and the correctness determination unit 154 controls the memory 151 to decode the provisional decoding pixel value of each pixel of the target line stored therein and provisional decoding additional information as a correct decoding result. The pixel value and the decoding additional information are output, and the process proceeds to step S78.

  In step S78, it is determined whether or not there is a line to be decoded next. When it is determined that there is a line, the process returns to step S72, and the line to be decoded next is newly set as a noticed line. The process is repeated.

  If it is determined in step S78 that there is no line to be decoded next, the decoding process ends.

  Note that the decoding process of FIG. 32 is performed for each frame.

  As described above, the encoding rule for obtaining the encoded data is destroyed based on the additional information, and a predetermined operation is performed on the embedded encoded data encoded by the destroyed encoding rule. Thus, the provisional encoded data is obtained, the provisional encoded data is decoded, and the result obtained by encoding the decoding result is determined to match the provisionally encoded data. Since the encoded data before being destroyed is restored, the destruction and restoration of this encoding rule is used to embed additional information and perform decoding without increasing the data amount of the encoded data. be able to.

  Here, the applicant of the present application has previously proposed a method of embedding additional information in an image using, for example, image correlation. That is, in the method proposed previously, for example, the lines constituting the frame are replaced based on the additional information, and the line of the frame after the replacement is highly correlated with the adjacent lines in the original frame. It is used to return to the original position, but in the method using this correlation, it may be difficult to embed additional information in some cases. That is, the method using the correlation is simply to pay attention to a line with a frame in which additional information is embedded, and replace the line having the highest correlation with the target line next to the target line. However, depending on the image, the line with the highest correlation with the target line may not be the line that should be located next to the target line. If the line is replaced by embedding the additional information, it is difficult to embed the additional information because the correlation cannot be used to return to the original position.

  On the other hand, in the above-described method, there is no such decoding failure.

In the embodiment of FIG. 26, the decoder 12 is configured by providing only one system of the encoding rule restoration unit 122 and the decoding unit 122, and sequentially changing the parameters output from the parameter controller 124, Although provisionally encoded data corresponding to the parameter of each value is sequentially obtained, the decoder 12 is configured such that, for example, as shown in FIG. 33, the M system encoding rule restoring units 121 _{1 to} 121 _M and the decoding unit 122 _{1 to} 122 _M are provided, and different value parameters are given to the encoding rule restoration unit 121 _m (m = 1, 2,..., M) of each system to correspond to the parameters of each value. It is possible to obtain provisionally encoded data at the same time.

  In the above-described case, the encoding unit 91 in FIG. 19 performs vector quantization. However, the encoding unit 91 performs, for example, predictive encoding other than vector quantization. It is possible. In this case, the embedding unit 92 can embed the additional information by operating the prediction residual based on the predictive coding based on the additional information.

  Further, in the above-described case, the VQ residual of each line is shifted based on the additional information to embed the additional information. However, the additional information is, for example, a value representing the VQ residual. Can be embedded by shifting or the like.

  In the above-described case, a code book is created for each frame and included in the encoded data. However, the code book is created in advance, and the embedded compression encoder 11 and decoder 12 respectively. It is possible to memorize it. Furthermore, the code book can be created for each of a plurality of frames or for each predetermined area of one frame.

  Next, the series of processes described above can be performed by hardware or software. When a series of processing is performed by software, a program constituting the software is installed in a general-purpose computer or the like.

  Therefore, FIG. 34 shows a configuration example of an embodiment of a computer in which a program for executing the above-described series of processing is installed.

  The program can be recorded in advance in a hard disk 205 or ROM 203 as a recording medium built in the computer.

  Alternatively, the program is stored in a removable recording medium 211 such as a floppy (registered trademark) disk, a CD-ROM (Compact Disc Read Only Memory), an MO (Magneto optical) disk, a DVD (Digital Versatile Disc), a magnetic disk, or a semiconductor memory. It can be stored (recorded) temporarily or permanently. Such a removable recording medium 211 can be provided as so-called package software.

  The program is installed on the computer from the removable recording medium 211 as described above, or transferred from the download site to the computer wirelessly via a digital satellite broadcasting artificial satellite, LAN (Local Area Network), The program can be transferred to a computer via a network such as the Internet. The computer can receive the program transferred in this way by the communication unit 208 and install it in the built-in hard disk 205.

  The computer includes a CPU (Central Processing Unit) 202. An input / output interface 210 is connected to the CPU 202 via the bus 201, and the CPU 202 operates the input unit 207 including a keyboard, a mouse, a microphone, and the like by the user via the input / output interface 210. When a command is input as a result of this, the program stored in a ROM (Read Only Memory) 203 is executed accordingly. Alternatively, the CPU 202 also receives a program stored in the hard disk 205, a program transferred from a satellite or a network, received by the communication unit 208 and installed in the hard disk 205, or a removable recording medium 211 attached to the drive 209. The program read and installed in the hard disk 205 is loaded into a RAM (Random Access Memory) 204 and executed. Thereby, the CPU 202 performs processing according to the above-described flowchart or processing performed by the configuration of the above-described block diagram. Then, the CPU 202 outputs the processing result from the output unit 206 configured with an LCD (Liquid Crystal Display), a speaker, or the like, for example, via the input / output interface 210, or from the communication unit 208 as necessary. Transmission, and further recording on the hard disk 205 is performed.

  Here, in this specification, the processing steps for describing a program for causing a computer to perform various types of processing do not necessarily have to be processed in time series according to the order described in the flowchart, but in parallel or individually. This includes processing to be executed (for example, parallel processing or processing by an object).

  Further, the program may be processed by a single computer, or may be processed in a distributed manner by a plurality of computers. Furthermore, the program may be transferred to a remote computer and executed.

  In this embodiment, image data is a target for encoding. However, data to be encoded is not limited to image data, and audio data (audio data), a computer program, etc. It is possible to adopt various types of data.

It is a figure which shows the structural example of one Embodiment of the embedding compression / decoding system to which this invention is applied. 2 is a block diagram illustrating a first configuration example of an embedded compression encoder 11. FIG. 3 is a block diagram illustrating a configuration example of a variable length encoding unit 23. FIG. It is a figure for demonstrating variable length coding. It is a figure which shows the relationship between the appearance frequency of a pixel value, and the code length of the code allocated to a pixel value. It is a figure which shows the example of a Huffman table. It is a figure explaining the code | symbol allocated according to the appearance frequency of a pixel value, and conversion of the code | symbol. It is a figure which shows the example of embedded encoding data. It is a figure explaining the method to decompress | restore embedded encoding data to the original encoding data. It is a flowchart explaining an embedded encoding process. It is a flowchart explaining a conversion table creation process. It is a flowchart explaining an encoding data conversion process. 3 is a block diagram illustrating a second configuration example of the embedded compression encoder 11. FIG. 3 is a block diagram illustrating a first configuration example of a decoder 12. FIG. It is a block diagram which shows the structural example of the variable-length decoding parts 52 and 56. FIG. 4 is a block diagram illustrating a configuration example of a Huffman table generation unit 53. FIG. 6 is a block diagram illustrating a configuration example of an inverse conversion table creation unit 54. FIG. It is a flowchart explaining a decoding process. 11 is a block diagram illustrating a third configuration example of the embedded compression encoder 11. FIG. It is a flowchart explaining an embedded encoding process. 3 is a block diagram illustrating a configuration example of an encoding unit 91. FIG. It is a figure explaining the encoding by vector quantization. 3 is a block diagram illustrating a configuration example of an embedding unit 92. FIG. It is a figure explaining embedding of additional information by shifting a line. It is a flowchart explaining an embedded encoding process. 11 is a block diagram illustrating a second configuration example of the decoder 12. FIG. It is a flowchart explaining a decoding process. 3 is a block diagram illustrating a configuration example of an encoding rule restoration unit 121. FIG. 22 is a block diagram illustrating a configuration example of a decoding unit 122. FIG. 3 is a block diagram illustrating a configuration example of a determination unit 123. FIG. It is a figure explaining the process of the determination part. It is a flowchart explaining a decoding process. 11 is a block diagram illustrating a third configuration example of the decoder 12. FIG. It is a block diagram which shows the structural example of one Embodiment of the computer to which this invention is applied.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Encoding apparatus, 2 Decoding apparatus, 3 Recording medium, 4 Transmission medium, 11
Embedded compression encoder, 12 decoder, 21 frame memory, 22 additional information memory, 23 variable length encoding unit, 24 conversion table creation unit, 25 coding rule destruction unit, 26 MUX, 31 frequency table creation unit, 32 Huffman Tree creation part, 33 Huffman table creation part, 34 encoding part,
41 Coding rule destruction unit, 51 DEMUX, 52 Variable length decoding unit, 53 Huffman table generation unit, 54 Inverse conversion table creation unit, 55 Coding rule restoration unit, 56 Variable length decoding unit, 61 Huffman tree creation unit, 62 Huffman Table creation unit, 63 decoding unit, 71 frequency table creation unit, 72 Huffman tree creation unit, 73 Huffman table creation unit, 81 comparison unit, 82 code association unit, 91 coding unit, 92 embedding unit, 101 frame memory, 102 Codebook creation unit, 103 vector quantization unit, 111 VQ code memory, 112 VQ residual memory, 113 codebook memory, 114 line shift unit, 115 compression unit, 116 MUX, 121, 121 _{1 to} 121 _M coding rule restoration parts, 122, 122 ₁ to 122 _M decoder, 1 Third determining unit, 124 parameter controller,
131 DEMUX, 132 decompression unit, 133 VQ code memory, 134 VQ residual memory, 135 codebook memory, 136 line shift unit, 141 vector inverse quantization unit, 142 addition unit, 151 memory,
152 encoding unit, 153 comparison unit, 154 correctness determination unit, 201 bus, 202 CPU, 203 ROM, 204 RAM, 205 hard disk, 206 output unit, 207 input unit, 208 communication unit, 209 drive, 210 I / O interface, 211 Removable recording medium

Claims (7)

Encoding means for quantizing the first data and outputting the resulting quantization code and quantization error as encoded data;
Embedded encoded data in which the second data is embedded by shifting the correspondence between each value of the quantization code as the encoded data and a quantization error by an operation amount corresponding to the second data A data processing apparatus comprising: an embedding means for generating Data processing device
The first data is quantized, and the resulting quantization code and quantization error are output as encoded data,
Embedded encoded data in which the second data is embedded by shifting the correspondence between each value of the quantization code as the encoded data and a quantization error by an operation amount corresponding to the second data A data processing method including generating a step. Encoding means for quantizing the first data and outputting the resulting quantization code and quantization error as encoded data;
Embedded encoded data in which the second data is embedded by shifting the correspondence between each value of the quantization code as the encoded data and a quantization error by an operation amount corresponding to the second data A recording medium on which is recorded a program for causing a computer to function as an embedding means for generating an image. Each value of the quantization code and the quantization error in the embedded encoded data in which the second data is embedded in the encoded data including the quantization code and the quantization error obtained by quantizing the first data Temporary encoded data decoding means for decoding temporary encoded data by shifting the correspondence relationship with a specific operation amount,
Temporary decoded data decoding means for decoding temporary decoded data from the temporary encoded data according to a predetermined code book;
By comparing the quantization code and the quantization error with the quantization error obtained by re-encoding the provisional decoded data decoded by the provisional decoded data decoding unit and the quantization error, Determining means for determining whether or not the encoded data is correct,
The temporary encoded data decoding means changes the manipulated variable according to the determination result by the determination means and decodes the temporary encoded data again,
When the determination means determines that the provisional encoded data is correct, the determination means outputs data corresponding to the manipulated variable as the second data, and the provisional decoded data from the provisional decoded data decoding means is Data processing device that outputs as decrypted data. The data processing apparatus according to claim 4, wherein the first data is image data. Data processing device
Each value of the quantization code and the quantization error in the embedded encoded data in which the second data is embedded in the encoded data including the quantization code and the quantization error obtained by quantizing the first data A temporary encoded data decoding step for decoding the temporary encoded data by shifting the correspondence relationship with a specific operation amount;
Temporary decoded data decoding step for decoding temporary decoded data from the temporary encoded data according to a predetermined codebook;
By comparing the quantization code and the quantization error with the quantization error obtained by re-encoding the temporary decoded data decoded in the temporary decoded data decoding step and the temporary error, A determination step for determining whether or not the encoded data is correct,
In the temporary encoded data decoding step, the operation amount is changed according to the determination result in the determination step and the temporary encoded data is decoded again,
In the determination step, when it is determined that the temporary encoded data is correct, the data corresponding to the manipulated variable is output as the second data, and the temporary decoded data decoded in the temporary decoded data decoding step is true. Data processing method to output as decrypted data. Each value of the quantization code and the quantization error in the embedded encoded data in which the second data is embedded in the encoded data including the quantization code and the quantization error obtained by quantizing the first data Temporary encoded data decoding means for decoding temporary encoded data by shifting the correspondence relationship with a specific operation amount,
Temporary decoded data decoding means for decoding temporary decoded data from the temporary encoded data according to a predetermined code book;
By comparing the quantization code and the quantization error with the quantization error obtained by re-encoding the provisional decoded data decoded by the provisional decoded data decoding unit and the quantization error, Determining means for determining whether or not the encoded data is correct,
The temporary encoded data decoding means changes the manipulated variable according to the determination result by the determination means and decodes the temporary encoded data again,
When the determination unit determines that the temporary encoded data is correct, the determination unit outputs the data corresponding to the operation amount as the second data, and the temporary decoded data from the temporary decoded data decoding unit is A recording medium on which a program for causing a computer to function as a data processing device to output as decoded data is recorded.

Images