JP2004140507A - Data encoder, data decoder, data encoding method, data decoding method, and program and recording medium allowing computer to realize the methods - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a data encoder capable of executing efficient encoding on the basis of a correlation of a plurality of data sequences to solve the difficulty of prior arts in the improvement of the encoding efficiency based on a dictionary. <P>SOLUTION: The image encoder 1000 includes: an input image data memory 1010 for storing high resolution image data and low resolution image data; an encoding pre-processing unit 1020 for classifying pixels of the high resolution image data into a plurality of sub data sequences on the basis of specified cross-referencing between the high resolution image data and the low resolution image data according to position coordinates of the pixels of the image data; an image encoding unit 1050 for applying encoding pixel data by each of the classified sub data groups and applying encoding processing to the low resolution image data; and an encoding data combining unit 1080 for combining encoded data resulting from encoding the data included in the sub data groups with encoded data resulting from encoding data included in a second data group. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a technique for processing image data, and more particularly, to a data encoding apparatus for compressing and encoding image data, a data encoding method, a data decoding apparatus for decompressing and decoding, and a data decoding method.
[0002]
[Prior art]
As a data encoding process, an encoding process based on a dictionary is widely used. This is because the encoding process determines whether each subsequence of data to be encoded matches any of the subsequences that have been processed by the encoder, and provides information specifying such a pattern. It encodes itself.
[0003]
If the codeword obtained in this way is shorter than the original subsequence, the effect of compressing data can be obtained. The lossless encoding process using such a principle is widely used because a certain compression effect can be obtained regardless of the type of data. For example, a Ziv-Lempel code (LZ77 code), which is one of the encodings based on such a dictionary, is used as an encoding algorithm of a compression tool gzip. Regardless of whether the data to be compressed is data representing an image or data representing a text, such a tool has a low computational load in the decryption process, and is generally used for a general-purpose reversible compression process. Widely used for.
[0004]
Such a dictionary-based encoding process is excellent from the viewpoint of versatility, but generally has a relevance to the encoding process of increasing the compression efficiency by utilizing the correlation of different data sequences. Low.
[0005]
An image encoding apparatus related to the present invention that executes encoding processing based on a dictionary, which enhances compression efficiency using correlation between different data sequences, will be described.
[0006]
FIG. 39 shows a block diagram of the image encoding device 100. The image encoding apparatus 100 includes an input image data memory 102 for storing image data, a high-resolution image for encoding at the resolution of the input image before encoding the image, and a low-resolution image having half the resolution. A high-resolution image data memory 106 for storing the generated high-resolution image data; a low-resolution image data memory 108 for storing the generated low-resolution image data; A data encoder 110 for encoding image data representing a high-resolution image and image data representing a low-resolution image, and a high-resolution code for storing encoded image data representing the encoded high-resolution image Encoded image data memory 112, and a low-resolution encoded image data for storing encoded image data representing an encoded low-resolution image. A memory 114; a coded data combiner 116 for combining coded image data representing the coded high-resolution image with coded image data representing the coded low-resolution image; A combined coded image data memory 118 for storing the combined coded image data.
[0007]
FIG. 40 is a flowchart showing a procedure of processing executed by image encoding apparatus 100.
[0008]
In step (hereinafter, step is referred to as S) 100, encoding preprocessor 104 converts input image 300 stored in input image data memory 102 into low-resolution image 304 and high-resolution image 302. The data is separated and stored in the low-resolution image data memory 108 and the high-resolution image data memory 106, respectively.
[0009]
FIG. 41 shows a low-resolution image 304 and a high-resolution image 302 processed by the image encoding device 100. In the image shown in FIG. 41, the input image 300 is represented by four pixels both vertically and horizontally for the sake of simplicity. However, this is for simplification of the description, and the size of the image to be coded is usually larger than this.
[0010]
In FIG. 41, each pixel of the input image 300, the high-resolution image 302, and the low-resolution image 304 is distinguished by a given number. As is evident from the numbers attached to the pixels, the high resolution image 302 is a copy of the input image 300 and the low resolution image 304 is created by １ ／ downsampling the input image 300.
[0011]
In S200, data encoder 110 encodes high-resolution image 302 stored in high-resolution image data memory 106 and low-resolution image 304 stored in low-resolution image data memory 108, and performs high-resolution encoding, respectively. The image data is stored in the image data memory 112 and the low-resolution coded image data memory 114. The order in which the pixels of the high-resolution image 302 and the low-resolution image 304 are encoded by the data encoder 110 is a normal raster scan order as shown in FIG.
[0012]
It is assumed that the data encoder 110 uses LZ77 encoding for encoding. This LZ77 encoding has the following operation principle.
(1) A symbol string of a fixed length (reference symbol string) extracted from the most recently coded one and a symbol string from the head of the uncoded symbol string (unreferenced coded string) are subjected to the longest matching method. To compare.
(2) Output information including the following as a codeword.
The number of the symbol in the reference symbol string where the longest match starts
・ How many symbols does the above match contain?
・ What is the first symbol that does not match in the unreferenced coded sequence?
(3) The length (the number of symbols) of a symbol string to be encoded at one time generally differs each time, but only the encoded symbol string is rewritten in such a way that the old reference symbol string is pushed out. Can be The head of the unencoded symbol string advances by the number of encoded symbol strings.
(4) The shorter the same pattern appears in the encoding target symbol string with higher frequency, the higher the encoding efficiency.
Note that the premise that "short identical patterns appear with high frequency" described in (4) above is the same as "high correlation between adjacent symbols included in the encoding target symbol string". It should be noted. This is because, for example, in encoding target data consisting of repetition of the same symbol string composed of two symbols, when a certain symbol is extracted, the next symbol is also determined, that is, the correlation coefficient becomes “1”. I understand. Also, it should be noted that the property of (4) is not limited to LZ77, but is a property generally applied to lossless encoding based on a dictionary.
[0013]
Since there are many literatures on LZ77 encoding, a more detailed description will not be repeated here.
[0014]
In S300, coded data combiner 116 combines the contents of low-resolution coded image data memory 114 and high-resolution coded image data memory 112 and stores them in combined coded image data memory 118.
[0015]
According to such an image encoding device 100, the high-resolution image 302 and the low-resolution image 304 are different data sequences having a clear correlation. In the coding of both, coding is efficiently performed using the correlation between the two. Furthermore, in decoding the low-resolution image 304, it is not necessary to decode the high-resolution image 302. The image encoding apparatus 100 is the simplest method of encoding such that the high-resolution image 302 and the low-resolution image 304 can be selectively obtained at the time of decoding. The pixels included in 304 are partially overlapped with each other, such as a pixel 310 shown in FIG. For this reason, the coding efficiency is deteriorated.
[0016]
Further, there is an image encoding apparatus improved to eliminate the overlapping pixels in order to improve the encoding efficiency.
[0017]
The block diagram of the improved image encoding device is the same as the block diagram shown in FIG. 39, and the flowchart of the improved image encoding device is the same as the flowchart shown in FIG.
[0018]
FIG. 43 shows a high-resolution image 502 and a low-resolution image 504 in the improved image encoding device. The input image 500 and the low-resolution image 504 are exactly the same as the input image 300 and the low-resolution image 304, respectively. However, the high-resolution image 502 is configured not to include the pixels included in the low-resolution image 504. If the high-resolution image 502 and the low-resolution image 504 are configured as described above, there is no pixel to be redundantly encoded, so that the encoding efficiency is improved.
[0019]
FIG. 44 shows an order of encoding the high-resolution image 502 in the data encoder 110 of the improved image encoding device. As described above, when encoding the high-resolution image 502 in the data encoder 110, there are the following problems.
[0020]
For encoding in the order shown in FIG. When the pixel 508 shown in FIG. 43 is encoded, the immediately preceding pixel is not the pixel adjacent to the left, but is another pixel 506 adjacent to the left. On the other hand, when encoding the pixel 308 of the high-resolution image 302 in FIG. 41, the pixel encoded immediately before is the pixel 306 adjacent to the left. This is because, in the high-resolution image 502 of FIG. 43, the pixel on the left of the pixel 508 is included in the low-resolution image 504 as the pixel 510. It is because it does not enter.
[0021]
As a characteristic of image data, it is known that the closer a pixel is, the higher the correlation is. However, in the improved image encoding device, when encoding the high-resolution image 502, some pixels have the highest correlation. Indicates that other pixels cannot be referenced. For this reason, it is difficult for the improved image encoding device to increase the encoding efficiency. This corresponds to general encoding processing based on a dictionary from the operating principle of LZ77.
[0022]
Even in such a case, of course, when the low-resolution image 504 and the high-resolution image 502 are sequentially encoded by the data encoder 110 and the processing is transferred from the low-resolution image 504 to the high-resolution image 502, This problem can be avoided by making the reference symbol string extremely long without initializing the reference symbol string. However, information indicating which substring the encoding target character string matches in the reference symbol string becomes long, which is not preferable in terms of encoding efficiency.
[0023]
Such two image encoding apparatuses are capable of encoding different data sequences having a correlation with each other by making use of the correlation by using an encoding method without an explicitly given probability model (conventional encoding method). (Method represented by an encoding method based on a dictionary).
[0024]
In connection with such a problem, Japanese Patent Laying-Open No. 7-336237 (Patent Document 1) discloses a method for compressing data information. The compression method disclosed in this publication is based on a Lempel-Ziv (LZ) compression process that creates a dictionary of phrases using preceding symbols and compresses the data stream using the dictionary. , The d context or d history h of a phrase is defined as the set of d bits immediately preceding it, as a function of the history associated with each of the phrases to be encoded, as a function of the dictionaries of the phrase to be encoded. Expanding and storing; comparing the compression efficiencies of the plurality of dictionaries with respect to previously encoded phrases to select a dictionary determined to provide the statistically highest potential compression; and the selected dictionary Encoding the next phrase using a particular history corresponding to the history of the next phrase.
[0025]
According to the compression method disclosed in this publication, by using different reference character strings depending on the encoding history at each time point, it is possible to achieve substantially the same effect by performing more efficient encoding. To do. This compression method is applied to encoding having a plurality of probability models by extending the range to general entropy encoding. This compression method is a method of performing encoding or decoding processing while selecting an explicitly given probability model based on encoded (at encoding) and decoded (at decoding) data.
[0026]
[Patent Document 1]
JP-A-7-336237
[0027]
[Problems to be solved by the invention]
However, the compression method disclosed in this publication aims to improve the efficiency of encoding a single data sequence. The above-described image encoding device and the improved image encoding device improve the efficiency of encoding using the correlation between different data sequences. For this reason, the compression method disclosed in this publication cannot be applied to the above-described image encoding device or the improved image encoding device. Therefore, even when these techniques are combined, a method of using the correlation of different data series in the encoding process while having the characteristic that the decoding process can be performed at high speed, especially like encoding based on a dictionary. Cannot be realized.
[0028]
The present invention has been made to solve the above-described problem, and performs efficient encoding based on a correlation of a plurality of data sequences, which has been difficult with conventional dictionary-based encoding. To provide a data encoding device capable of performing the above-described data encoding, and a data decoding method corresponding to the data encoding method. And a program and a recording medium for causing a computer to realize the methods.
[0029]
[Means for Solving the Problems]
A data encoding device according to a first invention stores a first data group including a plurality of data and a second data group including a plurality of data not included in the first data group. Storage means, determination means for determining the correspondence between the data included in the first data group and the data included in the second data group, and the first means based on the determined correspondence. Classification means for classifying a data group into a plurality of sub data groups; and first encoding means for encoding data included in the sub data group for each sub data group classified by the classification means And second encoding means for encoding data included in the second data group, encoded data obtained by encoding data included in the sub data group, and encoded data included in the second data group. Data and the encoded data And a coupling means for coupling to.
[0030]
According to the first invention, the association determined by the determining means is based on a correlation established between data included in the first data group and data included in the second data group. It is. The classification means classifies the first data group into a plurality of sub-data groups composed of data having a strong correlation based on such association. At this time, the data included in the classified one sub-data group has a strong correlation such that their values are close to each other. The first encoding means encodes data included in the sub data group for each classified sub data group, and the second encoding means encodes data included in the second data group. To process. The combining unit combines the encoded data obtained by encoding the data included in the sub-data group and the encoded data obtained by encoding the data included in the second data group. Thus, the first encoding unit performs the encoding process for each sub-data group composed of data having a strong correlation with each other, so that the entropy of the data included in this sub-data group can be reduced. As a result, coding efficiency can be improved by coding for each classification, and efficient coding based on the correlation of multiple data series, which was difficult with conventional dictionary-based coding. Can be provided.
[0031]
In a data encoding apparatus according to a second aspect of the present invention, in addition to the configuration of the first aspect, the determining means determines that data included in the first data group and data included in the second data group are data Means for determining the association so as to have a predetermined relationship with respect to the value of.
[0032]
According to the second invention, for example, in the case of still image data, the position value of a pixel is close, and in the case of moving image data, it is an adjacent frame. The association can be determined so as to have a relationship such as performing the association.
[0033]
In a data encoding device according to a third aspect of the present invention, in addition to the configuration of the second aspect of the invention, the deciding means includes a position in the image of the pixel data included in the first data group, and a position included in the second data group. Means for determining the association so as to have a relationship based on the position of the pixel data in the image.
[0034]
According to the third aspect, the data to be subjected to the encoding processing is data representing an image, and the data includes pixel data corresponding to a plurality of pixels constituting the image. The determining means determines the association based on the positions of the pixels in the image so that the images have a strong correlation with each other. In this way, since the data whose pixel positions are close to each other have a strong correlation with each other (the values of the pixel data are close to each other), the sub-data groups composed of these data are collectively subjected to the encoding process. Efficient encoding can be performed.
[0035]
In a data encoding device according to a fourth aspect of the present invention, in addition to the configuration of the third aspect, the determining means includes means for determining the association by a function using a position in an image.
[0036]
According to the fourth aspect, the correspondence between the data included in the first data group and the data included in the second data group is determined by a function using the XY coordinate values that are the positions of the pixels in the image. Can be.
[0037]
The data encoding apparatus according to a fifth aspect of the present invention further includes a dividing unit for dividing an image into a plurality of types of resolutions, in addition to the configuration of the third or fourth aspect. The first group of data includes data for high-resolution images, and the second group of data includes data for low-resolution images.
[0038]
According to the fifth aspect, for example, the dividing means creates high-resolution image data and low-resolution image data, and converts the high-resolution data into low-resolution pixels corresponding to the positions of the high-resolution pixels. Map to data. Thereby, a plurality of high-resolution data corresponding to one low-resolution data can be included in one sub-data group.
[0039]
The data encoding apparatus according to a sixth aspect of the present invention further includes, in addition to the configuration of the second aspect, division means for dividing a moving image into a plurality of frame images in a time series. The determining means has a relationship based on the data on the image of the specific frame included in the first data group and the data on the image earlier than the specific frame included in the second data group. Thus, means for determining the association is included.
[0040]
According to the sixth aspect, when encoding data representing a moving image, data at a certain position of a specific frame and data at the same position of a frame earlier than the specific frame are mutually different. Has a strong correlation. If these data are included in one sub-data group and subjected to the encoding process, the entropy of the data included in the sub-data group is reduced, and efficient encoding can be performed.
[0041]
In a data encoding apparatus according to a seventh aspect of the present invention, in addition to the configuration of any one of the first to sixth aspects, the classifying unit may be configured to convert the first data group into a sub-data group composed of data strongly related to each other. Including means for classifying.
[0042]
According to the seventh aspect, the sub-data group is composed of data having a strong relationship with each other, and the encoding process is performed for each of the sub-data groups. Therefore, the entropy of the data included in the sub-data group is reduced, and an efficient code is obtained. Can be implemented.
[0043]
In a data encoding apparatus according to an eighth aspect of the present invention, in addition to the configuration of any one of the first to sixth aspects, the classifying unit is configured to classify the first data group from data having mutually similar data values. Means for classifying the data into sub data groups.
[0044]
According to the eighth aspect, the sub-data group is composed of data whose data values are close to each other, and the encoding process is performed for each sub-data group. Therefore, the entropy of the data included in the sub-data group is reduced, and the efficiency is reduced. Encoding can be performed.
[0045]
According to a ninth aspect of the present invention, in addition to the configuration of any one of the first to eighth aspects, the data encoding apparatus further comprises a type of encoding process executed by the first encoding unit and a second encoding unit. Is a device that is the same type of encoding process as that performed by the device.
[0046]
According to the ninth aspect, the encoding process executed by the first encoding unit and the encoding process executed by the second encoding unit are the same type of encoding process. It is possible to provide a data encoding device with a small circuit scale by combining encoders that realize the means into one.
[0047]
A data encoding apparatus according to a tenth aspect of the present invention executes encoding processing based on a dictionary in addition to the configuration of the ninth aspect.
[0048]
According to the tenth aspect, efficient encoding can be executed in encoding processing based on a dictionary, which has been difficult in the past.
[0049]
The data decoding device according to an eleventh aspect includes a storage unit for storing the combined first encoded data group and the second encoded data group, and a data group stored in the storage unit. A dividing unit for dividing the data into a first encoded data group and a second encoded data group; and a sub data group by decoding encoded data included in the first encoded data group. A first decoding unit, a second decoding unit for decoding encoded data included in the second encoded data group to create a second data group, a sub data group and a second Creating means for creating a first data group based on the data group and according to the association.
[0050]
According to the eleventh aspect, the original data decoded by the data decoding device includes a first data group including a plurality of data and a second data group including a plurality of data not included in the first data group. And data groups. The sub-data group is a data group in which the first data group is classified based on the correspondence between the data included in the first data group and the data included in the second data group. The sub-data group is encoded into a first encoded data group for each sub-data group. This association is, for example, an association based on a correlation established between the data included in the first data group and the data included in the second data group. The first decoding means decodes the coded data included in the first coded data group divided by the dividing means to create a sub-data group, and the second decoding means generates a sub-data group. The encoded data included in the data group is decoded to create a second data group. The creating unit creates a first data group using the decoded sub-data group and the second data group, based on the association at the time of encoding. Thus, the original data can be created by decoding the encoded data. In the encoding process, data included in one classified sub-data group has a strong correlation such that their values are close to each other. Since the data included in the sub data group is subjected to the encoding process for each of the classified sub data groups, the entropy of the data included in the sub data group can be reduced, and the encoding efficiency can be improved. As a result, it is possible to provide a data decoding device corresponding to a data encoding device capable of executing efficient encoding based on the correlation of a plurality of data sequences, which has been difficult with conventional dictionary-based encoding. it can.
[0051]
In a data decoding device according to a twelfth aspect, in addition to the configuration of the eleventh aspect, the creating means includes a step of determining whether the data included in the first data group and the data included in the second data group correspond to each other. Means for creating the first data group so as to have a predetermined relationship with the value is included.
[0052]
According to the twelfth aspect, for example, in the case of still image data, since the position data of the pixel is close, and in the case of moving image data, it is an adjacent frame, so that the data values are approximate to each other. The first data group can be created so as to have a relationship such as the first data group.
[0053]
In a data decoding device according to a thirteenth aspect of the present invention, in addition to the configuration of the twelfth aspect, the creating means includes a position in the image of pixel data included in the first data group, and a position included in the second data group. Means for creating the first data group so as to have a relationship based on the position of the pixel data in the image is included.
[0054]
According to the thirteenth aspect, the original data to be decoded is data representing an image, and the data includes pixel data corresponding to a plurality of pixels forming the image. The creating unit creates the first data group based on the positions of the pixels in the image so as to have a strong correlation with each other. In this manner, in the encoding process, data having close pixel positions have a strong correlation with each other (pixel data values are close to each other). It is possible to provide a data decoding device that can decode the encoded data encoded in (1).
[0055]
In a data decoding device according to a fourteenth aspect, in addition to the configuration of the thirteenth aspect, the creating means includes means for creating a first data group by a function using a position in an image.
[0056]
According to a fourteenth aspect, there is provided a data decoding device capable of decoding encoded data that has been classified and encoded based on a correspondence determined by a function using XY coordinate values that are pixel positions in an image. Can be.
[0057]
A data decoding device according to a fifteenth aspect of the present invention further includes a designation unit for designating a resolution, in addition to the configuration of the thirteenth aspect.
[0058]
According to the fifteenth aspect, the original data includes data representing an image at a plurality of resolutions, and the specifying means can specify up to which resolution the image data is to be decoded.
[0059]
In a data decoding device according to a sixteenth aspect of the present invention, in addition to the configuration of the twelfth aspect, the creating means includes: data on an image of a specific frame included in the first data group; Means for generating a first group of data based on data for an image earlier than a particular frame to be generated.
[0060]
According to the sixteenth aspect, when decoding processing is performed on original data representing a moving image, data at a certain position of a specific frame and data at the same position of a frame earlier than the specific frame are mutually different. Has a strong correlation. It is possible to provide a data decoding device capable of decoding encoded data that has been efficiently encoded by including these data in one sub-data group.
[0061]
A data decoding device according to a seventeenth aspect of the present invention provides, in addition to the configuration of any one of the eleventh to sixteenth aspects, a type of decoding processing executed by the first decoding means and a type of decoding processing executed by the second decoding means The type of the decryption process is an apparatus that is the same process.
[0062]
According to the seventeenth aspect, the decoding process executed by the first decoding unit and the decoding process executed by the second decoding unit are the same type of decoding process. Can be combined into one to provide a data decoding device with a small circuit scale.
[0063]
A data decoding device according to an eighteenth aspect of the present invention, in addition to the configuration of the seventeenth aspect, performs the decoding process based on a dictionary.
[0064]
According to the eighteenth aspect, it is possible to provide a data decoding device capable of efficiently decoding encoded data based on a dictionary, which has been difficult in the past.
[0065]
In a data encoding method according to a nineteenth aspect, a first data group including a plurality of data and a second data group including a plurality of data not included in the first data group are prepared in advance. A preparing step, a determining step of determining a correspondence between the data included in the first data group and the data included in the second data group, and a step of determining the first data group based on the determined correspondence. A classification step of classifying the data into a plurality of sub data groups, a first encoding step of encoding data included in the sub data group for each sub data group classified in the classification step, and a second data A second encoding step of encoding data included in the group, encoded data in which data included in the sub data group is encoded, and a code in which data included in the second data group are encoded. Data And a combining step of engagement.
[0066]
According to the nineteenth aspect, the association determined in the determining step is an association based on a correlation established between data included in the first data group and data included in the second data group. And so on. In the classification step, the first data group is classified into a plurality of sub-data groups composed of data having a strong correlation based on such association. At this time, the data included in the classified one sub-data group has a strong correlation such that their values are close to each other. In a first encoding step, data included in the sub data group is encoded for each classified sub data group, and in a second encoding step, data included in the second data group is encoded. Perform encoding processing. In the combining step, the coded data obtained by coding the data included in the sub-data group and the coded data obtained by coding the data included in the second data group are combined. Thus, in the first encoding step, the encoding process is performed for each sub-data group composed of data having a strong correlation with each other, so that the entropy of data included in this sub-data group can be reduced. As a result, coding efficiency can be improved by coding for each classification, and efficient coding based on the correlation of multiple data series, which was difficult with conventional dictionary-based coding. Can be provided.
[0067]
A data decoding method according to a twentieth aspect includes: a preparing step of preparing a combined first coded data group and a second coded data group in advance; and a data group prepared in the preparing step. A dividing step of dividing the data into a first encoded data group and a second encoded data group, and a first step of decoding encoded data included in the first encoded data group to create a sub data group. A decoding step, a second decoding step of decoding encoded data included in the second encoded data group to create a second data group, and a sub data group and a second data group. Creating a first data group in accordance with the association.
[0068]
According to the twentieth aspect, the original data decoded by the data decoding method includes a first data group including a plurality of data and a second data group including a plurality of data not included in the first data group. And data groups. The sub-data group is a data group in which the first data group is classified based on the correspondence between the data included in the first data group and the data included in the second data group. The sub-data group is encoded into a first encoded data group for each sub-data group. This association is, for example, an association based on a correlation established between the data included in the first data group and the data included in the second data group. In a first decoding step, encoded data included in the first encoded data group, which has been divided in the dividing step, is decoded to create a sub data group, and in a second decoding step, a second data group is generated. The encoded data included in the encoded data group is decoded to create a second data group. In the creating step, the first data group is created based on such association using the decoded sub data group and the second data group. Thus, the original data can be created by decoding the encoded data. In the encoding process, data included in one classified sub-data group has a strong correlation such that their values are close to each other. Since the data included in the sub data group is subjected to the encoding process for each of the classified sub data groups, the entropy of the data included in the sub data group can be reduced, and the encoding efficiency can be improved. As a result, it is possible to provide a data decoding method corresponding to a data encoding method that can perform efficient encoding based on the correlation of a plurality of data sequences, which has been difficult with conventional dictionary-based encoding. it can.
[0069]
A program according to a twenty-first invention is a program for causing a computer to realize at least one of the data encoding method according to the nineteenth invention and the data decoding method according to the twentieth invention.
[0070]
According to the twenty-first aspect, there is provided a data encoding method and a data encoding method capable of performing efficient encoding based on a correlation of a plurality of data sequences, which have been difficult with conventional dictionary-based encoding. A program for causing a computer to execute any of the corresponding data decoding methods can be provided.
[0071]
A recording medium according to a twenty-second invention is a computer-readable recording medium on which the program according to the twenty-first invention is recorded.
[0072]
According to the twenty-second invention, a data encoding method and a data encoding method capable of performing efficient encoding based on a correlation of a plurality of data sequences, which are difficult with conventional dictionary-based encoding, It is possible to provide a recording medium on which a program for causing a computer to implement any of the corresponding data decoding methods is recorded.
[0073]
A data encoding device according to a twenty-third aspect includes a storage unit for storing a plurality of data groups ordered in advance, and a storage unit for storing a plurality of ordered data groups from each data included in one data group. Determining means for determining a correspondence relation to data included in the data group in the order, and classifying one data group into a sub-data group based on the corresponding data in the next data group based on the correspondence relation Classifying means, first coding means for coding each of the sub-data groups classified based on the classifying means, and a last-order data group among the plurality of ordered data groups And a second encoding means for encoding processing.
[0074]
According to the twenty-third aspect, it is possible to recursively classify data into sub-data groups and code them while referring to the next data group (when applied to an image, the resolution is one step coarser). it can.
[0075]
A data decoding device according to a twenty-fourth aspect includes a storage unit configured to store a combined encoded data group and a storage unit configured to divide the combined encoded data group into a plurality of ordered encoded data groups. Division means, first decoding means for decoding the last-ordered coded data group among the plurality of ordered coded data, and ordered coded data group other than the last-ordered coded data group And a second decoding means for decoding the data to generate a sub-data group, and a sub-data group included in a next data group from each data included in one data group included in the plurality of ordered encoded data. Determining means for determining the correspondence to the data to be processed, and integrating means for integrating the sub-data group into the data group based on the corresponding data of the next-order data group based on the correspondence relationship. Including.
[0076]
According to the twenty-fourth aspect, a code obtained by recursively classifying data into sub-data groups while referring to the next data group (when applied to an image, the resolution is one step coarser) is referred to. Decrypted data.
[0077]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the following description, the same components are denoted by the same reference numerals. Their names and functions are the same. Therefore, detailed description thereof will not be repeated.
[0078]
<First embodiment>
Hereinafter, an image encoding device according to the first embodiment of the present invention will be described.
[0079]
FIG. 1 shows a block diagram of image coding apparatus 1000 according to the present embodiment. As shown in FIG. 1, the image encoding apparatus 1000 includes an input image data memory 1010 for storing image data, and a high-resolution image for encoding at the resolution of the input image before encoding the image. , A pre-encoding processor 1020 for generating a low-resolution image having half the resolution thereof, a high-resolution image data memory 1030 for storing the generated high-resolution image data, and a low-resolution image memory for storing the generated low-resolution image data. A resolution image data memory 1040, an image encoder 1050 that encodes image data representing a high-resolution image and image data representing a low-resolution image, and an encoded image data representing an encoded high-resolution image. A high-resolution coded image data memory 1060 for storing, and coded image data representing a coded low-resolution image are stored. Data combining for coding low-resolution coded image data memory 1070 and coded image data representing a coded high-resolution image and coded image data representing a low-resolution coded image And a combined coded image data memory 1090 for storing the combined coded image data subjected to the combining process.
[0080]
In the image encoding device 1000, an input image data memory 1010, a high-resolution image data memory 1030, a low-resolution image data memory 1040, a high-resolution encoded image data memory 1060, and a low-resolution encoded image data memory 1070 The combined coded image data memory 1090 is generally realized by a storage device implemented in a computing system including a computer.
[0081]
In the image encoding apparatus 1000, the encoding pre-processor 1020, the image encoder 1050, and the encoded data combiner 1080 may be realized by software executed by a computer or by hardware. Is also possible. In the following description, the encoding preprocessor 1020, the image encoder 1050, and the encoded data combiner 1080 have specific hardware including a controller, and the controller executes predetermined software. The description will be made assuming that the hardware is controlled by executing the function to realize a predetermined function. However, instead of specific hardware, a general computer may be used to realize the predetermined function. Therefore, one essential aspect of the present invention is a program for causing a computer to realize such a function.
[0082]
The input image data stored in the input image data memory 1010, that is, the data that is encoded by the image encoding device 1000 according to the present embodiment will be described in detail. It is assumed that the input image data is a grayscale image of 256 gradations. Note that the image coding apparatus 1000 according to the present embodiment can be applied to the case of image data having a different number of gradations by appropriately changing the value of the corresponding parameter. Furthermore, since a color image is separated into YUV, RGB, and the like, resulting in a plurality of grayscale images, the image encoding device 1000 according to the present embodiment is applied only to grayscale images and monochrome images. It is not something to be done. That is, when applied to a color image, the process may be repeated as many times as necessary for the number of color components (usually 3 or 4). In the following description, the length of the register in each memory is assumed to be a necessary and sufficient length to represent a possible value, and will not be particularly described unless it is related to the format of the encoded image data. . The details of the structure of the image encoder 1050 will be described later.
[0083]
With reference to FIG. 2, the procedure of the image data encoding process of image encoding apparatus 1000 according to the present embodiment will be described.
[0084]
In S1000, image encoding apparatus 1000 creates high-resolution image data and low-resolution image data based on input image data stored in input image data memory 1010 using encoding preprocessor 1020. Then, they are stored in the high-resolution image data memory 1030 and the low-resolution image data memory 1040, respectively. The encoding preprocessor 1020 creates a high resolution image 1210 and a low resolution image 1220 based on the input image 1200, as shown in FIG. Here, the input image 1200 is shown as an image having 4 pixels horizontally and 4 pixels vertically, but this is for the purpose of simplifying the description. In practice, the size of the image to be coded is usually larger than this. In FIG. 3, each pixel of the input image 1200, the high-resolution image 1210, and the low-resolution image 1220 is distinguished by a given number. As is evident from the numbers attached to the pixels, the high-resolution image 1210 is a copy of the input image 1200 excluding the pixels included in the low-resolution image 1220, and the low-resolution image 1220 is one of the input images 1200. / 2 downsampling.
[0085]
Further, the encoding preprocessor 1020 may create a high-resolution image 1310 and a low-resolution image 1320 based on the input image 1300 as shown in FIG. As can be seen from FIGS. 3 and 4, each pixel of the high resolution image 1310 is the same as each pixel of the high resolution image 1210, but the value of each pixel of the low resolution image 1320 shown in FIG. Is given by the average of the values of the four pixels.
[0086]
3 and 4, when the high resolution image 1210 (or the high resolution image 1310) and the low resolution image 1220 (or the low resolution image 1320) are decoded, the input image 1200 (or the input image 1300) is given enough information to completely decode it. That is, in the case of FIG. 4, each pixel value of the low-resolution image 1320 is an average value of the corresponding four pixels of the input image 1300. Therefore, if the values of the three pixels of the high-resolution image 1310 are decoded, The value of one pixel is obtained by subtracting the sum of the values of the three corresponding pixels of the high-resolution image 1310 from the value obtained by quadrupling the corresponding pixel of the low-resolution image 1320 (to the extent that bit precision allows). ), The value of the remaining one pixel of the high resolution image 1310 can be obtained. In the case of FIG. 3, it is clear that if the information of the high-resolution image 1210 and the information of the low-resolution image 1220 are added, sufficient information for restoring the input image 1200 can be obtained. Of course, if complete lossless encoding is required, the method of FIG. 4 cannot be generally adopted due to bit precision issues.
[0087]
In S1100, image encoding apparatus 1000 encodes high-resolution image data and low-resolution image data using a predetermined compression method using image encoder 1050, and outputs the encoded data. Is stored in the high-resolution coded image data memory 1060 and the low-resolution coded image data memory 1070. Details of the processing procedure in the image encoder 1050 will be described later. The order in which the high-resolution image 1210 shown in FIG. 3 and the high-resolution image 1310 shown in FIG. 4 are encoded by the image encoder 1050 is a normal raster scan order as shown in FIG.
[0088]
At S1200, image coding apparatus 1000 uses coded data combiner 1080 to combine the data stored in high-resolution coded image data memory 1060 and low-resolution coded image data memory 1070, and combine them. It is stored in the encoded image data memory 1090.
[0089]
The combined coded image data memory 1090 combines the contents of the high-resolution coded image data memory 1060 shown in FIG. 6 and the contents of the low-resolution coded image data memory 1070 shown in FIG. Is stored. Note that the image coding apparatus 1000 according to the present embodiment causes the combined coded image data memory 1090 to store information that is transmitted to the decoding apparatus and that is necessary for decoding processing such as the vertical and horizontal sizes of an image. It does not exclude that.
[0090]
Hereinafter, details of the image encoder 1050 will be described.
FIG. 9 shows a block diagram of the image encoder 1050. Note that the image encoder 1050 of the image encoding device 1000 according to the present embodiment executes the LZ77 encoding process, but the image encoding device 1000 according to the present invention is limited to the LZ77 encoding process. It does not apply. The image encoding device 1000 according to the present invention can be applied to general encoding processing based on a dictionary.
[0091]
The image encoder 1050 includes a symbol string decomposer 1420, a symbol string buffer 1430, a symbol string counter 1440, a data encoder 1450, and a controller 1410 that controls these.
[0092]
The controller 1410 controls the entire image encoder 1050. The controller 1410 includes a CPU (Central Processing Unit), not shown, and a memory for storing a program executed by the CPU.
[0093]
Hereinafter, the Ith and Jth elements of the symbol string buffer 1430 are represented by S [I] [J], and the Ith value of the symbol string counter 1440 is represented by SC [I].
[0094]
The symbol string decomposer 1420 decomposes the contents of the high-resolution image data stored in the high-resolution image data memory 1030 and stores it in the symbol string buffer 1430. The symbol string buffer 1430 is a buffer having a two-dimensional array in which the first dimension is 3 and the second dimension is equal to the length of the high-resolution image data memory 1030. That is, the symbol string buffer 1430 has a two-dimensional array having a size of 3 × N. The “first dimension is 3” represents 3 × N “3”, and the “length of the high resolution image data memory 1030” is the number of words (= stored in the high resolution data memory 1030). The maximum number of pixels obtained).
[0095]
The symbol string buffer 1430 is a two-dimensional memory for classifying and storing each pixel included in the high-resolution data memory 1030 according to the value of a key calculated for each pixel (keys will be described later). Is an array. In the image encoding apparatus 1000 according to the present embodiment, the key value K is set to three values of 0 to 2, so that the symbol string buffer 1430 includes three columns. Therefore, the first dimension becomes “3”.
[0096]
The length of the second dimension is high-resolution image data so that data can be stored even if the required length is the maximum length (when the same key register 1428 value is calculated from all pixels). The length is set to the length of the memory 1030 (= the maximum number of pixels that can be stored). However, such a configuration of the symbol string buffer 1430 is an example, and the image encoding device according to the present invention is not limited to such a buffer configuration.
[0097]
The symbol string counter 1440 is a one-dimensional array buffer having three elements. These three elements are SC [0], SC [1], and SC [2]. The SC [I] (I = 0, 1, 2) itself does not indicate the value of the image data, but indicates the number of pixels included in each classification when each pixel is classified according to the key value. Things. SC [0], SC [1], and SC [2] indicate the location of the symbol string buffer 1430 where the next pixel is stored in the flowchart (FIG. 12) described later. As each pixel is stored in S [K] and [SC [K]] according to the value of the key, SC [K] is incremented according to the value of K. After the processing shown in the flowchart described later is completed, SC [0], SC [1], and SC [2] indicate the number of pixels included in each classification.
[0098]
The details of the structure of the symbol string decomposer 1420 for classifying and storing symbol strings in the symbol string buffer 1430 as described above will be described later.
[0099]
The data encoder 1450 encodes the three symbol strings stored in the symbol string buffer 1430 and the image data stored in the low-resolution image data memory 1040.
[0100]
With reference to FIG. 10, the procedure of the encoding process (the process in S1100 in FIG. 2) executed by the image encoder 1050 having the above structure will be described.
[0101]
At S1110, symbol string decomposer 1420 controlled by controller 1410 decomposes the contents of high-resolution image data memory 1030 and stores it in symbol string buffer 1430. At the same time, the symbol string decomposer 1420 generates a counter value of the symbol string counter 1440. The details of the processing procedure in the symbol string decomposer 1420 will be described later.
[0102]
In S1130, data encoder 1450 controlled by controller 1410 encodes the following three symbol strings stored in symbol string buffer 1430 by LZ77 encoding to obtain high-resolution encoded image data. It is stored in the memory 1060. At this time, the symbol string stored in the symbol string buffer 1430 is
Symbol string consisting of SC [0] symbols starting from S [0]
Symbol string consisting of SC [1] symbols starting from S [1]
Symbol string consisting of SC [2] symbols starting from S [2]
It is. Here, three symbol strings are stored because three types of keys, which will be described later, are set. Note that the image coding apparatus 1000 according to the present invention is not limited to the apparatus in which the types of keys are set to three. Furthermore, the image encoding device 1000 according to the present invention is not limited to the data encoding method of LZ77 encoding.
[0103]
FIG. 6 shows an example of the format of the encoded image data stored in the high-resolution encoded image data memory 1060. The data structure in which the encoded image data length 1510 is followed by the encoded image data 1520 is repeated by the number of encoded image data (three times in this example). The coded image data length 1510 is a 4-byte field indicating the length of the coded image data 1520 by the number of bytes.
[0104]
Subsequently, the encoded image data length 1530, the encoded image data 1540, the encoded image data length 1550, and the encoded image data 1560 are successively stored. The relationship between the coded image data length 1530 and the coded image data 1540 and the relationship between the coded image data length 1550 and the coded image data 1560 are all the same as those described above for the coded image data length 1510 and the coded image data 1520. The relationship is the same. Both the coded image data length 1530 and the coded image data length 1550 are 4-byte fields, like the coded image data length 1510. Note that the image encoding device according to the present invention is not limited to such a data format.
[0105]
In S1140, data encoder 1450 controlled by controller 1410 encodes the low-resolution image data stored in low-resolution image data memory 1040 by LZ77 encoding in the raster scan order shown in FIG. It is stored in the low-resolution coded image data memory 1070.
[0106]
As for the format of the low-resolution coded image data memory 1070, as shown in FIG. 7, coded data length 1610 is followed by coded image data 1620. It should be noted that with such a data structure, only low-resolution images can be individually decoded at the time of decoding without decoding the entire image. Also, here, the contents of the low-resolution image data memory 1040 are encoded in the same manner as the high-resolution image data memory 1030, in order to simplify the configuration of the data encoder 1450. Such simplicity is preferred for many applications. Note that the image coding apparatus according to the present invention can be applied to an apparatus to which two or more different coding systems are applied.
[0107]
Hereinafter, details of the symbol string decomposer 1420 will be described.
FIG. 11 shows a block diagram of the symbol string decomposer 1420. The symbol string decomposer 1420 includes a first counter 1424, a corresponding pixel number register 1426, a key register 1428, and a controller 1422 that controls these.
[0108]
The controller 1422 controls the entire symbol string decomposer 1420. The controller 1422 includes a CPU (not shown) and a memory for storing a program executed by the CPU.
[0109]
The first counter 1424 counts a variable I representing the number of elements (the number of pixels) of the high-resolution image data. The corresponding pixel number register 1426 stores an element (pixel) of the low-resolution image data memory 1040 which has a strong correlation and corresponds to an element (pixel) of the high-resolution image data memory 1030. The key register 1428 stores classified key values corresponding to elements (pixels) of the high-resolution image data.
[0110]
With reference to FIG. 12, the procedure of the decomposition processing (the processing of S1110 in FIG. 10) executed by the symbol string decomposer 1420 having the above structure will be described.
[0111]
In S1112, controller 1422 initializes the counter value of first counter 1424 (counter value = 0). Hereinafter, the counter value of the first counter 1424 is indicated by a variable I.
[0112]
In S1114, controller 1422 initializes all elements of symbol string counter 1440, which is a one-dimensional array having three elements (SC [K] = 0, K = 0, 1, 2).
[0113]
In S1116, controller 1422 calculates number J of the element (pixel) of low-resolution image data memory 1040 (strongly correlated) corresponding to the I-th element (pixel) of high-resolution image data memory 1030, and It is stored in the pixel number register 1426.
[0114]
The calculation of which pixel of the high-resolution image corresponds to which pixel of the low-resolution image mainly depends on how the high-resolution image and the low-resolution image are generated by the encoding preprocessor 1020. Here, associating the high-resolution image with the low-resolution image means associating pixels considered to have a high correlation between them. Any rule may be used as long as a rule is provided indicating that each pixel of the high resolution image corresponds to each of the pixels of the low resolution image. However, if the correlation between the corresponding pixels of the high-resolution image and the low-resolution image is low or not at all, efficient coding cannot be achieved.
[0115]
This correspondence will be described using the image example shown in FIG. Assuming that the width of the image is W (here, 4) and the element number of the low-resolution image data memory 1040 corresponding to the I-th element of the high-resolution image data memory 1030 is J, J is calculated as follows. In the following description, the width W of the image is assumed to be an even number for simplicity, and a value obtained by multiplying W by 3/2 is assumed to be V (V = 1.5 × W).
[0116]
When a number I is assigned to each element (each pixel) of the high-resolution image data memory 1030 from 0 in the raster scan order as shown in FIG. 5, the XY coordinates of the I-th element as a pixel on the original image Are divided according to the value of (I mod V) as follows. The XY coordinates are represented in the format (X, Y).
(A) When (I mod V) <W / 2
(2 × (I mod (W / 2)) + 1,2 × (I / V)) (1)
(B) When W / 2 ≦ (I mod V) <W
(I mod (W / 2), 2 × (I / V) +1) (2)
(C) When (I mod V) ≧ W
(I mod (W / 2) + W / 2, 2 × (I / V) +1) (3)
Hereinafter, it is assumed that the elements (pixels) of the low resolution image data memory 1040 are assigned numbers J from 0 in the raster scan order on the low resolution image.
[0117]
Assuming that the element number of the low-resolution image data memory 1040 corresponding to the I-th element of the high-resolution image data memory 1030 is J, the low-resolution image corresponds to １ ／ reduction of the original image, and its width is W / 2. From the equations (1), (2) and (3), J is obtained as follows.
(A) When (I mod V) <W / 2
J = (I mod (W / 2)) + (I / V) × (W / 2) (4)
(B) When W / 2 ≦ (I mod V) <W
J = (I mod (W / 2)) / 2+ (2 × (I / V) +1) / 2 × (W / 2) (5)
(C) When (I mod V) ≧ W
J = ((I mod (W / 2)) + W / 2) / 2 + ((2 × (I / V)) + 1) / 2 × (W / 2) (6)
In S1118, controller 1422 reads the element of the number indicated by the value of corresponding pixel number register 1426 in low-resolution image data memory 1040, converts the value, and stores it in key register 1428. The conversion rule at this time is given as follows. The value of the pixel read from the low resolution image data memory 1040 is P, and the value of the key register 1428 is K.
K = 0 (when P = 0)
K = 1 (when 1 ≦ P ≦ 254)
K = 2 (when P = 255)
Here, the reason why the value of the key register 1428 is set in this manner is that a background pixel (255) and a pixel having a pixel value of a plain text (0) are included in a lot of ordinary paper, and a low-resolution pixel is a background ( This is because if it is a character) pixel, it is considered that the probability that all the high resolution pixels are also background (character) pixels is high. This phenomenon is remarkable in a normal document, particularly in a background pixel.
[0118]
In the case where the value of the background pixel or the value of the character pixel fluctuates, it is desirable that the value of the pixel such that the value of the key register 1428 becomes 0 or 2 has a width. For example, the condition for setting the value K of the key register 1428 to 0 is “0 ≦ P <16”, the condition for setting the value K of the key register 1428 to 1 is “16 ≦ P <240”, and the value K of the key register 1428 is set to The condition for setting 2 may be set to be “240 ≦ P ≦ 255”.
[0119]
Here, the value K of the key register 1428 is designed to take three values, but this is an example, and the image encoding device according to the present invention is not limited to this.
[0120]
Further, another setting of the value K of the key register 1428 may be based on a value obtained by dividing the pixel value P by a certain integer. In this case, the range of values that the value K of the key register 1428 can take also varies depending on the integer value obtained by dividing the pixel value P. Various other variations can be considered for setting the value K of the key register 1428. In setting the value K of the key register 1428, the value K of the high resolution image data memory 1030 corresponding to the pixel read from the low resolution image data memory 1040 can be predicted by the value K of the key register 1428. It is important that the values of the pixels of the high-resolution image data memory 1030 classified by the value K of the key register 1428 concentrate on some values within the same classification. This makes it possible to reduce the entropy within each classification. Therefore, performing encoding for each classification leads to improvement in encoding efficiency as compared to encoding all pixels collectively. More generally, when one of two data series having a high correlation is used as a key and the other is classified into a sub-data series, the values of the elements of each created sub-data series are partially You will concentrate on the value. Therefore, the sum of entropy of each sub data sequence is lower than that of the original data sequence, and coding efficiency can be improved.
[0121]
The length of the elements in the first dimension direction of the symbol string buffer 1430 and the number of elements of the symbol string counter 1440 are all determined by the range of the value K of the key register 1428. Therefore, if the range of the value K of the key register 1428 is changed, such parameters should be changed.
[0122]
At S1120, controller 1422 substitutes the I-th element (pixel) of high-resolution image data memory 1030 into S [K] [SC [K]] of symbol string buffer 1430.
[0123]
At S1122, controller 1422 adds 1 to the K-th value SC [K] of symbol string counter 1440. In S1124, controller 1422 adds 1 to counter value I of first counter 1424.
[0124]
In S1126, controller 1422 compares counter value I of first counter 1424 with the number of pixels included in high-resolution image data memory 1030. If the counter value I of the first counter 1424 is equal to the number of pixels included in the high resolution image data memory 1030 (YES in S1126), this processing ends. If not (NO in S1126), the process returns to S1116, and the process for the next element (pixel) is performed.
[0125]
The operation of the image coding apparatus 1000 based on the above structure and flowchart will be described.
[0126]
The captured image data and the like are stored in the input image data memory 1010 as input image data. The encoding preprocessor 1020 creates high-resolution image data and low-resolution image data based on the input image data stored in the input image data memory 1010 (S1000). Thus, the high-resolution image 1210 shown in FIG. 3 or the high-resolution image 1310 shown in FIG. 4 is stored in the high-resolution image data memory 1030, and the low-resolution image 1220 shown in FIG. Alternatively, a low resolution image 1320 shown in FIG. 4 is stored.
[0127]
The image encoder 1050 encodes the high-resolution image data and the low-resolution image data using a predetermined compression method (S1100). At this time, first, the symbol string decomposer 1420 decomposes the contents of the high resolution image data memory 1030 and stores it in the symbol string buffer 1430 (S1110). The counter value I of the first counter 1424 is initialized (S1112), and the symbol string counter SC [K] is initialized (S1114). The number J of the element (pixel) of the low-resolution image data memory 1040 (highly correlated) corresponding to the I-th element (pixel) of the high-resolution image data memory 1030 is calculated. At this time, the number J is calculated from the variable I using the equations (1) to (6) (S1116). The J-th element in the low-resolution image data memory 1040 is read and converted into a key value K (S1118). The I-th element (pixel) of the high-resolution image data memory 1030 is assigned to S [K] [SC [K]] of the symbol string buffer 1430 (S1120).
[0128]
One is added to the K-th value SC [K] of the symbol string counter 1440 (S1122), and one is added to the counter value I of the first counter 1424 (S1124). Until the counter value I of the first counter 1424 becomes equal to the number of pixels included in the high resolution image data memory 1030 (YES in S1126), the processing from S1116 to S1124 is repeatedly executed, and all elements (pixels) ) Is performed.
[0129]
With such an operation, the elements (pixels) of the high-resolution image data memory 1030 are classified according to the values of the corresponding elements (pixels) of the low-resolution image data memory 1040. Here, when there is a correlation between the value of each pixel of the high resolution image data memory 1030 and the value of the corresponding pixel of the low resolution image data memory 1040, the value of the corresponding low resolution image data memory 1040 is approximated. For example, the value of each pixel in the original high-resolution image data memory 1030 is similar. Therefore, when the high-resolution image data memory 1030 is classified on the condition that the values of the corresponding pixels in the low-resolution image data memory 1040 are similar, it is considered that the pixel data values concentrate on some values in each classification. Can be Stated another way, each classification will have low entropy.
[0130]
Note that, in this classification, the above-described equations (1) to (6) are used. In the case of the image data illustrated in FIG. 3, for example, it is assumed that the pixel (00) and the pixel (22) are approximate values, and the pixel (02) and the pixel (20) are approximate values. In this case, each is converted to the same key value. In this case, pixel (01), pixel (10), pixel (11), pixel (23), pixel (32), pixel (32) of the high-resolution image corresponding to one of pixel (00) and pixel (22). 33) fall into the same category, and correspond to either pixel (02) or pixel (20), pixel (03), pixel (12), pixel (13), pixel (21), pixel (21) of the high resolution image. 30), indicating that pixels (31) fall into the same category. After that, by encoding the pixels of the high-resolution image for each of the above two classifications, it is possible to perform encoding more efficiently than encoding the pixels of the high-resolution image collectively. In this way, the encoded image data is stored in the high-resolution encoded image data memory 1060 in a format as shown in FIG.
[0131]
After that, the following three symbol strings stored in the symbol string buffer 1430 are subjected to encoding processing by LZ77 encoding by the data encoder 1450 and stored in the high-resolution encoded image data memory 1060 (S1130). ).
[0132]
The low-resolution image data stored in the low-resolution image data memory 1040 is coded by LZ77 coding and stored in the low-resolution coded image data memory 1070 (S1140). As a result, the encoded image data is stored in the low-resolution encoded image data memory 1070 in a format as shown in FIG.
[0133]
Further, the data stored in the high-resolution coded image memory 1060 and the data stored in the low-resolution coded image memory 1070 are combined by the coded data combiner 1080, and the data is stored in the combined coded image data memory 1090. (S1200).
[0134]
As described above, according to the image encoding device according to the present embodiment, the elements constituting the high-resolution image data are classified in association with the elements constituting the low-resolution image data. When the encoding process is executed for each of the classified high-resolution image data, high-resolution image data having a high correlation with each other can be collectively and efficiently encoded.
[0135]
This means that the values of the pixels in the low-resolution image are used as a key to classify the corresponding pixels in the high-resolution image, and the coding is performed for each classification. By classifying one of a plurality of data series as a key and the other as a sub-data series, the values of the elements of each created sub-data series are concentrated on some of the values within each. This is because the entropy of the sub data sequence has been reduced.
[0136]
Further, according to the image coding apparatus according to the present embodiment, the coding processing is performed because it can be applied to coding processing without an explicit probability model, such as dictionary-based coding. When decoding encoded image data, the decoding process is lightweight.
[0137]
Furthermore, according to the image encoding device according to the present embodiment, it is possible to configure so that decoding of a specific data sequence does not require decoding of another data sequence. In the case where the outline of the data obtained by integrating is obtained, the outline can be known by decoding only a specific data sequence.
[0138]
<Second embodiment>
Hereinafter, an image decoding apparatus according to the second embodiment of the present invention will be described. The image decoding device according to the present embodiment is a decoding device corresponding to the image encoding device according to the first embodiment.
[0139]
FIG. 13 shows a block diagram of image decoding apparatus 2000 according to the present embodiment. As shown in FIG. 13, the image decoding apparatus 2000 is configured such that the coded image data representing the coded high-resolution image and the coded image data representing the coded low-resolution image are combined. A combined coded image data memory 2010 for storing coded image data, and a coded image data representing a high resolution image and a code representing a low resolution image from the combined coded image data before decoding the combined coded image data Pre-decoding unit 2020 for separating separated high-resolution coded image data, high-resolution coded image data memory 2030 for storing separated high-resolution coded image data, and low-resolution storage for storing separated low-resolution coded image data The encoded image data memory 2040 decodes encoded image data representing a high-resolution image and encoded image data representing a low-resolution image. An image decoder 2050 for processing, a high-resolution image data memory 2060 for storing image data representing a decoded high-resolution image, and a low-resolution image memory for storing image data representing a decoded low-resolution image An image data memory 2070 and an image data combiner 2080 for combining image data representing a decoded high-resolution image and image data representing a decoded resolution image are included.
[0140]
In the image decoding device 2000, the combined encoded image data memory 2010, the high-resolution encoded image data memory 2030, the low-resolution encoded image data memory 2040, the high-resolution image data memory 2060, and the low-resolution image data memory 2070 And the output image data memory 2090 are generally realized by a storage device mounted on a computing system including a computer.
[0141]
In the image decoding device 2000, the pre-decoding processor 2020, the image decoder 2050, and the image data combiner 2080 can be realized by software executed by a computer or by hardware. . In the following description, the pre-decoding processor 2020, the image decoder 2050, and the image data combiner 2080 have specific hardware including a controller, and the controller executes predetermined software. The following description will be made on the assumption that the hardware is controlled by the computer to realize a predetermined function. However, instead of specific hardware, a general computer may be used to realize the predetermined function. Therefore, one essential aspect of the present invention is a program for causing a computer to realize such a function.
[0142]
The encoded image data stored in the combined encoded image data memory 2010, that is, the data that is decoded by the image decoding device 2000 according to the present embodiment, corresponds to the image encoding device according to the first embodiment. 1000 is the encoded image data that has been subjected to the encoding process.
[0143]
Referring to FIG. 14, the procedure of the image data decoding process of image decoding apparatus 2000 according to the present embodiment will be described.
[0144]
In S2000, image decoding apparatus 2000 uses decoding preprocessor 2020 to convert high-resolution encoded image data and low-resolution encoded image data from encoded image data stored in joint encoded image data memory 2010. Are separated and stored in the high-resolution coded image data memory 2030 and the low-resolution coded image data memory 2040, respectively. At this time, the high-resolution coded image data shown in FIG. 6 and the low-resolution coded image data shown in FIG. 7 are separated from the combined coded image data shown in FIG.
[0145]
In S2100, image decoding apparatus 2000 uses image decoder 2050 to decompress the high-resolution encoded image data and the low-resolution encoded image data into a decompression scheme corresponding to the image encoding apparatus according to the first embodiment. , And the decoded data is stored in the high-resolution image data memory 2060 and the low-resolution image data memory 2070. Details of the processing procedure in the image decoder 2050 will be described later.
[0146]
At S2200, image decoding apparatus 2000 combines the data stored in high-resolution image data memory 2060 and low-resolution image data memory 2070 using image data combiner 2080, and stores the combined data in output image data memory 2090. Let it.
[0147]
Hereinafter, the details of the image decoder 2050 will be described.
FIG. 15 shows a block diagram of the image decoder 2050. The image decoder 2050 of the image decoding device 2000 according to the present embodiment executes the LZ77 encoding process used in the image encoding device 1000 according to the first embodiment. The decoding device 2000 is not limited to the LZ77 encoding process. An image decoding device 2000 according to the present invention can be applied to a decoding process corresponding to a general encoding process based on a dictionary.
[0148]
The image decoder 2050 includes a data decoder 2420, a symbol string combiner 2430, a symbol string buffer 2440, a symbol string counter 2450, and a controller 2410 that controls these.
[0149]
The controller 2410 controls the entirety of the image decoder 2050. The controller 2410 includes a CPU (not shown) and a memory for storing a program executed by the CPU.
[0150]
The data decoder 2420 decodes the low-resolution encoded image data separated from the encoded image data and stored in the low-resolution encoded image data memory 2040 and the encoded image data stored in the symbol string buffer 2440. I do.
[0151]
Note that the symbol string buffer 2440 and the symbol string counter 2450 have the same configurations as the symbol string buffer 1430 and the symbol string counter 1440 in the image encoder 1050 of the image encoding apparatus 1000 according to the first embodiment described above. Therefore, detailed description here will not be repeated. Details of the structure of the symbol string combiner 2430 will be described later.
[0152]
With reference to FIG. 16, the procedure of the decoding process (the process of S2100 in FIG. 14) executed by the image decoder 2050 having the above structure will be described.
[0153]
In S2110, data decoder 2420 controlled by controller 2410 decodes the low-resolution coded image data stored in low-resolution coded image data memory 2040.
[0154]
In S2130, data decoder 2420 controlled by controller 2410 decodes the following three symbol strings stored in symbol string buffer 2440 in accordance with LZ77 encoding. At this time, the symbol strings stored in the symbol string buffer 2440 and subjected to decoding processing are the same as in the image encoding apparatus 1000 according to the first embodiment described above.
Symbol string consisting of SC [0] symbols starting from S [0]
Symbol string consisting of SC [1] symbols starting from S [1]
Symbol string consisting of SC [2] symbols starting from S [2]
It is.
[0155]
Further, in image decoding apparatus 2000 according to the present embodiment, encoded image data in low-resolution encoded image data memory 2040 and a symbol string in symbol string buffer 2440 are decoded by the same method. This is for simplifying the configuration of the data decoder 2420. Such simplicity is preferred for many applications. Note that the image decoding apparatus according to the present invention can be applied to an apparatus to which two or more different decoding methods are applied.
[0156]
In S2140, symbol string combiner 2430 controlled by controller 2410 creates high-resolution image data based on the decoded low-resolution image data and the decoded symbol string. The details of the processing procedure in the symbol string combiner 2430 will be described later.
[0157]
Hereinafter, details of the symbol string combiner 2430 will be described.
FIG. 17 is a block diagram of the symbol string combiner 2430. The symbol string combiner 2430 includes a first counter 2434, a corresponding pixel number register 2436, a key register 2438, and a controller 2432 that controls these.
[0158]
The controller 2432 controls the entire symbol string combiner 2430. The controller 2432 includes a CPU (not shown) and a memory for storing a program executed by the CPU.
[0159]
The first counter 2434 counts a variable I representing the number of elements (the number of pixels) of the high-resolution image data. The corresponding pixel number register 2436 stores an element (pixel) of the low-resolution image data memory 2070 having a strong correlation corresponding to the element (pixel) of the high-resolution image data memory 2060. The key register 2438 stores classified key values corresponding to elements (pixels) of the high-resolution image data.
[0160]
With reference to FIG. 18, the procedure of the combining process (the process of S2140 in FIG. 16) executed by the symbol string combiner 2430 having the above structure will be described.
[0161]
In S2142, controller 2432 initializes the counter value of first counter 2434 (counter value = 0). Hereinafter, the counter value of the first counter 2434 is indicated by a variable I.
[0162]
In S2144, controller 2432 initializes all elements of symbol string counter 2450, which is a one-dimensional array having three elements (SC [K] = 0, K = 0, 1, 2).
[0163]
In S2146, controller 2432 calculates the number J of the element (pixel) of low-resolution image data memory 2070 (strongly correlated) corresponding to the I-th element (pixel) of high-resolution image data memory 2060, and It is stored in the pixel number register 2436. The calculation of which pixel of the high-resolution image corresponds to which pixel of the low-resolution image is performed by the encoding preprocessor 1020 in the image encoding device 1000 according to the first embodiment described above. It mainly depends on how the image was generated. The calculation of the number J from the variable I in S2136 is performed according to the equations (1) to (6) described in the first embodiment.
[0164]
In S2148, controller 2432 reads the element of the low-resolution image data indicated by the value of corresponding pixel number register 2436, converts the value, and stores the converted value in key register 2438. The conversion rules at this time are given as follows, as in the first embodiment. Let the pixel value of the low resolution image data be P and the value of the key register 1428 be K.
K = 0 (when P = 0)
K = 1 (when 1 ≦ P ≦ 254)
K = 2 (when P = 255)
As in the first embodiment, the types of K are not limited to three types, and the value of P defining the value of K is not limited as described above.
[0165]
In S2150, controller 2432 substitutes the value of S [K] [SC [K]] in symbol string buffer 2440 for the I-th element (pixel) of the high-resolution image data.
[0166]
At S2152, controller 2432 adds 1 to the K-th value SC [K] of symbol string counter 2450. In S2154, controller 2432 adds 1 to counter value I of first counter 2434.
[0167]
In S2156, controller 2432 compares counter value I of first counter 2434 with the number of pixels of the high-resolution image data. If the counter value I of the first counter 2434 is equal to the number of pixels of the high-resolution image data (YES in S2156), this processing ends. If not (NO in S2156), the process returns to S2146, and the process for the next element (pixel) is further performed.
[0168]
The operation of the image decoding apparatus 2000 based on the above structure and flowchart will be described.
[0169]
The combined encoded image data encoded by the image encoding apparatus 1000 according to the first embodiment is stored in the combined encoded image data memory 2010 as input data. The pre-decoding processor 2020 separates high-resolution coded image data and low-resolution coded image data from the combined coded image data stored in the combined coded image data memory 2010 (S2000). Thus, the high-resolution coded image data memory 2030 stores the high-resolution coded image data shown in FIG. 6, and the low-resolution image data memory 2040 stores the low-resolution coded image data shown in FIG. You.
[0170]
The image decoder 2050 decodes the high-resolution image data and the low-resolution image data using an expansion method corresponding to the compression method executed in advance by the image decoding apparatus 1000 (S2100). At this time, first, the data decoder 2420 decodes the low-resolution coded image data stored in the low-resolution coded image data memory 2040 and stores it in the low-resolution image data memory 2070 (S2110). Next, the data decoder 2420 decodes the symbol string stored in the symbol string buffer 2440 (S2130).
[0171]
The symbol string combiner 2430 combines the data with the decoded low-resolution image data and the symbol string based on the content stored in the symbol string to create high-resolution image data. At this time, first, the counter value I of the first counter 2434 is initialized (S2142), and the symbol string counter SC [K] is initialized (S2144). The number J of the element (pixel) of the low-resolution image data (strongly correlated) corresponding to the I-th element (pixel) of the high-resolution image data is calculated. At this time, the number J is calculated from the variable I using the equations (1) to (6) (S2146). The J-th element of the low-resolution image data is read and converted into a key value K (S2148). The value of S [K] [SC [K]] in the symbol string buffer 2450 is assigned to the I-th element (pixel) of the high-resolution image data (S2150).
[0172]
One is added to the K-th value SC [K] of the symbol string counter 2450 (S2152), and one is added to the counter value I of the first counter 2434 (S2154). Until the counter value I of the first counter 2434 becomes equal to the number of pixels of the high-resolution image data (YES in S2156), the processing from S2146 to S2154 is repeatedly executed, and the processing for all elements (pixels) is performed. Is performed.
[0173]
With such an operation, the element (pixel) of the high-resolution image data is processed according to the value of the element (pixel) of the corresponding low-resolution image data, and as shown in FIG. 2520 to form an output image 2530.
[0174]
As described above, according to the image decoding apparatus according to the present embodiment, elements constituting high-resolution image data are classified in association with elements constituting low-resolution image data, and the classified high-resolution image is classified. By executing the encoding process for each data, it is possible to efficiently decode the encoded image data.
[0175]
<Third embodiment>
Hereinafter, an image decoding device according to the third embodiment of the present invention will be described. The image decoding device according to the present embodiment is a decoding device corresponding to the image encoding device according to the first embodiment, and has, in addition to the functions of the image decoding device according to the second embodiment, It has a function of selecting which of a low-resolution image and a high-resolution image to decode at the time of decoding. Otherwise, the configuration is the same as that of the image decoding device 2000 according to the above-described second embodiment. Therefore, detailed description thereof will not be repeated here.
[0176]
Referring to FIG. 20, a block diagram of image decoding apparatus 3000 according to the present embodiment is shown. As shown in FIG. 20, the image decoding device 3000 includes a selector 3010 connected to an image data combiner 2080 in addition to the configuration of the image decoding device 2000 according to the above-described second embodiment. Based on the content stored in the selector 3010, it is determined whether to decode the low-resolution image or the high-resolution image. If the content stored in the selector 3010 is “0”, high-resolution image data for outputting a high-resolution image is generated. Otherwise, low-resolution image data for outputting a low-resolution image is generated. The other structure is the same as the structure of the image decoding device 2000 according to the above-described second embodiment. Their functions are the same. Therefore, detailed description thereof will not be repeated here.
[0177]
With reference to FIG. 21, a description will be given of a procedure of image data decoding processing of image decoding apparatus 3000 according to the present embodiment. In the flowchart shown in FIG. 21, the same processes as those in the flowchart shown in FIG. 14 described above are denoted by the same step numbers. Their functions are the same. Therefore, detailed description thereof will not be repeated here.
[0178]
In S3000, image decoding apparatus 3000 decodes high-resolution image data or low-resolution image data using image decoder 2050. Details of the processing in S3000 will be described later.
[0179]
In S3100, image decoding apparatus 3000 outputs the decoded high-resolution image data or low-resolution image data, and displays the high-resolution image or the low-resolution image.
[0180]
Referring to FIG. 22, a block diagram including image decoder 2050 and selector 3010 in image decoding apparatus 3000 according to the present embodiment is shown. As shown in FIG. 21, the selector 3010 is connected to a controller 2410 connected to the image decoder 2050. The controller 2410 determines whether to decode the high-resolution image or the low-resolution image based on the content stored in the selector 3010, and controls the data decoder 2420 to execute a necessary decoding process. The other structure is the same as the structure of the image decoding device 2000 according to the above-described second embodiment. Their functions are the same. Therefore, detailed description thereof will not be repeated here.
[0181]
The procedure of the decoding process (the process of S3000 in FIG. 21) executed by the image decoder 2050 having the above structure will be described with reference to FIG.
[0182]
In S3010, data decoder 2420 controlled by controller 2410 decodes the low-resolution coded image data stored in low-resolution coded image data memory 2040.
[0183]
In S3020, controller 2410 determines whether or not the content of selector 3010 is “0”. If the content of selector 3010 is “0” (YES in S3020), the process proceeds to S3030. Otherwise (NO at S3020), the process proceeds to S3040.
[0184]
In S3030, data decoder 2420 controlled by controller 2410 decodes the three symbol strings stored in symbol string buffer 2440 in accordance with LZ77 encoding.
[0185]
In S3040, controller 2410 determines whether or not the content of selector 3010 is “0”. If the content of selector 3010 is “0” (YES in S3040), the process proceeds to S2140. Otherwise (NO at S3040), this process ends. Note that the processing in S2140 is the same as the processing in the image decoding device 2000 according to the above-described second embodiment.
[0186]
The operation of the image decoding device 3000 based on the above structure and flowchart will be described.
[0187]
The pre-decoding processor 2020 separates the combined encoded image data into high-resolution encoded image data and low-resolution encoded image data (S2000). The low-resolution encoded image data is decoded by the data decoder 2420, and the low-resolution image data is stored in the low-resolution image data memory 2070 (S3010).
[0188]
If “0” is stored in selector 3010 (YES in S3020), the symbol string stored in symbol string buffer 2440 is decoded by data decoder 2420 (S3030). Further, if “0” is stored in selector 3010 (YES in S3040), high-resolution image data is created based on the decoded low-resolution image data and the symbol string (S2140). The created high-resolution image data is stored in the output image data memory 2090.
[0189]
On the other hand, if “0” is not stored in selector 3010 (NO in S3020), decoding processing of the symbol string stored in symbol string buffer 2440 is not performed, and high-resolution image data is not created.
[0190]
As described above, according to the image decoding apparatus according to the present embodiment, when the user wants to know the entire image at a coarse resolution quickly and when he wants to know the details of the image at a fine resolution, It is possible to select whether to create a resolution image or a high-resolution image. Such a function exhibits a great effect in an image browsing system and the like.
[0191]
Note that the image decoding apparatus according to the present embodiment selects one of the two types of resolution images, that is, the high resolution and the low resolution. An image may be selected.
[0192]
<Fourth embodiment>
Hereinafter, an image encoding device according to the fourth embodiment of the present invention will be described. The image encoding apparatus according to the first embodiment decomposes an image into two-layer resolutions of a high resolution and a low resolution and executes an encoding process, whereas the image encoding apparatus according to the present embodiment The device has a function of decomposing an image into three or more layers of resolution and executing an encoding process. Otherwise, the configuration is the same as that of the image encoding device 1000 according to the above-described first embodiment. Therefore, detailed description thereof will not be repeated here. In the following description, it is assumed that the image encoding apparatus according to the present embodiment performs encoding processing on image data having (M + 1) types of resolutions.
[0193]
The image encoding apparatus according to the present embodiment recursively repeats the procedure of decomposing and encoding the other data sequence using one of the correlated data sequences as a key, thereby recursively resolving hierarchies of three or more hierarchies. , The encoding process is executed efficiently.
[0194]
FIG. 24 shows a block diagram of image coding apparatus 4000 according to the present embodiment. As shown in FIG. 24, the image coding apparatus 4000 has a resolution of 10% instead of the high-resolution image data memory 1030 and the low-resolution image data memory 1040 of the image coding apparatus 1000 according to the first embodiment. It has (M + 1) data memories of the 0th level image data memory 4020, the first level image data memory 4030, and the Mth level image data memory 4040 from the highest.
[0195]
Further, the image coding apparatus 4000 has a high resolution instead of the high-resolution coded image data memory 1060 and the low-resolution coded image data memory 1070 of the image coding apparatus 1000 according to the first embodiment. From the side, it has (M + 1) data memories of a 0th level coded image data memory 4060, a first level coded image data memory 4070, and an Mth level coded image data memory 4080.
[0196]
Further, this image encoding device 4000 is different from the image encoding device 1000 according to the first embodiment in that the image encoding device 1000 is connected to a data memory instead of the encoding preprocessor 1020. It has an encoding preprocessor 4010 with a different relationship. This image coding apparatus 4000 differs from the image coding apparatus 1050 of the image coding apparatus 1000 according to the first embodiment in the connection relationship with the data memory from the image coding apparatus 1050. An image encoder 4050 is provided. This image coding apparatus 4000 has a connection relationship with a data memory with the coded data combiner 1080 instead of the coded data combiner 1080 of the image coding apparatus 1000 according to the first embodiment. It has a different coded data combiner 4090.
[0197]
Referring to FIG. 25, a procedure of the image data encoding process of image encoding apparatus 4000 according to the present embodiment will be described.
[0198]
In S4000, image coding apparatus 4000 uses (M + 1) types of resolutions (0th level to 0th level) based on input image data stored in input image data memory 1010 using encoding preprocessor 4010. (M level) image data is created and stored in the 0th level image data memory 4020, the first level image data memory 4030, and the Mth level image data memory 4040, respectively.
[0199]
In the process of S4000, for example, as shown in FIG. 4, from the L-th level image (input image 1300) on the higher resolution side, the (L + 1) -th level image whose resolution level is lower by one than that (Low resolution image 1320) is created. Further, the (L + 1) -th level image 1320 is used as a high-resolution side image, and an (L + 2) -th level image having a lower resolution level than that of the (L + 1) -th level image is created. When such a process is repeated (M + 1) times, as shown in FIG. 26, the first level image, the second level image, and the second level image are arranged in order from the highest resolution to the lowest resolution with respect to the input image (the 0th level image). ,..., The (M−1) th level image, and the Mth level image (M + 1) images.
[0200]
That is, an image having a 0-level resolution (corresponding to a high-resolution image) and an image having a first-level resolution (corresponding to a low-resolution image) are created from an input image. An image having a first level of resolution and an image having a second level of resolution are created using the image as an input. As conceptually shown in FIG. 26, by repeating this process, an image having the Mth level of resolution can be reached. Even in the case where the input image and the high-resolution image are different from each other, an image having the Mth level of resolution can be obtained by performing the same recursive processing. The image processed in this way is not limited to the image shown in FIG.
[0201]
In S4100, image encoding apparatus 4000 encodes image data of each level of resolution by a predetermined compression method using image encoder 4050, and encodes the encoded data. , The 0th level encoded image data memory 4060, the first level encoded image data memory 4070, and the Mth level encoded image data memory 4080. Details of the processing procedure in the image encoder 4050 will be described later.
[0202]
In S4200, image coding apparatus 4000 uses coded data combiner 4090, to store the 0th level coded image data memory 4060, the first level coded image data memory 4070, and the Mth level coded image The data stored in the data memory 4080 is combined to create combined encoded image data corresponding to an image of a predetermined level of resolution, and the created encoded image data is stored in the combined encoded image data memory 1090. Let it.
[0203]
FIG. 27 shows a block diagram of image encoder 4050 in image encoding apparatus 4000 according to the present embodiment. As shown in FIG. 27, the image encoder 4050 includes a level register 4052 connected to the controller 1410 in addition to the structure of the image encoder 1050 of the image encoding device 1000 according to the above-described first embodiment. Further included. The level register 4052 stores the resolution level (the higher the number, the lower the resolution). The controller 1410 increments the data stored in the level register 4052 for each process of lowering the resolution by one. When the data reaches a predetermined resolution (M), the image data of that level of resolution is obtained. And ends the processing. The other structure is the same as the structure of the image encoding device 1000 according to the above-described first embodiment. Their functions are the same. Therefore, detailed description thereof will not be repeated here.
[0204]
With reference to FIG. 28, the procedure of the encoding process (the process of S4100 in FIG. 25) executed by the image encoder 4050 having the above structure will be described.
[0205]
At S4110, controller 1410 initializes level register L (L = 0). In S4120, symbol string decomposer 1420 controlled by controller 1410 decomposes the image data having the L-th level of resolution into symbol strings. The decomposed symbol string is stored in the symbol string buffer 1430.
[0206]
At S 4130, data encoder 1450 controlled by controller 1410 performs encoding processing on the symbol string stored in symbol string buffer 1430. At S4140, controller 1410 adds 1 to level register L. At S4150, controller 1410 compares level register L with constant M. If level register L is equal to constant M (YES in S4150), the process proceeds to S4160. If not (NO in S4150), the process returns to S4120, and the process for the image data of the next level of resolution is executed.
[0207]
In S4160, data encoder 1450 controlled by controller 1410 encodes the image data having the resolution of the L-th (= M) level. The combined encoded image data created by repeatedly executing such processing is stored in the combined encoded image data memory 1090.
[0208]
FIG. 29 shows an example of the data format of the combined encoded image data memory 1090. The number-of-layers field 4500 is a 4-byte field indicating how many layers have been encoded. Here, the value (M + 1) is stored as a 4-byte integer. Subsequent to the hierarchy number field 4500, coded image data 4510 corresponding to the Mth level resolution coded image corresponding to the Mth level follows, and thereafter, in the order of decreasing resolution levels (in order of increasing resolution itself). , The first level resolution encoded image data 4520, and the 0th level resolution encoded image data 4530. For the generalized L-th level resolution encoded image data, the contents of the L-th level resolution encoded image data memory are used as they are.
[0209]
As described above, according to the image encoding apparatus according to the present embodiment, it is possible to perform hierarchical decomposition of an image and efficient encoding. Also, by placing data corresponding to a low-resolution image at the beginning of a data string and reading only a part from the beginning of the encoded image data, only that part is decoded and a low-resolution image is displayed. Thus, the outline of the image can be displayed to the user quickly.
[0210]
<Fifth embodiment>
Hereinafter, an image decoding apparatus according to the fifth embodiment of the present invention will be described. The image decoding device according to the present embodiment is a decoding device corresponding to the image encoding device according to the fourth embodiment.
[0211]
An image decoding device according to a fifth embodiment of the present invention will be described. While the image decoding device according to the second embodiment decomposes an image into two-layer resolutions of a high resolution and a low resolution and executes a decoding process, the image decoding device according to the present embodiment includes: It has a function of decomposing an image into three or more layers of resolution and executing decoding processing. Otherwise, the configuration is the same as that of the image decoding device 2000 according to the above-described second embodiment. Therefore, detailed description thereof will not be repeated here. In the following description, it is assumed that the image decoding apparatus according to the present embodiment decodes (M + 1) types of image data.
[0212]
The image decoding apparatus according to the present embodiment recursively performs a procedure of decomposing and encoding the other data series using one of the correlated data series as a key, so as to recursively increase the resolution to three or more layers. On the other hand, decoding processing is performed on the coded image data that has been efficiently coded.
[0213]
FIG. 30 shows a block diagram of image decoding apparatus 5000 according to the present embodiment. As shown in FIG. 30, this image decoding device 5000 is different from the above-described image decoding device 2000 according to the second embodiment in that the high-resolution coded image data memory 2030 and the low-resolution coded image data memory 2040 are replaced by It has (M + 1) data memories of a 0th level coded image data memory 5020, a first level coded image data memory 5030, and an Mth level coded image data memory 5040, in order of higher resolution.
[0214]
Further, this image decoding device 5000 replaces the high-resolution image data memory 2060 and the low-resolution image data memory 2070 of the image decoding device 2000 according to the second embodiment described above, It has (M + 1) data memories of a level image data memory 5060, a first level image data memory 5070, and an Mth level image data memory 5080.
[0215]
Further, the image decoding device 5000 is different from the image decoding device 2000 according to the second embodiment in that the decoding processor 2020 differs from the decoding processor 2020 in the connection relationship with the data memory. It has a pre-processor 5010. This image decoding device 5000 is different from the image decoder 2050 of the image decoding device 2000 according to the second embodiment in that the image decoder 2050 has a different connection relationship with a data memory from the image decoder 2050. Having. This image decoding device 5000 is different from the image data combiner 2080 of the image decoding device 2000 according to the second embodiment in that the image data combiner 2080 differs from the image data combiner 2080 in the connection relationship with the data memory. It has a coupler 5090.
[0216]
Referring to FIG. 31, the procedure of the image data decoding process of image decoding apparatus 5000 according to the present embodiment will be described.
[0217]
In S5000, image decoding apparatus 5000 uses decoding pre-processor 5010 based on the coded image data stored in combined coded image data memory 2010 to provide (M + 1) types of resolutions (0th level to 0th level). (Mth level) encoded image data is created and stored in the 0th level encoded image data memory 5020, the first level encoded image data memory 5030, and the Mth encoded level image data memory 5040, respectively. Remember. At this time, coded image data of (M + 1) kinds of resolutions (0th level to Mth level) is created based on the data format shown in FIG. 29 described in the fourth embodiment. That is, the process of S5000 is a process of decomposing the encoded data into data corresponding to each resolution.
[0218]
In S5100, image decoding apparatus 5000 uses image decoder 5050 to convert the encoded image data of each level of resolution into a compression scheme used in image encoding apparatus 4000 according to the fourth embodiment. Then, the decoded data is stored in the 0th level image data memory 5060, the first level image data memory 5070, and the Mth level image data memory 5080, respectively. The details of the processing procedure in the image decoder 5050 will be described later.
[0219]
At S5200, image decoding device 5000 uses image data combiner 5090 to store the data stored in 0th level image data memory 5060, first level image data memory 5070, and Mth level image data memory 5080. To generate image data of a predetermined level of resolution, and store the generated image data in the output image data memory 2090.
[0220]
FIG. 32 shows a block diagram of image decoder 5050 in image decoding apparatus 5000 according to the present embodiment. As shown in FIG. 32, the image decoder 5050 further includes a level register 5052 connected to the controller 2410 in addition to the structure of the image encoder 1050 of the image decoding device 1000 according to the second embodiment. Including. The level register 5052 stores the resolution level (the higher the number, the lower the resolution). The controller 2410 decrements the data stored in the level register 5052 each time the resolution is increased by one with a predetermined resolution (M) as an initial value, and when the data becomes “0”, The encoded image data having the resolution of the level is decoded, and the process is terminated. The other structure is the same as the structure of the image decoding device 2000 according to the above-described second embodiment. Their functions are the same. Therefore, detailed description thereof will not be repeated here.
[0221]
The procedure of the decoding process (the process of S5100 in FIG. 31) executed by the image decoder 5050 having the above structure will be described with reference to FIG.
[0222]
In S5110, data decoder 2420 controlled by controller 2410 decodes the coded image data having the Mth level of resolution. The decoded image data is stored in the M-th level image data memory 5080.
[0223]
In S5120, controller 2410 substitutes (M-1) for level register L. This content is stored in the level register 5052. In S5130, data decoder 2420 controlled by controller 2410 decodes the data in the symbol string buffer having the L-th level of resolution. In S5140, symbol string combiner 2430 controlled by controller 2410 creates image data of the L-th level of resolution.
[0224]
In S5150, controller 2410 subtracts 1 from level register L. This content is stored in the level register 5052. At S5160, controller 2410 determines whether or not level register L is negative. If level register L is negative (YES in S5160), this process ends. If not (NO in S5160), the process returns to S5130, and the process for the encoded image data of the next level of resolution is executed.
[0225]
Here, the processing of S5100 (S5110 to S5160) will be described in confirmation. After decoding the encoded image data of the M-th level of resolution (S5110), the data stored in the (L + 1) -th level image data memory The L-th level resolution image is decoded (created) from the contents stored in the symbol string buffer 2440 (decoded in S5130) while referring to the (L + 1) level resolution image (S5140), and the L-th level image is obtained. A series of processing of storing the image data in the level image data memory is sequentially performed for L = M−1, M−2,. That is, such decoding processing is sequentially performed, and the encoded data is decoded with reference to images of different levels of resolution. That is, images of each resolution are not independently decoded.
[0226]
In the image decoding device 5000 according to the present embodiment, as shown in FIG. 34, the images of the 0th level, the images of the first level, the images of the second level, .., Output image data can be created based on the (M−1) th level image and the (M + 1) th image of the Mth level image. That is, the decoding process is performed recursively by following the resolution layers one by one. The (M-1) th level image is restored by being combined with the Mth level image, and the (M-2) th level image is restored using the restored (M-1) th level image. Is repeated to indicate that the 0th level image is restored.
[0227]
As described above, according to the image decoding apparatus according to the present embodiment, it is possible to decode coded image data obtained by dividing an image hierarchically and coding efficiently. Also, by placing data corresponding to a low-resolution image at the beginning of a data string and reading only a part from the beginning of the encoded image data, only that part is decoded and a low-resolution image is displayed. Thus, the outline of the image can be displayed to the user quickly.
[0228]
Note that the image decoding apparatus according to the present embodiment is provided with the selector of the image decoding apparatus according to the third embodiment, and the selector is set to what resolution image data is to be decoded. May be decoded into image data of a required resolution.
[0229]
<Sixth Embodiment>
Hereinafter, an image encoding device according to the sixth embodiment of the present invention will be described. The image encoding apparatus according to the first embodiment decomposes an image into two-layer resolutions of a high resolution and a low resolution and executes an encoding process, whereas the image encoding apparatus according to the present embodiment The apparatus has a function of executing a coding process on a continuous frame by using a correlation between frames included in the moving image with a coding target as a moving image. Otherwise, the configuration is the same as that of the image encoding device 1000 according to the above-described first embodiment. Therefore, detailed description thereof will not be repeated here. In the following description, it is assumed that the image encoding device according to the present embodiment performs encoding processing on image data of (M + 1) frames.
[0230]
The image coding apparatus according to the present embodiment replaces the resolution hierarchy in the image coding apparatus according to the fourth embodiment with continuous frames.
[0231]
FIG. 35 shows a block diagram of image coding apparatus 6000 according to the present embodiment. As shown in FIG. 35, the image encoding device 6000 replaces the high-resolution image data memory 1030 and the low-resolution image data memory 1040 of the image encoding device 1000 according to the above-described first embodiment with frames. It has (M + 1) data memories of a 0th frame image data memory 6020, a first frame image data memory 6030, and an Mth frame image data memory 6040, from the earliest time series.
[0232]
Further, the image encoding device 6000 replaces the high-resolution encoded image data memory 1060 and the low-resolution encoded image data memory 1070 of the image encoding device 1000 according to the first embodiment described above with frames. It has (M + 1) data memories of the 0th frame coded image data memory 6060, the first frame coded image data memory 6070, and the Mth frame coded image data memory 6080 from the earliest in the sequence.
[0233]
Further, the image encoding device 6000 includes an imaging device 6010 that captures an image instead of the encoding preprocessor 1010 of the image encoding device 1000 according to the above-described first embodiment. This image coding device 6000 is different from the image coding device 1050 of the image coding device 1000 according to the first embodiment in the connection relationship with the data memory from the image coding device 1050. It has an image encoder 6050. This image coding device 6000 has a connection relationship with a data memory with the coded data combiner 1080 instead of the coded data combiner 1080 of the image coding device 1000 according to the first embodiment. It has a different, coded data combiner 6090.
[0234]
With reference to FIG. 36, the procedure of the image data encoding process of the image encoding device 6000 according to the present embodiment will be described.
[0235]
In S6000, image coding apparatus 6000 stores the image data of the L-th frame captured by imaging apparatus 6010 in the L-th frame image data memory in chronological order. This process is executed until L changes from 0 to M.
[0236]
In S6100, image encoder 6050 of image encoding device 6000 encodes the image data stored in the L-th frame image data memory and stores the encoded image data in the L-th frame encoded image data memory. This process is executed until L changes from 0 to M. At this time, the use of correlation between consecutive frames can be most simply realized by associating pixels at the same physical position in consecutive frames. Using the value of the pixel referenced from the temporally preceding frame as a key, the pixel referenced from the temporally succeeding frame is encoded. This is exactly the same as encoding the pixels of the high-resolution image using the values of the pixels referenced from the low-resolution image as a key in the image encoding device according to the first embodiment. It is. As a method of determining the corresponding pixel, it is conceivable to determine the pixel more precisely by using a motion vector as performed in moving image encoding such as MPEG. The image encoding device according to the present invention is not limited to such a device, and the corresponding pixels may be obtained in any manner.
[0237]
In S6200, encoded data combiner 6090 of image encoding apparatus 6000 combines the encoded image data of each frame and stores the combined image data in combined encoded image data memory 1090.
[0238]
As described above, according to the image coding apparatus according to the present embodiment, by applying to the coding processing of a moving image and coding by using the correlation between different frames, the correlation between frames is obtained. Utilization of the method enables efficient encoding of a moving image composed of continuous frames.
[0239]
<Seventh embodiment>
Hereinafter, an image decoding apparatus according to the seventh embodiment of the present invention will be described. The image decoding device according to the present embodiment is a decoding device corresponding to the image encoding device according to the sixth embodiment.
[0240]
An image decoding device according to a seventh embodiment of the present invention will be described. While the image decoding device according to the second embodiment decomposes an image into two-layer resolutions of a high resolution and a low resolution and executes a decoding process, the image decoding device according to the present embodiment includes: It has a function of executing a decoding process of coded image data obtained by coding a continuous frame by using a correlation between frames included in the moving image with a decoding target as a moving image. Otherwise, the configuration is the same as that of the image decoding device 2000 according to the above-described second embodiment. Therefore, detailed description thereof will not be repeated here. In the following description, it is assumed that the image decoding apparatus according to the present embodiment performs a decoding process on image data of (M + 1) frames.
[0241]
The image decoding device according to the present embodiment replaces the resolution hierarchy in the image decoding device according to the fifth embodiment with continuous frames.
[0242]
FIG. 37 shows a block diagram of image decoding apparatus 7000 according to the present embodiment. As shown in FIG. 37, this image decoding device 7000 is different from the above-described image decoding device 2000 according to the second embodiment in that the high-resolution coded image data memory 2030 and the low-resolution coded image data memory 2040 are replaced by (M + 1) data memories of the 0th frame coded image data memory 7020, the first frame coded image data memory 7030, and the Mth frame coded image data memory 7040 from the earliest time series of the frame Having.
[0243]
Further, the image decoding device 7000 replaces the high-resolution image data memory 2060 and the low-resolution image data memory 2070 of the image decoding device 2000 according to the second embodiment with It has (M + 1) data memories of a 0th frame image data memory 7060, a first frame image data memory 7070, and an Mth frame image data memory 7080.
[0244]
Furthermore, this image decoding device 7000 has a different connection relationship with the data memory from this pre-decoding processor 2020 instead of the pre-decoding processor 2020 of the image decoding device 2000 according to the second embodiment. A decoding preprocessor 7010 is provided. This image decoding device 7000 is different from the image decoder 2050 of the image decoding device 2000 according to the second embodiment in that the image decoding device 7000 has a different connection relationship with a data memory from the image decoder 2050. Having. The image decoding device 7000 has a display device 7090 for displaying a moving image instead of the image data combiner 2080 and the output image data memory 2090 of the image decoding device 2000 according to the second embodiment.
[0245]
With reference to FIG. 38, a procedure of the image data decoding process of the image decoding device 7000 according to the present embodiment will be described.
[0246]
In S7000, image decoding apparatus 7000 separates the encoded image data encoded by image encoding apparatus 6000 according to the above-described sixth embodiment into encoded image data of (M + 1) frames. Then, the image data is sequentially stored in the L-th frame coded image data memory. This process is executed until L changes from 0 to M.
[0247]
In S7100, image decoder 7050 of image decoding apparatus 7000 decodes the encoded image data stored in the L-th frame encoded image data memory, and stores the decoded image data in the L-th frame image data memory. This process is executed until L changes from 0 to M.
[0248]
In S7200, display device 7090 of image decoding device 7000 sequentially outputs and displays the image data of (M + 1) frames stored in the (M + 1) frame image data memory in chronological order.
[0249]
As described above, according to the image decoding device of the present embodiment, it is possible to decode encoded image data obtained by efficiently encoding a moving image by using the correlation between different frames.
[0250]
The embodiments disclosed this time are to be considered in all respects as illustrative and not restrictive. The scope of the present invention is defined by the terms of the claims, rather than the description above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims.
[Brief description of the drawings]
FIG. 1 is a block diagram of an image encoding device according to a first embodiment of the present invention.
FIG. 2 is a flowchart showing a procedure of a process executed by the image encoding device according to the first embodiment of the present invention.
FIG. 3 is a diagram illustrating a first example of separating an input image into a high-resolution image and a low-resolution image in the image encoding device according to the first embodiment of the present invention.
FIG. 4 is a diagram illustrating a second example in which the input image is separated into a high-resolution image and a low-resolution image in the image encoding device according to the first embodiment of the present invention.
FIG. 5 is a diagram showing a scan order when encoding a high-resolution image in the image encoding device according to the first embodiment of the present invention.
FIG. 6 is a diagram showing a format of data stored in a high-resolution coded image data memory in the image coding device according to the first embodiment.
FIG. 7 is a diagram showing a format of data stored in a low-resolution coded image data memory in the image coding device according to the first embodiment.
FIG. 8 is a diagram showing a format of data stored in a combined encoded image data memory in the image encoding device according to the first embodiment.
FIG. 9 is a block diagram of an image encoder in the image encoding device according to the first embodiment.
FIG. 10 is a flowchart illustrating a procedure of a process performed by an image encoder in the image encoding device according to the first embodiment.
FIG. 11 is a block diagram of a symbol string decomposer in the image encoding device according to the first embodiment.
FIG. 12 is a flowchart illustrating a procedure of a process performed by a symbol string decomposer in the image encoding device according to the first embodiment.
FIG. 13 is a block diagram of an image decoding device according to a second embodiment of the present invention.
FIG. 14 is a flowchart illustrating a procedure of a process executed by the image decoding device according to the second embodiment.
FIG. 15 is a block diagram of an image decoder in the image decoding device according to the second embodiment.
FIG. 16 is a flowchart illustrating a procedure of a process performed by an image decoder in the image decoding device according to the second embodiment.
FIG. 17 is a block diagram of a symbol string combiner in the image decoding device according to the second embodiment.
FIG. 18 is a flowchart illustrating a procedure of a process performed by a symbol string combiner in the image decoding device according to the second embodiment.
FIG. 19 is a diagram illustrating an example of generating an output image by combining a high-resolution image and a low-resolution image in the image decoding device according to the second embodiment.
FIG. 20 is a block diagram of an image decoding device according to a third embodiment of the present invention.
FIG. 21 is a flowchart illustrating a procedure of a process performed by the image decoding device according to the third embodiment.
FIG. 22 is a block diagram of an image decoder in the image decoding device according to the third embodiment.
FIG. 23 is a flowchart showing a procedure of processing executed by an image decoder in the image decoding device according to the third embodiment.
FIG. 24 is a block diagram of an image encoding device according to a fourth embodiment of the present invention.
FIG. 25 is a flowchart illustrating a procedure of a process performed by the image encoding device according to the fourth embodiment.
FIG. 26 is a diagram for explaining an operation of an encoding preprocessor in an image encoding device according to a fourth embodiment.
FIG. 27 is a block diagram of an image encoder in an image encoding device according to a fourth embodiment.
FIG. 28 is a flowchart illustrating a procedure of a process performed by an image encoder in an image encoding device according to a fourth embodiment.
FIG. 29 is a diagram illustrating a format of data stored in a combined encoded image data memory in the image encoding device according to the fourth embodiment.
FIG. 30 is a block diagram of an image decoding device according to a fifth embodiment of the present invention.
FIG. 31 is a flowchart showing a procedure of processing executed by the image decoding device according to the fifth embodiment.
FIG. 32 is a block diagram of an image decoder in the image decoding device according to the fifth embodiment.
FIG. 33 is a flowchart illustrating a procedure of processing executed by an image decoder in the image decoding device according to the fifth embodiment.
FIG. 34 is a diagram for explaining an operation of the image data combiner according to the fifth embodiment.
FIG. 35 is a block diagram of an image encoding device according to a sixth embodiment of the present invention.
FIG. 36 is a flowchart showing a procedure of processing executed by the image encoding device according to the sixth embodiment.
FIG. 37 is a block diagram of an image encoding device according to a seventh embodiment of the present invention.
FIG. 38 is a flowchart illustrating a procedure of processing executed by the image encoding device according to the seventh embodiment.
FIG. 39 is a block diagram of an image encoding device related to the present invention.
FIG. 40 is a flowchart showing a procedure of a process executed by the image encoding device related to the present invention.
FIG. 41 is a diagram illustrating a first example of separating an input image into a high-resolution image and a low-resolution image in an image encoding device according to the present invention.
FIG. 42 is a diagram showing the order of scanning when encoding a high-resolution image and a low-resolution image in the image encoding device according to the present invention.
FIG. 43 is a diagram illustrating a second example in which an input image is separated into a high-resolution image and a low-resolution image in an image encoding device according to the present invention.
FIG. 44 is a diagram illustrating a scanning order when encoding a high-resolution image in the image encoding device related to the present invention.
[Explanation of symbols]
1000 image encoding device, 1010 input image data memory, 1020 encoding preprocessor, 1030 high-resolution image data memory, 1040 low-resolution image data memory, 1050 image encoder, 1060 high-resolution encoded image data memory, 1070 low Resolution encoded image data memory, 1080 encoded data combiner, 1090 joint encoded image data memory, 1200, 1300 input image, 1210, 1310 high resolution image, 1220, 1320 low resolution image, 1230, 1240, 1250 pixels, 1410 Controller, 1420 symbol string decomposer, 1422 controller, 1424 first counter, 1426 corresponding pixel number register, 1428 key register, 1430 symbol string buffer, 1440 symbol string counter, 1450 data code Device, 2000 image decoding device, 2010 joint coded image data memory, 2020 decoding preprocessor, 2030 high resolution coded image data memory, 2040 low resolution coded image data memory, 2050 image decoder, 2060 high resolution image data memory , 2070 low-resolution image data memory, 2080 image data combiner, 2090 output image data memory, 2410 controller, 2420 data decoder, 2430 symbol string combiner, 2432 controller, 2434 first counter, 2436 corresponding pixel number register, 2428 key Register, 2430 symbol string combiner, 2440 symbol string buffer, 2450 symbol string counter, 2510 high resolution image, 2520 low resolution image, 2530 output image, 3000 image decoding device, 3010 selector, 4 00 image encoding device, 4010 encoding preprocessor, 4020 0th level image data memory, 4030 first level image data memory, 4040 Mth level image data memory, 4050 image encoder, 4052 level register, 4060 0th level coded image data memory, 4070 1st level coded image data memory, 4080 Mth level coded image data memory, 4090 coded data combiner, 5000 image decoding device, 5010 decoding preprocessor 5020 0th level encoded image data memory, 5030 1st level encoded image data memory, 5040 Mth level encoded image data memory, 5050 image decoder, 5052 level register, 5060 0th level image data Memory, 5070 first level image data Memory, 5080 Mth level image data memory, 5090 image data combiner, 6000 image encoding device, 6010 imaging device, 6020 0th frame image data memory, 6030 first frame image data memory, 6040 Mth frame Image data memory, 6050 image encoder, 6060 0th frame encoded image data memory, 6070 first frame encoded image data memory, 6080 Mth frame encoded image data memory, 6090 encoded data combiner, 7000 image decoding device, 7010 decoding preprocessor, 7020 0th frame coded image data memory, 7030 first frame coded image data memory, 7040 Mth frame coded image data memory, 7050 image decoder, 7060 0th frame image data Memory, 7070 first frame image data memory, 7080 Mth frame image data memory, 7090 display device.

Claims (24)

A data encoding device that encodes data,
Storage means for storing a first data group including a plurality of data, and a second data group including a plurality of data not included in the first data group;
Determining means for determining the correspondence between the data included in the first data group and the data included in the second data group;
Classifying means for classifying the first data group into a plurality of sub-data groups based on the determined association;
A first encoding unit configured to encode data included in the sub data group for each sub data group classified by the classification unit;
A second encoding unit for encoding data included in the second data group,
A data code including a combination unit for combining encoded data obtained by encoding data included in the sub-data group and encoded data obtained by encoding data included in the second data group; Device. The determining means determines the association such that the data included in the first data group and the data included in the second data group have a predetermined relationship with respect to the value of the data. The data encoding device according to claim 1, further comprising: The data is data representing an image, the data includes pixel data corresponding to a plurality of pixels constituting the image,
The determining unit has the relationship based on a position of the pixel data included in the first data group in the image and a position of the pixel data included in the second data group in the image. 3. The data encoding device according to claim 2, further comprising means for determining the association. 4. The data encoding device according to claim 3, wherein said determining means includes means for determining said association by a function using a position in said image. The data encoding device further includes a dividing unit for dividing the image into a plurality of types of resolution images,
The first data group includes data about a high-resolution image,
The data encoding device according to claim 3, wherein the second data group includes data on a low-resolution image. The data is data representing a moving image,
The data encoding device further includes a dividing unit for dividing the moving image into a plurality of frame images according to a time series,
The determining unit is configured to perform processing based on data on an image of a specific frame included in the first data group and data on an image earlier than the specific frame included in the second data group. 3. The data encoding apparatus according to claim 2, further comprising: means for determining the association so as to have the relationship. The data encoding device according to any one of claims 1 to 6, wherein the classification unit includes a unit configured to classify the first data group into a sub data group including data having a strong relationship with each other. The data according to any one of claims 1 to 6, wherein the classification unit includes a unit configured to classify the first data group into a sub-data group including data in which values of data are similar to each other. Encoding device. 9. The method according to claim 1, wherein the type of the encoding process executed by the first encoding unit and the type of the encoding process executed by the second encoding unit are the same. A data encoding device according to claim 1. The data encoding device according to claim 9, wherein the encoding process is an encoding process based on a dictionary. A data decoding device for decoding encoded data,
Storage means for storing the combined first encoded data group and second encoded data group;
Division means for dividing the data group stored in the storage means into the first encoded data group and the second encoded data group;
First decoding means for decoding a coded data included in the first coded data group to create a sub data group;
Second decoding means for decoding the encoded data included in the second encoded data group to create a second data group,
The original data decoded by the data decoding device includes a first data group including a plurality of data and a second data group including a plurality of data not included in the first data group. The sub-data group is a data group in which the first data group is classified based on the correspondence between the data included in the first data group and the data included in the second data group; Each sub data group is encoded into a first encoded data group,
The data decoding device, further comprising a creating unit for creating the first data group according to the association based on the sub data group and the second data group. The creating means may be configured to store the first data in the first data group and the data in the second data group such that the first data group has a predetermined relationship with respect to a data value. The data decoding device according to claim 11, further comprising means for creating a group. The original data is data representing an image, and the original data includes pixel data corresponding to a plurality of pixels forming the image,
The creating unit has the relationship based on a position of the pixel data included in the first data group in the image and a position of the pixel data included in the second data group in the image. 13. The data decoding apparatus according to claim 12, further comprising means for creating the first data group. 14. The data decoding apparatus according to claim 13, wherein the creation unit includes a unit for creating the first data group by a function using a position in the image. The original data includes data representing the image at a plurality of resolutions,
14. The data decoding device according to claim 13, wherein the data decoding device further includes a designation unit for designating the resolution. The data is data representing a moving image,
The creation unit is configured to perform processing based on data on an image of a specific frame included in the first data group and data on an image earlier than the specific frame included in the second data group. 13. The data decoding apparatus according to claim 12, further comprising: means for generating the first data group. The data decoding according to any one of claims 11 to 16, wherein the type of the decoding process executed by the first decoding unit and the type of the decoding process executed by the second decoding unit are the same. apparatus. The data decoding device according to claim 17, wherein the decoding process is a decoding process based on a dictionary. A data encoding method for encoding data, the method comprising:
A preparation step of preparing in advance a first data group including a plurality of data and a second data group including a plurality of data not included in the first data group;
A determining step of determining correspondence between data included in the first data group and data included in the second data group;
A classifying step of classifying the first data group into a plurality of sub-data groups based on the determined association;
A first encoding step of encoding data included in the sub data group for each sub data group classified in the classification step;
A second encoding step of encoding data included in the second data group;
A data encoding method comprising: combining encoded data in which data included in the sub-data group is encoded and encoded data in which data included in the second data group is encoded. . A data decoding method for decoding encoded data,
Preparing a combined first encoded data group and a second encoded data group in advance;
A dividing step of dividing the data group prepared in the preparing step into the first encoded data group and the second encoded data group;
A first decoding step of decoding encoded data included in the first encoded data group to create a sub data group;
Decoding a coded data included in the second coded data group to create a second data group; and
The original data decoded by the data decoding method includes a first data group including a plurality of data, and a second data group including a plurality of data not included in the first data group. The sub-data group is a data group in which the first data group is classified based on the correspondence between the data included in the first data group and the data included in the second data group; Each sub data group is encoded into a first encoded data group,
The data decoding method further includes a creating step of creating the first data group according to the association based on the sub data group and the second data group. A program for causing a computer to realize at least one of the data encoding method according to claim 19 and the data decoding method according to claim 20. A computer-readable recording medium recording the program according to claim 21. A data encoding device that encodes data,
Storage means for storing a plurality of pre-ordered data groups;
Determining means for determining the correspondence relationship from the data included in one data group to the data included in the next data group, of the plurality of ordered data groups,
Classification means for classifying the one data group into sub-data groups based on corresponding data of the next-order data group based on the correspondence relationship,
First encoding means for performing encoding processing on each of the sub-data groups classified based on the classification means;
Second encoding means for encoding the last data group among the plurality of ordered data groups. A data decoding device for decoding encoded data,
Storage means for storing the combined encoded data group;
Dividing means for dividing the combined encoded data group into a plurality of ordered encoded data groups;
First decoding means for decoding a last-order encoded data group among the plurality of ordered encoded data,
Second decoding means for decoding the ordered encoded data group other than the last-order encoded data group to generate a sub data group;
Determining means for determining a correspondence relationship from respective data included in one data group included in the plurality of ordered encoded data to data included in the next-order data group,
An integration unit for integrating the sub-data group into the data group based on the corresponding data of the next-order data group based on the correspondence relationship.

Images