JP4133357B2 - Image coding apparatus and image coding method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
  The present invention relates to an image encoding apparatus.andImage coding methodTo the lawIn particular, an encoding apparatus for encoding image data with different resolutions reduced in stages and its methodTo the lawRelated.
[0002]
[Prior art]
Conventionally, as an image encoding method, there is a method of generating image data of a plurality of resolutions from one image data and encoding it hierarchically. In this image coding method, the image data itself is set to the lowest layer, and the process of generating image data having a lower resolution than that layer as the upper layer image data is repeated, so that the higher the layer image data, the higher the resolution. Low, so-called pyramid type multi-level image data is generated. Then, the generated image data of each layer is encoded for compression (see, for example, Patent Document 1).
[0003]
[Patent Document 1]
Japanese Patent No. 3337160
[0004]
[Problems to be solved by the invention]
However, in the image encoding method for encoding image data of a plurality of layers in a pyramid type, all of the image data having a resolution corresponding to each of the plurality of layers generated from the image data is encoded. The amount of data becomes enormous. For this reason, there is a problem that the amount of data after encoding becomes larger than when only the image data of the lowest hierarchy is encoded.
[0005]
In addition, in the image decoding method for decoding data in which image data of a plurality of layers of a pyramid type is encoded, if only a layer corresponding to a required resolution is specified and only the encoded data of that layer is decoded, Image data with a desired resolution can be obtained. However, all of the image data having the resolution corresponding to the designated hierarchy must be decoded. For this reason, when only a local region of an image is required, it is necessary to decode even an unnecessary part, and there is a problem that unnecessary time is consumed for decoding.
[0006]
The present invention has been made to solve the above-mentioned problems, and one of the objects of the present invention is to efficiently encode pyramid-type image data of a plurality of hierarchies and improve the compression rate. An encoding device is provided.
[0008]
[Means for Solving the Problems]
  In order to achieve the above-described object, according to one aspect of the present invention, an image encoding device reduces the image data by thinning out part of the pixel data included in the image data.Low resolution,Image reduction means for generating image data having a higher hierarchical level;
  Selecting means for selecting processing image data to be processed in order from image data having a higher hierarchical level;
  Processed image data to be processedAboveA dividing unit that divides the pixel data included in the image data having a higher hierarchical level than the processed image data into blocks including at least one;
  AboveFor each block divided by the dividing means, pixel data excluding pixel data included in image data having a higher hierarchical level among pixel data included in the block.As encoding target pixel dataEncoding means for encoding in a recoverable manner; and,
  Reordering means for reordering the encoding target pixel data arranged two-dimensionally into pixel data arranged one-dimensionally in a predetermined order;With,
  The encoding means is encoded using an encoding method based on an adaptive dictionary method including a reference unit and an encoding unit,
  The rearranging means arranges the encoding target pixel data in the horizontal direction in the order adjacent to the horizontal direction of the pixel positions arranged in two dimensions, and when there is no pixel data in the horizontal direction, the pixel in the next row Processing to arrange data in the horizontal direction, or arrange in the vertical direction in the order adjacent to the vertical direction of the pixel positions arranged in two dimensions, and when there is no pixel data in the vertical direction, the pixel data of the next column is Are arranged in a one-dimensional manner by repeating the processing to be arranged until the encoding target pixel data in the block is exhausted,
  The dividing means calculates the product of the number of pixels included in one row in the horizontal direction or one column in the vertical direction of the blocks arranged in one dimension by the rearranging means and the number of bits of pixel data representing gradation. Dividing into blocks each smaller than the number of data stored in the reference section of the encoding scheme and the number of data stored in the encoding section of the encoding scheme.
[0009]
  According to the present invention, processed image data to be processed is selected in order from image data having a higher hierarchical level, and the processed image data includes at least pixel data included in image data having a higher hierarchical level than the processed image data. The pixel data is divided into blocks including one, and for each divided block, pixel data excluding pixel data included in image data having a higher hierarchical level among pixel data included in the block is included.As encoding target pixel dataEncoded to be recoverable.By the function of the rearranging means, the encoding target pixel data arranged two-dimensionally is rearranged to the pixel data arranged one-dimensionally in a predetermined order. The encoding means is encoded using an encoding method based on an adaptive dictionary method including a reference unit and an encoding unit. By the function of the rearrangement means, the encoding target pixel data is arranged in the horizontal direction in the order adjacent to the horizontal direction of the pixel positions arranged in two dimensions, and when there is no pixel data in the horizontal direction, the pixel in the next row When the processing for arranging the data in the horizontal direction or in the vertical direction of the pixel positions arranged in two dimensions is arranged in the vertical direction, and there is no pixel data in the vertical direction, the pixel data in the next column is The processing arranged in the vertical direction is repeated until the coding target pixel data in the block is exhausted, thereby arranging in one dimension. Due to the function of the dividing means, the product of the number of pixels included in one row in the horizontal direction or one column in the vertical direction of the block arranged in one dimension by the sorting means and the number of bits of the pixel data representing the gradation is Is divided into blocks smaller than the number of data stored in the reference unit of the encoding method and the number of data stored in the encoding unit of the encoding method.. When encoding image data with a lower hierarchical level, pixel data included in higher image data is removed, so that pixel data to be encoded can be reduced, and the compression rate is improved. be able to. Further, since encoding is performed in units of blocks, partial decoding can be performed when decoding.In addition, since the pixel data to be encoded is rearranged in a predetermined order, the compression rate can be further improved. Furthermore, since all of the adjacent pixel data is stored in one row in the horizontal direction or one column in the vertical direction of the block in the reference portion of the encoding method, the compression rate can be improved, and the code of the encoding method can be improved. Since all of the adjacent pixel data is stored in the conversion unit in one row in the horizontal direction or one column in the vertical direction, the encoding can be efficiently performed.As a result, it is possible to provide an image encoding device that efficiently encodes image data and improves the compression rate.
[0018]
  According to yet another aspect of the invention,imageThe encoding method has a low resolution by thinning out part of the pixel data included in the image data to reduce the image data.,Generating image data having a higher hierarchy level;
  Selecting processing image data to be processed in order from image data having a higher hierarchical level;
  Processed image data to be processedAboveDividing the pixel data included in the image data having a higher hierarchical level than the processed image data into blocks including at least one;
  AboveFor each block divided by the dividing step, pixel data excluding pixel data included in image data having a higher hierarchical level among pixel data included in the block.As encoding target pixel dataEncoding reversibly and,
  Rearranging the encoding target pixel data arranged two-dimensionally into pixel data arranged one-dimensionally in a predetermined order;IncludingSee
  The step of encoding is performed using an encoding method based on an adaptive dictionary method including a reference unit and an encoding unit,
  The rearranging step arranges the encoding target pixel data in the horizontal direction in the order adjacent to the horizontal direction of the two-dimensionally arranged pixel positions, and when there is no pixel data in the horizontal direction, the pixel in the next row Processing to arrange data in the horizontal direction, or arrange in the vertical direction in the order adjacent to the vertical direction of the pixel positions arranged in two dimensions, and when there is no pixel data in the vertical direction, the pixel data of the next column is Are arranged in a one-dimensional manner by repeating the processing to be arranged until the encoding target pixel data in the block is exhausted,
  In the dividing step, a product of the number of pixels included in one row in the horizontal direction or one column in the vertical direction of the blocks arranged in one dimension in the rearrangement step and the number of bits of pixel data representing gradation is obtained. Dividing into blocks smaller than the number of data stored in the reference section of the encoding scheme and the number of data stored in the encoding section of the encoding scheme.
[0019]
  According to the present invention, when encoding image data with a lower hierarchical level, pixel data included in higher image data is removed, so that pixel data to be encoded can be reduced. . Further, since encoding is performed in units of blocks, partial decoding can be performed when decoding.In addition, since the pixel data to be encoded is rearranged in a predetermined order, the compression rate can be further improved. Furthermore, since all of the adjacent pixel data is stored in one row in the horizontal direction or one column in the vertical direction of the block in the reference portion of the encoding method, the compression rate can be improved, and the code of the encoding method can be improved. Since all of the adjacent pixel data is stored in the conversion unit in one row in the horizontal direction or one column in the vertical direction, the encoding can be efficiently performed.As a result, an image encoding method for efficiently encoding image data can be provided.
[0022]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the following description, the same parts are denoted by the same reference numerals. Their names and functions are also the same. Therefore, detailed description thereof will not be repeated.
[0023]
FIG. 1 is a block diagram showing an outline of functions of an image conversion apparatus 100 according to one embodiment of the present invention. Referring to FIG. 1, an image conversion apparatus 100 includes an encoding processing unit 104 that encodes image data to compress it, and a decoding processing unit 105 that decodes data encoded by the encoding processing unit 104. A program input unit 102 to which programs executed by the encoding processing unit 104 and the decoding processing unit 105 are input, an image input unit 103 to which image data to be processed is input, and output decoded image data An image output unit 106 for storing an image, a storage unit 107 for storing a program and data, and an instruction input unit 101 for receiving an operation input from a user.
[0024]
The image conversion apparatus 100 also includes a scanner 108 that optically reads image data, converts it into electronic data, and outputs it, a display 110 and a printer 111 for outputting image data, an external storage device 112, and a network interface. 109 is connected.
[0025]
The image input unit 103 is connected to the scanner 108, the network interface 109, or the external storage device 112, from which image data is input. The image input unit 103 is also connected to the encoding processing unit 104 and outputs the input image data to the encoding processing unit 104.
[0026]
The encoding processing unit 104 is connected to the program input unit 102 and executes an encoding program received from the program input unit 102 to compress the image data output from the image input unit 103. A process for encoding data is executed. This process includes a process of generating image data having a plurality of resolutions forming a pyramid hierarchy from the image data, and an encoding process for compressing the generated image data having a plurality of resolutions. The encoding processing unit 104 is connected to the storage unit 107 and stores the encoded data in the storage unit 107. The encoding processing unit 104 is also connected to the instruction input unit 101, and executes the above-described processing based on a user operation received by the instruction input unit 101.
[0027]
The decoding processing unit 105 is connected to the program input unit 102 and the storage unit 107, and by executing the decoding program received from the program input unit 102, the data decoded by the encoding processing unit 104 is stored in the storage unit 107. Read from and decrypt. If data encoded by another image conversion device is received via the network interface 109 or read from the external storage device 112 and stored in the storage unit 107, such code is used. It is also possible to decrypt the digitized data.
[0028]
The decoding processing unit 105 is also connected to the instruction input unit 101, and decodes the encoded data based on a user operation accepted by the instruction input unit 101.
[0029]
The external storage device 112 reads or reads image data, encoded image data, or a program from a recording medium such as a flexible disk (FD), a compact disk (CD), a mini disk (MD), or a hard disk (HD). Write. The external storage device 112 is connected to the image output unit 106, the storage unit 107, the program input unit 102, and the image input unit 103, and writes data or programs transmitted from these to a recording medium. Further, the image data recorded on the recording medium or the encoded image data is read out and transmitted to the storage unit 107 or the image input unit 103. Further, the program recorded on the recording medium is read and transmitted to the program input unit 102.
[0030]
The network interface 109 is an interface for connecting the image conversion apparatus 100 and a communication network such as a local area network or the Internet. The image conversion apparatus 100 can communicate with another computer connected to a communication network, and transmits / receives image data, encoded image data, or a program to / from another computer.
[0031]
The scanner 108 is connected to the image input unit 103 and outputs the read image data to the image input unit 103. The display 110 and the printer 111 are connected to the image output device 106 and display or print image data output from the image output device 106.
[0032]
The image conversion apparatus 100 configured as described above can be realized by a generally known computer. Since the hardware configuration and operation of the computer itself are well known, description thereof will not be repeated here.
[0033]
Next, processing executed by the encoding processing unit 104 of the image conversion apparatus 100 in the present embodiment will be described. The encoding processing unit 104 executes processing for generating image data with a plurality of resolutions forming a pyramid hierarchy from image data, and encoding processing for compressing the generated image data with a plurality of resolutions. First, a process for generating image data having a plurality of resolutions will be described.
[0034]
The encoding processing unit 104 generates pyramid-type multi-level image data by generating image data with low resolution based on image data to be processed. Image data consists of data in which pixel data is two-dimensionally arranged. The pixel data may be binary or multi-valued. FIG. 2 is a diagram showing image data with reference numerals indicating pixel positions. Referring to FIG. 2, pixel data is represented by a matrix of 8 columns in the X direction in the figure and 8 rows in the Y direction in the figure. The two-digit number shown in the figure indicates the position of the pixel, the tens place indicates the row, and the first place indicates the column. In the figure, a matrix having 0 to 7 columns in the X direction and 0 to 7 rows in the Y direction is shown. For example, “00” indicates a pixel in 0 row and 0 column, “13” indicates a pixel in 1 row and 3 columns, and “77” indicates a pixel in 7 rows and 7 columns.
[0035]
FIG. 3 is a diagram showing two-layer pixel data of a pyramid type. The image data shown in FIG. 3 indicates pixel data of each layer generated based on the image data shown in FIG. Therefore, the position of the pixel of the image data shown in FIG. 3 is shown by the position attached to each pixel of the image data shown in FIG.
[0036]
FIG. 3A shows the image data of the second hierarchy, and FIG. 3B shows the image data of the first hierarchy. The image data of the first layer has the same resolution as the image data shown in FIG. 2, and the image data of the second layer has half the resolution in the X direction and the Y direction with respect to the image data of the first layer. . The image data of the second layer is lower in resolution than the image data of the first layer and is image data of the upper layer.
[0037]
The image data of the second hierarchy is generated by executing a reduction process for thinning out the pixels of the image data shown in FIG. The pixels to be thinned are determined by the resolution of the image data before processing and the image data to be generated. When the resolution is halved in the X direction, every other interval may be thinned out. When the resolution is 2/3 in the X direction, every other interval may be thinned out.
[0038]
FIG. 4 is a diagram showing pyramid-type three-layer pixel data. The image data shown in FIG. 4 indicates pixel data of each layer generated based on the image data shown in FIG. Therefore, the position of the pixel of the image data shown in FIG. 4 is shown by the position attached to each pixel of the image data shown in FIG.
[0039]
4A shows the image data of the third hierarchy, FIG. 4B shows the image data of the second hierarchy, and FIG. 4C shows the image data of the first hierarchy. The image data of the first layer has the same resolution as the image data shown in FIG. 2, and the image data of the second layer has half the resolution in the X direction and the Y direction with respect to the image data of the first layer. . The image data of the third layer has a resolution of 1/2 in the X direction and the Y direction with respect to the image data of the second layer.
[0040]
Therefore, the image data of the second hierarchy is lower-layer image data having a lower resolution than the image data of the first hierarchy. Similarly, the image data of the third layer is image data of an upper layer having a lower resolution than the image data of the second layer. The image data of the higher hierarchy than the image data of the first hierarchy includes the image data of the second hierarchy and the image data of the third hierarchy.
[0041]
The image data of the second hierarchy is generated by executing a reduction process for thinning out the pixels of the image data shown in FIG. 2, and the image data of the third hierarchy is subjected to a reduction process of thinning out the pixels of the image data of the second hierarchy. Generated by executing.
[0042]
Next, the encoding process will be described. In the encoding process, the image data of each layer is divided into blocks, and the divided blocks are encoded. First, block division will be described with reference to FIGS. 3 and 4. The condition for block division is that each block of image data in a certain layer includes at least one pixel of image data in a higher layer. First, it demonstrates using FIG. The image data of the first hierarchy shown in FIG. 3B is divided into four in the X direction and four in the Y direction, and blocks A00 to A33 are generated. Blocks can also be represented as matrices. For this reason, in the reference numerals attached to the blocks, “A” indicates a block of image data in the first layer, and then the tens place of the number is the position in the column direction (X direction), and the first place is the row direction (Y Direction). For example, the block A00 indicates the position of the 0th row and the 0th column of the first layer image data. Here, each block is a block having two pixels in the X direction and two pixels in the Y direction.
[0043]
The image data of the second hierarchy shown in FIG. 3A is divided into two in the X direction and two in the Y direction, and blocks B00 to B11 are generated. In the code attached to the block, “B” indicates a block of image data in the second layer, and then the tens place of the number is the position in the column direction (X direction), and the first place is the row direction (Y direction). Indicates the position. For example, the block B00 indicates the position of the 0th row and the 0th column of the second layer image data. Here, each block is a block having two pixels in the X direction and two pixels in the Y direction.
[0044]
As is apparent from FIG. 3, each of the blocks included in the second layer image data includes four blocks of the lower first layer image data. Each of the blocks A00 to A33 of the first layer image data includes one pixel of the second layer image data. For example, the block A00 includes the pixel data at the position “00” of the second layer image data, the block A21 includes the pixel data at the position “42” of the second layer image data, and the block A33. Includes pixel data at position “66” of the image data in the second layer.
[0045]
In the image data of the third hierarchy shown in FIG. 4, the image data of the first hierarchy and the second hierarchy are the same as the image data of the second hierarchy shown in FIG. The image data of the third hierarchy shown in FIG. 4A is not divided into blocks because the number of pixels is small. In the reference numerals attached to the blocks, “A” indicates a block of image data in the first layer, “B” indicates a block of image data in the second layer, and “C” indicates a block of image data in the third layer. In the figures after each, the tens place indicates the position in the column direction (X direction), and the first place indicates the position in the row direction (Y direction). For example, the block A00 indicates the position of the 0th row and the 0th column of the first layer image data. Here, each block is a block having two pixels in the X direction and two pixels in the Y direction.
[0046]
As is apparent from FIG. 4, each block included in the third layer image data includes four blocks of the second layer image data, and each block included in the second layer image data includes the second layer image data. Four blocks of image data in one layer are included. Each of the blocks A00 to A33 of the first layer image data includes one pixel of the second layer image data. The relationship between the blocks of the first layer image data and the pixels of the second layer image data is the same as in the case of the image data of the second layer described with reference to FIG. Each of the blocks B00 to B11 of the second layer image data includes one pixel of the third layer image data. The block B00 includes the pixel data at the position “00” of the third layer image data, the block B01 includes the pixel data at the position “04” of the third layer image data, and the block B10 includes the block B10. The pixel data at the position “40” of the image data of the third hierarchy is included, and the pixel data at the position “44” of the image data of the third hierarchy is included in the block B11.
[0047]
In this embodiment, the block size in each layer is a power of 2, but the present invention is not limited to this as long as the block in the upper layer includes a plurality of blocks in the lower layer.
[0048]
Thus, the image data divided | segmented into the block is encoded for every block. In this encoding, the target pixel data does not include the pixel data included in the higher-level image data. For the encoding, a recoverable encoding method based on a dictionary is used. In the present embodiment, the LZ77 encoding method is used. Furthermore, since the encoding method based on the dictionary encodes data arranged in one dimension, it is necessary to rearrange pixel data arranged in two dimensions in one dimension. For this reason, in the present embodiment, the pixel data arranged two-dimensionally is arranged one-dimensionally for each block. More specifically, the pixel data in the block is a process of arranging pixel data excluding pixel data included in the higher order among the pixel data included in the block so as to be continuous in the vicinity of the pixel data included in the higher order. It is arranged in one dimension by repeating until there is no more.
[0049]
In image data, pixels around a certain pixel often have the same value, and are often arranged with pixel data having the same value in the horizontal and vertical directions. In the present embodiment, since the image data is divided into blocks, the number of pixel data continuous in the horizontal direction (X direction) is smaller than in the case where the image data is not divided into blocks. For this reason, pixel data arranged one-dimensionally is arranged relatively close to pixel data adjacent in the vertical direction (Y direction). Therefore, in the LZ77 encoding method, all the pixel data adjacent in the horizontal direction of the block are stored in the reference unit, so that the compression rate can be improved. The horizontal size of the block is preferably smaller than the number of data stored in the reference unit.
[0050]
This will be described in more detail. The number of pixels in the horizontal direction (X direction) of image data is W (dots), the number of bits of pixel data (representing gradation) is k (bits), the length of the reference portion in the LZ77 algorithm is N (bits), and LZ77. Let F (bit) be the length of the encoding part. When the LZ77 algorithm is applied in the conventional raster order (pixel order) in which the entire image data is arranged one-dimensionally in line units, the pixel at the position “00” (hereinafter referred to as “target pixel”) at the upper left of the image data is focused. When the pixel at the position “10” in the next row in the same column is encoded, k × W <N must be satisfied in order for the pixel of interest to be in the reference portion. Similarly, when the pixel at the position “20” in the next row in the same column is encoded, k × 2W <n must be satisfied in order for the pixel of interest to be in the reference portion. This means that the probability that the pixel in the next row exists in the reference unit when encoding is very small just by one pixel away from the pixel in the vicinity of the vertical direction (Y direction).
[0051]
Since the number of pixels in the horizontal direction of the block in this embodiment is L (dots), L <W is established. By doing in this way, the probability that the pixel which exists in the vicinity of the vertical direction exists in a reference part becomes high. Since the image data has a two-dimensional correlation, it can be said that the probability that the longest match in the encoding unit becomes longer increases as the probability that pixels in the vertical vicinity exist in the reference unit increases. Furthermore, it can be said that L pixels in the vicinity have a high probability of matching in the same series of symbol strings. Therefore, it is better that the L pixels can be compared with the reference unit when encoding is performed. Therefore, encoding can be efficiently performed if the length is within the length of the encoding unit, that is, if L × k <F is satisfied. In addition, when L = 1 (dot), the raster order is the same as that of the vertical type, and thus the effect of folding is not obtained. Therefore, it is desirable that the number L of pixels on one side of the block satisfy the following expression. k <L × k <F
1 <L <F / k, L <W (1)
In addition, in an adaptive dictionary method such as the LZ77 algorithm, if there is no dictionary to be referred to in the initial state, expansion may occur rather than compression, and effective compression can be obtained for sufficiently long symbol strings. The reason why an effective compression can be obtained for a sufficiently long symbol string is that a dictionary to be referred to is sufficiently obtained. In other words, if there is a symbol string that can sufficiently accommodate the size of the reference portion of the LZ77 algorithm, effective compression can be obtained. Therefore, effective compression can be obtained by setting a block size such that the capacity (number of pixels) in one block to be encoded is equal to or larger than the size of the reference portion. For example, when describing the lowest hierarchy, if the block size is S (dot), one pixel is k (bit), the size of the reference part of the LZ77 algorithm is N (dot), and the block is a square, The block size S can be obtained from the equation.
[0052]
S × S × k × 3/4> N
S> √ (4 × N / (3 × k)) (2)
In the one-dimensional arrangement, the pixel data is arranged in the vertical direction (Y direction), and when there is no more pixel data in the vertical direction, the process of arranging the pixel data in the next column in the vertical direction is performed. You may make it arrange in one dimension by repeating until there is no data. In this case, it is desirable that the vertical size of the block is smaller than the number of data stored in the reference unit.
[0053]
Furthermore, the one-dimensional arrangement may be made in the order of horizontally adjacent pixels and vertically adjacent pixels, or in the order of vertically adjacent pixels and horizontally adjacent pixels. In any case, adjacent pixels in the vertical direction and adjacent pixels in the horizontal direction may be arranged at an interval smaller than the number of data stored in the reference unit, and such an arrangement order may be determined in advance.
[0054]
FIG. 5 is a diagram illustrating pixel data to be encoded in image data of two layers. FIG. 5 shows the two-layer image data shown in FIG. Referring to FIG. 5 (a), with respect to the image data of the second layer, all the pixel data are to be encoded. This is because there is no image data of a higher hierarchy. Referring to FIG. 5B, the image data of the first layer is a reference for encoding pixel data excluding the pixel data included in the image data of the second layer. For example, the pixel data included in the block A00 includes the positions “00”, “01”, “10”, and “11”, but the pixel data at the position “00” is included in the image data of the second hierarchy. It is not a target for encoding. Thus, the compression rate can be improved by reducing the number of data to be encoded in the lower layer.
[0055]
FIG. 6 is a diagram showing pixel data to be encoded in image data of three layers. FIG. 6 shows the three-layer image data shown in FIG. With reference to FIG. 6A, for the image data of the third layer, all the pixel data are to be encoded. This is because there is no image data of a higher hierarchy. Referring to FIG. 6 (b), the second layer image data is pixel data except for the pixel data included in the third layer image data. For example, the pixel data included in the block B00 includes the positions “00”, “02”, “20”, and “22”, but the pixel data at the position “00” is included in the image data of the third hierarchy. It is not a target for encoding. Referring to FIG. 6C, the image data of the first layer is pixel data except for the pixel data included in the image data of the second layer and the third layer. For example, the pixel data included in the block A00 includes the positions “00”, “01”, “10”, and “11”, but the pixel data at the position “00” is included in the image data of the second hierarchy. It is not a target for encoding. The pixel data included in the block A01 includes positions “02”, “03”, “12”, and “13”, but the pixel data at the position “02” is included in the image data of the second hierarchy. It is not a target for encoding. As described above, in the image data of the lower hierarchy (for example, the first hierarchy), the pixel data to be encoded is the pixel data included in the image data of the higher hierarchy (second and third hierarchy). Not included.
[0056]
FIG. 7 is a flowchart showing a flow of encoding processing executed by the encoding processing unit 104 of the image conversion apparatus 100 according to the present embodiment. Here, for the sake of explanation, an example will be described in which the image data shown in FIG. 2 is encoded with image data of two layers. Referring to FIG. 7, in the encoding process, first, designation of a hierarchy to be encoded is accepted (step S01). The designation of the hierarchy is input by the user operating the instruction input unit 101. Here, since image data is described in two layers, what is specified in step S02 is two layers.
[0057]
Then, designation of the block size to be encoded is accepted (step S02). The designation of the block size is input by the user operating the instruction input unit 101. As described above, it is desirable that the number of pixels on one side of the block satisfies the formula (1) and the block size satisfies the formula (2). Here, it is assumed that a block size of 2 × 2 in length × width is designated. When there is no instruction in step S01 or step S02, the initial setting stored in the program input to the program input unit 102 is used.
[0058]
Next, the image data to be encoded is read from the image input unit or storage unit 107 (step S03). Then, the highest level hierarchy is set (step S04). In step S04, image data of each layer is generated by executing the reduction process by thinning out the image data input in step S03 for the number of layers (two layers) specified in step S01. The image data of the hierarchy is set as the image data to be processed. Here, the image data input in step S03 is the image data of the lowest first layer. Since the designated hierarchy is two, image data of the second hierarchy is generated. Furthermore, the image data of the second highest hierarchy is set as the image data to be processed. Next, the block size in the image data to be processed is set (step S05). In step S05, the block size of the image data to be processed is set based on the block size specified in step S02. Here, the block size of 2 × 2 in length × width shown in FIG. 3A is set. Then, the image data to be processed is divided into blocks based on the set block size (step S06).
[0059]
Next, a non-encoded block among the divided blocks is set (step S07). Here, it is assumed that the block B00 is set from the four blocks B00, B01, B10, and B11 shown in FIG. The pixel data of the set block is rearranged (step S08), and the rearranged pixel data is encoded according to the LZ77 algorithm (step S09). The rearrangement of the pixel data is in the order of the arrows shown in FIG.
[0060]
In the next step S10, it is determined whether or not encoding has been completed for all blocks included in the image data to be processed (step S10). If encoding has not been completed for all blocks, one of those blocks is set as a processing target block, and the processing from step S07 to step S09 is repeatedly executed for the set block. When the processing from step S07 to step S09 is completed for all the blocks of the image data of the processing pair, the process proceeds to step S11.
[0061]
In step S11, it is determined whether or not lower level image data exists. If image data of a lower hierarchy exists, the image data of the lower hierarchy is set as image data to be processed (step S12), and the processes of steps S05 to S10 are executed. If there is no lower-layer image data, that is, if the processing in steps S05 to S10 is executed for the first-layer image data, the processing ends. At this time, the encoded data is input from the encoding processing unit 104 to the storage unit 107 and stored therein.
[0062]
When stored in the storage unit 107, the encoded data is classified and stored for each layer and each block. Information for this classification is included in the header portion of the encoded data. That is, this header portion includes information on which block in which hierarchy corresponds to which encoded data. Therefore, by obtaining the information of this header part, it is possible to obtain only encoded data of an arbitrary block of an arbitrary hierarchy.
[0063]
When the image data of the first layer is the processing target, the block size in step S05 is 2 × 2 in the vertical and horizontal directions. This is the block size specified in step S02. In addition, in the encoding in step S09, pixel data included in the first layer image data is not subjected to encoding. Further, the rearrangement in step S08 is performed in the order indicated by the arrows in FIG.
[0064]
When the image data is encoded as image data of three layers, the processes from step S05 to step S11 are performed in the order of the image data of the respective layers in the order of the image data of the third layer, the image data of the second layer, and the image data of the first layer. Repeatedly executed on data.
[0065]
When the image data of the third hierarchy is the processing target, the block size in step S05 is 2 × 2 in the vertical and horizontal directions. In the encoding in step S09, all pixel data included in the third-layer image data are targeted. Further, the rearrangement in step S08 is rearranged in the order indicated by the arrows in FIG.
[0066]
When the image data of the second hierarchy is the processing target, the block size in step S05 is 2 × 2 in the vertical and horizontal directions. In the encoding in step S09, the pixel data included in the third-layer image data is not a target. Further, the rearrangement in step S08 is performed in the order indicated by the arrows in FIG.
[0067]
When the image data of the first layer is the processing target, the block size in step S05 is 2 × 2 in the vertical and horizontal directions. In the encoding in step S09, pixel data included in the second and third layer image data is not a target. Further, the rearrangement in step S08 is rearranged in the order indicated by the arrows in FIG.
[0068]
As described above, the image conversion apparatus 100 according to the present embodiment does not redundantly encode the upper layer image data that has already been encoded, when encoding the lower layer image data. Therefore, the pixel data to be encoded does not increase from the number of pixels of the input image data, that is, the lowest-order image data (first layer image data). More specifically, referring to FIG. 5B, the lowest-order first-layer image data is composed of three-fourths of pixel data of the original image data. For this reason, the compression capacity can be reduced by reducing the number of pixel data to be encoded.
[0069]
Further, the order in which the pixel data arranged in two dimensions is rearranged in one dimension is the order of adjacent pixels in the image data of each layer, so that similar series of pixels become continuous. Thus, in a method of converting a variable-length symbol string such as LZ77 into a fixed-length or variable-length code word, a longer variable-length symbol string can be converted into a fixed-length or variable-length code word. In other words, since the encoding is efficiently performed by an encoding method based on a dictionary that searches for the longest matching sequence, the compression capacity can be reduced.
[0070]
FIG. 8 is a flowchart showing a flow of decoding processing executed by the decoding processing unit 105 of the image conversion apparatus 100 according to the present embodiment. In the decoding processing unit 105, a decoding processing program is input from the program input unit 102 to the decoding processing unit 105. The decryption processing program is input from the external storage device 112 or the network interface 109. Further, it may be stored in the storage unit 107.
[0071]
Referring to FIG. 8, in the decoding process, designation of a hierarchy level desired to be decoded is accepted (step S21). The hierarchy level indicates a hierarchy corresponding to the resolution of an image to be decoded. If the relationship between the hierarchical level and the resolution is stored in the header portion of the encoded data, the header portion can be read and displayed on the display 110 from the image output portion 106 to notify the user. . In addition, when the user designates the resolution, the hierarchy level may be determined based on the information stored in the header part.
[0072]
Next, designation of an area to be decoded is accepted (step S22). The designation of the hierarchy level and the designation of the decoding target area are input by the user operating the instruction input unit 101. The decoding target area may be a block unit that is a processing unit at the time of encoding, or may include a plurality of blocks.
[0073]
Next, encoded data corresponding to the designated hierarchical level and area is read (step S23). The encoded data corresponding to the designated hierarchical level and area is specified from information included in the header portion of the encoded data. The encoded data corresponding to the designated hierarchical level and area includes, in addition to the encoded data corresponding to the block included in the designated area at the designated hierarchical level, the image of the upper hierarchy including the block. Encoded data corresponding to a block of data is also included. More specifically, data encoded with three-layer image data will be described with reference to FIG. 6. If the designated hierarchy level is the first hierarchy and the designated area is block A00, the code to be read is read out. In addition to the encoded data obtained by encoding the pixel data of the block A00, the encoded data is obtained by encoding the pixel data of the block B00 of the second hierarchy image data including the block A00 and the block C00 of the image data of the first hierarchy. Encoded data is included. Hereinafter, a case where these encoded data are read will be described as an example.
[0074]
Next, the encoded data of the highest hierarchy is set as the processing target data (step S24).
[0075]
In the next step S25, an undecoded block is set among the set processing target data. Then, the encoded data of the set block is decoded (step S26). This encoded data is decoded to obtain pixel data arranged in a one-dimensional manner. In step S26, the decoded pixel data is assigned to the pixel positions of the blocks included in the image data of the first hierarchy having the lowest hierarchy level. Thereby, the pixel data arranged in one dimension is arranged in two dimensions.
[0076]
In step S27, it is determined whether or not all blocks of the set processing target data have been decoded. If true, the process proceeds to step S28, and if false, the process proceeds to step S25. That is, step S25 and step S26 are repeatedly executed until all the blocks of the set processing target data are decoded.
[0077]
Next, the decoded block images are combined (step S28). This process is executed when the encoded data to be processed is composed of a plurality of blocks. Information for combining blocks, that is, block position information is stored in the header portion of the encoded data, and is combined with reference to this information. The position information can be expressed by a matrix such as A00 and A01, for example.
[0078]
Then, the image obtained by combining the decoded block images is interpolated with the decoded image of the higher hierarchy. In step S24, the encoded data of the highest hierarchy corresponds to the data encoded from the image data of the third hierarchy, and when the processing from step S25 to step S28 is executed, the process shown in FIG. The decoded image data is decoded. In this case, nothing is executed in step S29 because there is no higher hierarchy than the image data of the third hierarchy.
[0079]
In the next step S30, it is determined whether or not decoding has been performed up to the designated hierarchy level. If true, the process is terminated, and if false, the process proceeds to step S31. The decoded image data is output from the decoding processing unit 105 to the image output unit 106 or the storage unit 107.
[0080]
In step S31, the encoded data to be processed is set to the encoded data of the next lower hierarchy, and the process proceeds to step S25. The encoded data of the second hierarchy, more specifically, the data obtained by encoding the block B00 of the image data of the second hierarchy is the processing target data. In step S27, the data of the block B00 shown in FIG. Pixel data is decoded. Since the pixel data at the position “00” is not included in the block B00, the pixel data at the position “00” of the higher-level image data is interpolated in step S29. As a result, the interpolated image data includes pixel data at positions “00”, “02”, “20”, and “22”.
[0081]
Further, in step S31, encoded data of the third hierarchy, more specifically, data obtained by encoding the block A00 of the image data of the third hierarchy is set as processing target data. In step S27, FIG. The pixel data of the block A00 shown in FIG. Since the pixel data at the position “00” is not included in the block A00, the pixel data at the position “00” of the image data of the higher hierarchy is interpolated in step S29. As a result, the interpolated image data includes pixel data at positions “00”, “01”, “10”, and “11”. In the above-described processing, the pixel data at the positions “02”, “20”, and “22” are decoded, but do not need to be used because they are not included in the designated area.
[0082]
As described above, since the image conversion apparatus 100 according to the present embodiment decodes encoded image data in units of blocks, only a local area of the image data can be decoded. In addition, since unnecessary areas are not decoded, the processing speed of the decoding process can be increased.
[0083]
The embodiment disclosed this time should be considered as illustrative in all points 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 functional block diagram illustrating an outline of functions of an image conversion apparatus according to one embodiment of the present invention.
FIG. 2 is a diagram showing image data with reference numerals indicating pixel positions.
FIG. 3 is a diagram showing two-layer pixel data of a pyramid type.
FIG. 4 is a diagram showing three-layer pixel data of a pyramid type.
FIG. 5 is a diagram illustrating pixel data to be encoded in image data of two layers.
FIG. 6 is a diagram illustrating pixel data to be encoded in image data of three layers.
FIG. 7 is a flowchart showing a flow of encoding processing executed by an encoding processing unit of the image conversion apparatus according to the present embodiment.
FIG. 8 is a flowchart showing a flow of decoding processing executed by a decoding processing unit of the image conversion apparatus according to the present embodiment.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 100 Image converter, 101 Instruction input part, 102 Program input part, 103 Image input part, 104 Coding process part, 105 Decoding process part, 106 Image output apparatus, 106 Image output part, 107 Storage part, 108 Scanner, 109 Network Interface, 110 display, 111 printer, 112 external storage device.

Claims (2)

Low resolution by reducing image data by thinning out a part of the pixel data included in the image data, and image reduction means hierarchical level to generate image data of the upper,
Selection means for selecting processing image data to be processed in order from image data having a higher hierarchical level;
A dividing unit that divides processed image data to be processed into blocks including at least one pixel data included in image data having a higher hierarchical level than the processed image data;
For each block divided by the dividing unit, pixel data excluding pixel data included in image data having a higher hierarchical level among pixel data included in the block is encoded so as to be reconstructable as encoding target pixel data. Encoding means for ,
Rearrangement means for rearranging the encoding target pixel data arranged two-dimensionally into pixel data arranged one-dimensionally in a predetermined order ,
The encoding means is encoded using an encoding method based on an adaptive dictionary method including a reference unit and an encoding unit,
The rearranging means arranges the encoding target pixel data in the horizontal direction in the order adjacent to the horizontal direction of the pixel positions arranged in two dimensions, and when there is no pixel data in the horizontal direction, the pixel in the next row Processing to arrange data in the horizontal direction, or arrange in the vertical direction in the order adjacent to the vertical direction of the pixel positions arranged in two dimensions, and when there is no pixel data in the vertical direction, the pixel data of the next column is Are arranged in a one-dimensional manner by repeating the processing to be arranged until the encoding target pixel data in the block is exhausted,
The dividing means calculates the product of the number of pixels included in one row in the horizontal direction or one column in the vertical direction of the blocks arranged in one dimension by the rearranging means and the number of bits of pixel data representing gradation. An image encoding apparatus that divides the data into blocks smaller than the number of data stored in a reference unit of an encoding scheme and the number of data stored in an encoding unit of the encoding scheme . And step lower resolution, the hierarchical level to generate image data of the upper by reducing the image data by thinning out a part of the pixel data included in the image data,
Selecting processing image data to be processed in order from image data having a higher hierarchical level;
Dividing the processed image data to be processed into blocks including at least one pixel data included in image data having a higher hierarchical level than the processed image data;
For each block divided by the dividing step, pixel data excluding pixel data included in image data having a higher hierarchical level among the pixel data included in the block is encoded so as to be recoverable as encoding target pixel data. Steps ,
A parallel changing step to the pixel data arranged one-dimensionally in a predetermined given sequence of said coded pixel data arranged in a two-dimensional look-containing,
The step of encoding is performed using an encoding method based on an adaptive dictionary method including a reference unit and an encoding unit,
The rearranging step arranges the encoding target pixel data in the horizontal direction in the order adjacent to the horizontal direction of the two-dimensionally arranged pixel positions, and when there is no pixel data in the horizontal direction, the pixel in the next row Processing to arrange data in the horizontal direction, or arrange in the vertical direction in the order adjacent to the vertical direction of the pixel positions arranged in two dimensions, and when there is no pixel data in the vertical direction, the pixel data of the next column is Are arranged in a one-dimensional manner by repeating the processing to be arranged until the encoding target pixel data in the block is exhausted,
In the dividing step, a product of the number of pixels included in one row in the horizontal direction or one column in the vertical direction of the blocks arranged in one dimension in the rearrangement step and the number of bits of pixel data representing gradation is obtained. An image encoding method in which the number of data stored in the reference unit of the encoding scheme and the number of data stored in the encoding unit of the encoding scheme are each divided into smaller blocks .

Images