JP2003204439A - Image encoding apparatus and image encoding method - Google Patents

Abstract

(57) [Summary] [Image] An image encoding method capable of efficiently encoding a mixed image including areas having different required resolution levels such as a character area and a photograph area in a state where a high-quality image can be restored. An apparatus and a control method thereof are provided. An image input from an image input unit is divided into a plurality of tiles by a tile division unit.
The image reducing unit 104 generates a reduced image from the tile. Then, for each tile, the area information is input from the area information input unit 102, the area information is determined by the area determination unit 105, and the tile or the reduced tile is selected by the selector 106. Further, the discrete wavelet transform unit 107 performs a discrete wavelet transform on the selected tile and decomposes the tile into a predetermined number of subbands. The coefficients of the subbands are quantized by the transform coefficient quantization unit 108, coded by the bit plane coding unit 109, and converted into a bit sequence by the code sequence forming unit 110.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image coding apparatus for coding image data in which characters and photographs are mixed and a control method thereof.

[0002]

2. Description of the Related Art In recent years, the resolution of an image captured by an image input device has been increasing with the improvement of the technology relating to the image input device such as a digital camera and a scanner. An image captured using a conventional image input device has a low resolution and a small amount of image data, so that it does not cause a big problem in various image processing such as transmission and storage of image data.

[0003]

However, as the resolution of the image captured by the image input device becomes higher, the data amount of the image also becomes enormous.
For this reason, there are problems that it takes a lot of time to transmit an image and a large storage capacity is required to store the image.

In order to solve this problem, usually, at the time of image transmission and storage, high efficiency coding is used to remove image redundancy or to process the image within a visually permissible range. The amount of data is being reduced. An encoding method that can completely reproduce the original image by decoding the encoded image data is called lossless encoding. Although a visually similar image is obtained, the original image is completely reproduced. A coding method that cannot do this is called lossy coding.

In lossy encoding, it is important to reduce the code amount by changing a portion where deterioration is not visually noticeable, but this largely depends on the characteristics of an image. There are various types of images, and images generated by scanning a silver salt photograph of a person, landscape, etc. with a scanner, or a natural image generated by directly shooting a person, landscape, etc. with a digital camera, Characters generated by raster scanning and inputting character / line information on media such as paper
Examples thereof include line images, images drawn by computer graphics (CG), and CG images generated by rendering a constructed three-dimensional model.

In the various types of image data described above, the resolution and the number of gradations required to obtain good image quality are usually different. Generally, a character / line image requires higher resolution than a natural image.

Therefore, as a highly efficient encoding method,
Conventionally, the method of utilizing the wavelet transform has been used. In this method, first, an image to be encoded is divided into a plurality of frequency bands (subbands) by using the discrete wavelet transform. Next, the transform coefficient of each subband is quantized and entropy-coded by various methods to generate a code string. An example of the process of wavelet transform will be described with reference to FIG. 7. As a method of the wavelet transform, as shown in FIG. 7 for the original image shown in FIG.
As shown in (b), a one-dimensional conversion process is applied in the vertical direction, and then in the horizontal direction as shown in FIG. 7 (c) to divide into four subbands. To be Further, it is general to repeat division into four subbands only for the low frequency subband LL shown in FIG. 7C.

FIG. 8 illustrates a two-dimensional wavelet transform.
It is a figure which shows the example of division | segmentation of a subband at the time of repeating and repeating. One of the advantages of image coding using wavelet transform is that it is easy to realize stepwise decoding of spatial resolution. As shown in FIG. 8, the two-dimensional wavelet transform is performed twice to decompose into seven subbands, and the coefficients of each subband are sequentially encoded and transmitted from the low frequency subband LL to the high frequency subband HH2. In this case, the decoding side can decode the restored image with the horizontal / vertical 1/4 resolution at the stage of receiving the transform coefficient of the low-frequency subband LL. Also, sub-band LL,
When LH1, HL1, and HH1 are received, the restored image with the horizontal / vertical 1/2 resolution can be decoded. Furthermore, when subbands LH2, HL2, and HH2 are received, the restored image of the original resolution can be decoded. In this way, the decoded image can be generated by selecting the resolution for decoding.

Further, as seen in a mixed image of characters and photographs, as a coding method in the case where parts having different required resolutions are mixed in the image, the image is called a tile and is a rectangular coded independently. By dividing into regions and considering the required resolution for each tile, for tiles that do not require high resolution, the coded data corresponding to the high frequency subband is discarded, or the coefficient of the high frequency subband is not encoded. There is a method to change the resolution for each tile. For example, when each tile is divided into 7 subbands as shown in FIG. 8, high resolution is required for tiles belonging to the character / line image area, and therefore L
L, LH1, HL1, HH1, LH2, HL2, HH2
For the tiles belonging to the natural image area, low resolution is sufficient, so LL and LH
Only four subbands of 1, HL1 and HH1 are encoded.

However, in the above-mentioned conventional high-efficiency encoding method, in the case of encoding image data including regions having different required resolutions in a natural image, a mixed image of characters and line images, etc., a high resolution is required. For a region that does not require, for example, a natural image region, the restored image obtained on the decoding side is a low-frequency component obtained by the wavelet transform, so blunt edges and disappearance of thin lines occur, and it is not always good to reproduce. It didn't look like an image.

The present invention has been made in consideration of such circumstances, and a high-quality image can be restored from a mixed image including regions having different required resolution levels such as a character region and a photo region. It is an object of the present invention to provide an image coding apparatus and a control method thereof that can be coded efficiently in.

[0012]

In order to solve the above-mentioned problems, the present invention inputs a dividing means for dividing an image into a predetermined area to generate a small image, and area information for each divided area. For the input means and each divided area,
Generating means for generating a reduced image at a predetermined reduction ratio, and decomposing means for frequency-converting the small image or reduced image for each region based on the region information of each divided region, and decomposing into a predetermined number of subbands. , Encoding means for encoding the decomposed transform coefficients of each sub-band.

[0013]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings. In the present embodiment, as an image to be encoded, image data obtained by reading a black-and-white document in which a character region and a photo region are mixed with a scanner or the like and expressing it with an 8-bit luminance value is used. The photographic area may include the entire area other than the character area expressed in gray scale. In addition, the image data that can be encoded in the present invention is not limited to an 8-bit image, and image data that expresses a luminance value with a bit number of 4 bits, 10 bits, 12 bits, or the like can be applied. In addition, each pixel has a plurality of color components such as RGB and CMYK, or luminance and chromaticity such as YCbCr.
It is also applicable to color image data represented by color difference components. In the case of such color image data, each of the component data in the color image data may be regarded as black and white image data.

<First Embodiment> FIG. 1 shows a first embodiment of the present invention.
FIG. 3 is a block diagram showing a configuration of an image encoding device for encoding an image according to the embodiment of FIG. In FIG. 1, an image input unit 101 is a device for inputting image data encoded by the image encoding device according to the present embodiment.

Here, it is assumed that the image data input in the present embodiment is a high-resolution image, the image size in the horizontal direction is X pixels, and the image size in the vertical direction is Y pixels. Then, the x-th pixel data in the horizontal direction and the y-th pixel data in the vertical direction of the input image data are set to P (x, y).
Express. Here, 0 ≦ x <X and 0 ≦ y <Y. For example, as a device that realizes the image input unit 101, there is an imaging part such as a scanner or a digital camera. The image pickup portion includes an image pickup device such as a CCD and various image adjustment circuits such as gamma correction and shading correction.

Next, the tile division unit 103 is connected to the image input unit 101, stores the image data input from the image input unit 101 in a buffer (not shown) as appropriate, and stores the input image data in a predetermined manner. Divide into tiles of width TW and height TH. Here, it is assumed that the horizontal and vertical image sizes X and Y of the image data are predetermined multiples of the horizontal and vertical image sizes TW and TH of the tile.

FIG. 2 is a diagram for explaining image data divided into a plurality of tiles. As shown in FIG. 2, the tile including the pixel at the upper left corner that is the origin of the image data is tile T0, and tiles adjacent to tile T0 are numbered T1, T2, ..., Tn in raster scan order. give. Here, the x-th pixel value in the horizontal direction and the y-th pixel value in the vertical direction in each tile are represented as Ti (x, y). However, i
= 0, ..., N, 0 ≦ x <TW, and 0 ≦ y <TH.
Ti (0,0) represents the pixel value at the upper left corner of the tile Ti.

The tile division unit 103 is further connected to the image reduction unit 104 and the selector 106, and outputs the divided image data of each tile to each device in the raster scanning order of tiles T0 to Tn.

On the other hand, the area information input section 102 receives high resolution coded area information indicating a range of an area for which high resolution is required for the image data input from the image input section 101. FIG. 3 is a schematic diagram for explaining the image data to be encoded and the high resolution encoded area according to the present embodiment. As shown in FIG. 3, it is assumed that the region requiring high resolution in this embodiment is a character region. Then, in a rectangular area including the character area in the image data,
Pixel position (UL _x , UL _y ) at the upper left corner and pixel position (L at the lower right corner) based on the image size at high resolution
Two pieces of coordinate information (R _x , LR _y ) are input from the area information input unit 102 as high resolution encoded area information.

The area determination unit 105 includes an area information input unit 10
2 and the pixel position (UL _x , UL _y ) in the upper left corner and the pixel position (LR _x , L) in the lower right corner of the high resolution encoded area in the image data input from the area information input unit 101.
R _y ), it is determined whether each tile Ti generated by the tile dividing unit 103 is a tile including a high-resolution encoded area or a tile not including a high-resolution encoded area, and each tile Ti The judgment result is output to. As this output, the determination result for the tile Ti is F (Ti), and if the tile Ti includes a high-resolution coded area, then F (T
i) = 1, and if not included, F (Ti) = 0. FIG. 4 is a diagram showing an example of a case where two determination results are output for a tile of image data by the method described above.

The image reduction unit 104 is a tile division unit 103.
The reduced tile STi connected to the selector 106 and each tile output from the tile division unit 103 is reduced in size to 1/2 in both the horizontal and vertical directions.
To generate.

The image reduction process in the image reduction unit 104 is a reduction process that emphasizes the image quality of the reduced image. Any method may be used as long as it can obtain a reduced image having a higher image quality than the LL subband in the discrete wavelet transform unit 107, which will be described later. However, since many methods have been proposed, description thereof will be omitted here. The x-th pixel value in the horizontal direction and the y-th pixel value in the vertical direction in the reduced tile STi are represented as STi (x, y). However, i = 0, ...,
n, 0 ≦ x <TW / 2, 0 ≦ y <TH / 2, and S
Ti (0,0) represents the pixel value at the upper left corner of the reduced tile STi.

The selector 106 further includes the area determination unit 10
Connected to 5. In the selector 106, for the tile Ti of interest, based on the area determination information F (Ti) from the area determination unit 105, the image data of the reduced tile STi output from the image reduction unit 104 or the tile division unit 103 outputs. One of the image data of the tiles Ti to be selected is output. For example, when the area determination information F (Ti) regarding the tile Ti is 0, that is, when the tile Ti is a photo area that does not include the high-resolution encoded area that is a character area, the image data of the reduced tile STi is selected. Is output. On the other hand, when the area determination information F (Ti) regarding the tile Ti is 1, that is, when the tile Ti includes a high-resolution encoded area that is a character area,
The image data of the high resolution tile Ti is selected and output.

Further, in FIG. 1, the discrete wavelet transform section 107 is connected to the selector 106 and the area determination section 105. Then, the discrete wavelet transform unit 10
Reference numeral 7 performs two-dimensional discrete wavelet transform while appropriately storing the image data of the tile Ti or the reduced tile STi input from the selector 106 in an internal buffer (not shown), and decomposes the image into a plurality of subbands. Output the coefficient of each subband.

The number of applications of the two-dimensional discrete wavelet transform performed by the discrete wavelet transform unit 107 is
Area determination information F (T
It depends on i). That is, F (Ti)
Is 0, one-dimensional two-dimensional discrete wavelet transform is performed on the tile STi. Figure 5 shows tile STi 4
It is a schematic diagram for explaining the case where it decomposes | disassembles into one subband. As shown in FIG. 5, the tile STi is LL, H
It is decomposed into four subbands L1, LH1, and HH1. Further, FIG. 6 is a schematic diagram for explaining a case where the tile Ti is decomposed into seven subbands. As shown in FIG. 6, when F (Ti) is 1, LL, HL1, and LH are obtained by performing the two-dimensional discrete wavelet transform twice.
It is decomposed into seven subbands of 1, HH1, HL2, LH2, and HH2.

Thereafter, the coefficient of each subband is C (S, x,
y). However, S represents a subband, and LL and HL
1, LH1, HH1, HL2, LH2, or HH2. Further, (x, y) represents the coefficient position in the horizontal direction and the vertical direction when the coefficient position at the upper left corner in each subband is (0, 0).

The two-dimensional discrete wavelet transform used in this embodiment will be described here. The two-dimensional discrete wavelet transform is realized by applying a one-dimensional transform (filter process) to each of the horizontal and vertical directions. FIG. 7 is a schematic diagram for explaining an application example of the two-dimensional discrete wavelet transform for the image to be encoded. FIG. 7A shows an image to be encoded,
A one-dimensional discrete wavelet transform is applied to this image in the vertical direction and decomposed into a low frequency subband L and a high frequency subband H as shown in FIG. 7B. Next, a one-dimensional discrete wavelet transform is applied to each subband in the horizontal direction to decompose into four subbands LL, HL, LH, and HH as shown in FIG. 7C.

In the image coding apparatus according to this embodiment, N
The one-dimensional discrete wavelet transform for each of the one-dimensional signals x (n), (n = 0, ..., N-1) is performed using the following equation.

H (n) = x (2n + 1)-(x (2n) + x (2n + 2)) / 2 (1) l (n) = x (2n) + (h (n-1) + h (n)) / 4 (2) where h (n) is the conversion coefficient of the high frequency subband, l
(N) represents the conversion coefficient of the low frequency subband. It should be noted that both ends x of the one-dimensional signal x (n) required in the calculation of the above equation
(N) and (n <0 and n ≧ N) are obtained from the values of the one-dimensional signal x (n) and (0 ≦ n <N) using a known method.

On the other hand, for the sub-band Ti, the two-dimensional discrete wavelet transform is repeatedly applied to the sub-band LL obtained by the above-mentioned two-dimensional discrete wavelet transform, and as shown in FIG. , HL1, LH1, HH1, HL2, LH2, H
It decomposes into 7 subbands of H2. FIG. 8 is a diagram for explaining the two-dimensional discrete wavelet transform for the subband Ti. The sub-band LL in FIG.
7 is a re-decomposition of the subband LL of FIG. 7C, and the two are not the same.

The transform coefficient quantization unit 108 is connected to the discrete wavelet transform unit 107, and the generated quantization coefficient C (S, x, y) of each subband is determined for each subband by a quantization step delta ( Quantize using S).
When the quantized coefficient value is represented by Q (S, x, y), the quantization process performed by the transform coefficient quantization unit 108 can be represented by the following equation.

Q (S, x, y) = sign {C (S, x, y)} × floor {| C (S, x, y) | / delta (S)} (3) where sign { I} is a function that represents the sign of the integer I, and returns 1 when I is positive and -1 when I is negative. Also, floor {R} represents the maximum integer value that does not exceed the real number R.

On the other hand, the bit plane coding unit 109
The tiles LL, HL1, and LH are connected to the transform coefficient quantization unit 108, and the tiles STi generated through the discrete wavelet transform unit 107 and the transform coefficient quantization unit 108.
1, HL, HL1, LH1, HH1, HL2, LH for 4 subbands of HH1 or tile Ti
2, quantized coefficient values Q of 7 subbands of HH2
It is a device that encodes (S, x, y) to generate a code string.

As a coding method in the bit plane coding unit 109, there is known a method of improving the random accessibility by dividing the coefficients of each subband into blocks and coding them separately. In the form, coding is performed in subband units for the sake of simplicity. That is, the quantized coefficient values Q (S,
(x, y) is encoded by expressing the absolute value of the quantized coefficient value Q (S, x, y) in the subband in a natural binary number, and from the upper digit to the lower digit in the bit plane direction. Is prioritized and binary arithmetic coding is performed. In the present embodiment, the quantized coefficient values Q (S, x,
The description will be made by describing the bit of the nth digit from the bottom when y) is expressed in natural binary as Cn (S, x, y). Also, 2
A variable n representing a digit of a decimal number is called a bit plane number, and the bit plane number n has the LSB (least significant bit) as the 0th digit.

FIG. 9 is a flow chart for explaining the operation procedure of the bit plane coding unit 109 for coding the subband S. First, the absolute value of the coefficient in the sub-band S to be encoded is checked, and its maximum value Mabs
(S) is obtained (step S901). Next, the number of digits N _BP (S) required to represent the maximum value Mabs (S) in binary number is calculated based on the following equation (step S9).
02).

N _BP (S) = ceil {log ₂ (Mabs (S))} (4) where ceil {R} represents the smallest integer value equal to or greater than the real number R.

Then, the number of significant digits N _BP (S) is substituted for the bit plane number n (step S903). further,
Subtract 1 from the bit plane number n (step S90
4). Then, the bit plane n is encoded using the binary arithmetic code (step S905). In this embodiment, QM-Coder is used as the arithmetic code. still,
A certain state (context) using QM-Coder
Procedure to encode the binary symbol generated in
Regarding the initialization procedure and termination procedure for arithmetic coding processing, the international standard for still images, ITU-T Recommendation T.81 |
Since it is explained in detail in ISO / IEC10918-1 recommendation etc., the explanation is omitted here.

In the present embodiment, each bit is arithmetically encoded with a single context for the sake of simplicity. Then, the arithmetic encoder is initialized at the start of encoding each bit plane, and the termination processing of the arithmetic encoder is performed at the end.

Further, for each coefficient, immediately after the first coded "1", the sign of the coefficient is represented by 0 or 1 and arithmetically coded. In this embodiment, 0 is set when the sign of the coefficient is positive, and 1 is set when the sign is negative. For example, when the coefficient is -5 and the number of significant digits N _BP (S) of the subband S to which this coefficient belongs is 6, the absolute value of the coefficient is binary 0.
It is represented by 00101 and is encoded from the upper digit to the lower digit by encoding each bit plane. Then, when encoding the second bit plane (fourth digit from the top), the first "
1 "is encoded, and immediately after this, a positive / negative sign when negative"
1 "is arithmetically encoded.

Further, the bit plane number n is compared with 0 to determine whether n is 0 (step S90).
6). As a result, when n is 0 (Yes), that is,
If the LSB plane has been encoded in step S905, the subband encoding process ends. On the other hand, when n is not 0 (No), the process of step S904 is performed.

By the processing of the above-mentioned procedure, all the coefficients of the subband S are encoded to generate the code string CS (S, n) corresponding to each bit plane n. Code string forming unit 1
10 is connected to the bit plane encoding unit 109, the generated code string is sent to the code string forming unit 110,
It is temporarily stored in a buffer (not shown) in the code string forming unit 110. The code string forming unit 110 is further connected to the code output unit 111. Then, the code string forming unit 110
Is a predetermined order after all the code strings of the coefficients of all subbands of the tile Ti or the reduced tile STi for all tiles (T0 to Tn) output from the bit plane encoding unit 109 are stored in the internal buffer. Reads the code string stored in the internal buffer. Then, necessary additional information is inserted into the read code string to form a final code string that is the output of the present coding apparatus, and is output to the code output unit 111.

FIG. 10 is a diagram showing the structure of a code string generated by the code string forming unit 110. Code string forming unit 1
The final code string generated in 10 includes a tile header and coded data of tiles T0 to Tn. In addition, in the encoded data of each tile, a tile header storing identification information for identifying which tile in the image data, additional information such as the number of discrete wavelet transforms applied to each tile, and a level 0 And level 1 or level 0, level 1 and level 2 in three layers.

FIG. 11 is a schematic diagram showing the structure of the encoded data of the tile STi when F (Ti) = 0. Further, FIG. 12 is a schematic diagram showing the configuration of encoded data of tile Ti when F (Ti) = 1. 11 and 1
2, the encoded data of level 0 is CS (LL, N) obtained by encoding the coefficients of the LL subband.
_{It is} composed of code strings from _BP (LL) -1) to CS (LL, 0).

Level 1 is LH1, HL1, HH
Code sequence C obtained by encoding the coefficients of each subband of 1
From S (LH1, N _BP (LH1) -1) to CS (LH1,
0) and CS (HL1, N _BP (HL1) -1) to CS
(HL1,0) and CS (HH1, N _BP (HH1)-
1) to CS (HH1,0). Further, level 2 has a code string CS (LH2, NH) obtained by coding the coefficients of the subbands LH2, HL2, and HH2.
_{From BP} (LH2) -1) to CS (LH2,0), CS
From (HL2, N _BP (HL2) -1) to CS (HL2,
0) and CS (HH2, N _BP (HH2) -1) to CS
(HH2,0).

The code output unit 111 is a device that is connected to the code string forming unit 110 and outputs the code string generated by the code string forming unit 110 to the outside of the device. Examples of the code output unit 111 include a storage device such as a hard disk and a memory, a network line interface, and the like.

FIG. 13 is a flow chart for explaining the operation procedure of the image coding apparatus according to the above-mentioned embodiment. First, from the image input unit 101, high resolution X pixels x
Image data of Y pixels is input (step S13).
1). The input image data is divided by the tile division unit 103 into tiles having a size of TW × TH (step S132). Then, for each of the divided tiles, a reduced image is generated with a reduction ratio of 1/2 in both the vertical and horizontal directions (step S133).

On the other hand, area information about each divided area is input from the area information input unit 102, and which of the tiles or reduced tiles is selected for each area based on the area information of each divided area. However, the area determination unit 105
Is determined (step S134). Then, the selector 106 selects the determined tile (step S135). Furthermore, the discrete wavelet transform unit 1
At 07, the discrete wavelet transform of the image data of the selected tile is performed and decomposed into a predetermined number of subbands (step S136).

Further, the decomposed transform coefficient of each sub-band is quantized by the transform coefficient quantizing unit 108 (step S137). The quantized transform coefficient is coded by the bit plane coding unit 109 (step S
138), the bit string is formed in the code string forming unit 110 (step S139). Then, the generated encoded data is output from the code output unit 111 to the outside of the device (step S140).

That is, the image coding apparatus according to the present invention inputs the dividing means (tile dividing unit 103) for dividing an image into predetermined regions to generate a small image, and the region information for each divided region. Input means (area information input section 1
02, a region determination unit 105), a generation unit (image reduction unit 104) that generates a reduced image at a predetermined reduction ratio for each divided region, and region information for each divided region. Decomposing means (selector 106, discrete wavelet transforming unit 107) for frequency-converting the small image or reduced image into a predetermined number of subbands,
Coding means (bit plane coding section 109) for coding the decomposed transform coefficient of each sub-band is provided.

Further, the image coding apparatus according to the present invention is characterized in that the means (discrete wavelet transform section 107) for frequency-transforming an image into a predetermined number of sub-bands uses discrete wavelet transform. Furthermore, the image coding apparatus according to the present invention further includes a quantizing unit (transform coefficient quantizing unit 108) that quantizes a predetermined coefficient of the transform coefficients of the decomposed subbands of the divided predetermined area. It is characterized by Furthermore, the image coding apparatus according to the present invention is further characterized by further including adding means (code string forming unit 110) for adding predetermined header information to the encoded bit string.

As described above, in the present embodiment, the input image data is divided into a plurality of tiles Ti, and it is judged in tile units whether or not the tiles are high-resolution encoded areas. Tiles that belong to
For the tiles that do not belong to the high-resolution encoded area, the appropriate reduced tile STi is discrete-wavelet-transformed and encoded. This makes it possible to obtain coded data that includes information for high resolution reproduction for areas that require high resolution images and suitable low resolution restored images for areas that do not require high resolution. It is possible to generate encoded data as various encoded data.

<Second Embodiment> FIG. 14 is a block diagram showing the arrangement of an image coding apparatus according to the second embodiment of the present invention. 14, devices having the same operations and functions as those of the components used in the image encoding device according to the first embodiment shown in FIG. 1 are denoted by the same reference numerals, and the description thereof will be omitted.

The image encoding apparatus in this embodiment also encodes monochrome image data in which the brightness value of each pixel is represented by 8 bits, as in the first embodiment. Further, in the present embodiment, the image data is not limited to 8 bits, but 4 bits, 1
It can also be applied to the case of encoding image data expressing a brightness value with a bit number such as 0 bit or 12 bits. Furthermore, it is also applicable to the case of encoding color image data in which each pixel is expressed by a plurality of color components such as RGB and CMYK, or luminance and chromaticity / color difference components such as YCbCr.
In this case, each component in the color image data may be regarded as monochrome image data. Further, the size of the input image data is set to X pixels × Y pixels as in the first embodiment, and the vertical and horizontal sizes of tiles generated by dividing the image data into a predetermined size (TW × TH). The reduced image is a half of the above. In the present embodiment, X,
Both Y are multiples of 4.

Each section of the image coding apparatus according to this embodiment will be described below with reference to the block diagram shown in FIG. The image input unit 101 is a device for inputting encoded image data, and is connected to a tile division unit 103 that divides the input image data into tiles of a predetermined size. The tile dividing unit 103 is an image reducing unit 104 that generates a predetermined reduced image from the divided tiles.
An area determination unit that determines area information such as whether the tile is a character area or a photograph area, and a selector 106 that determines a required resolution for each tile.

Further, the selector 106 is further connected to the discrete wavelet transform unit 1404 which performs discrete wavelet transform on the image data of the tile determined to be the image reduction unit 104. The discrete wavelet transform unit 1404
Further, it is connected to the bit plane coding unit 109 and outputs the converted image data for coding. The bit plane encoding unit 109 includes a code output unit 11
The code string forming unit 1403 connected to 1 forms a code string based on the least significant bit determined by the least significant BP determining unit 1402.

FIG. 15 is a flow chart for explaining the operation procedure of the image coding apparatus according to the second embodiment shown in FIG. First, as in the first embodiment,
The image data to be encoded by the image encoding apparatus according to the present embodiment is input from the image input unit 101 in raster scan order (step S151). The input image data is divided by the tile division unit 103 into a plurality of tiles having a predetermined width TW and height TH (step S1).
52). The generated tiles T0 to Tn have area determination units 140 connected to the tile division unit 103 in raster scan order.
1, the selector 106, and the image reduction unit 104.

In the area determination unit 1401, the tile division unit 1
The tile Ti generated in 03 is stored in an internal buffer (not shown), and the pixel value in the tile is checked. Then, it is determined whether the tile is a tile including the high-resolution encoded area or a tile not including the high-resolution encoded area, and the determination result F (Ti) is output for each tile.

In determining the tile Ti, edge determination processing is performed for each pixel value Ti (x, y) of the tile Ti, and edge determination information E indicating whether each pixel is an edge portion or not.
(Ti (x, y)) is generated. The value of the generated edge determination information E (Ti (xx, y)) is the pixel T
When i (x, y) is determined to be an edge portion, 1 is given, and when it is determined not to be an edge portion, 0 is given. Then, the above-described edge determination is performed for all the pixels in the tile Ti (step S153a). Then the number of edges in the tile Ti, ie E (Ti
The sum of (x, y)) is calculated (step S153).
b) The value is compared with a predetermined threshold value Th, and when the sum of the number of edges is a value equal to or larger than the threshold value Th, it is determined that the tile Ti is an area including the high resolution coding area. Then, 1 is output as the determination result F (Ti). On the other hand, when the value of the total sum is equal to or less than the threshold value Th, it is determined that the tile Ti does not include the high resolution coding area, and 0 is output as the determination result F (Ti) (step S153c). ).

On the other hand, the image reducing unit 104 generates a reduced image having a half width and a half height from the divided tiles at the same time as, or before and after the above-described area determination (step S154). Further, in the selector 106, similarly to the first embodiment, the tile Ti of interest or the tile Ti of interest is reduced by the image reduction unit 104 based on the determination result F (Ti) of the area determination unit 1401. The generated STi is selected and output to the discrete wavelet transform unit 1404 (step S155).

The discrete wavelet transform unit 1404 outputs the transform coefficient obtained by performing the discrete wavelet transform of the tile Ti or the reduced tile STi selected by the selector 106 to the bit plane coding unit 109 (step S).
156). The process in the discrete wavelet transform unit 1404 is the same as that in the discrete wavelet transform unit 107 in the first embodiment except for the filter used for the transform. In the discrete wavelet transform unit 1404, the one-dimensional discrete wavelet transform for N one-dimensional signals x (n), (n = 0, ..., N−1) is performed by the following equation.

H (n) = x (2n + 1)-floor {(x (2n) + x (2n + 2)) / 2} (5) l (n) = x (2n) + floor {(h (n-1) + h (n) + 2) / 4} (6) where floor {R} is a function that obtains a maximum integer value that does not exceed R, and h (n) is a high-frequency subband. , L (n) represents the coefficient of the low frequency subband. It should be noted that both ends x (n) and (n <0 and n ≧ N) of the one-dimensional signal x (n) required in the calculation of the above equation are set to 1 by a known method.
It is obtained from the values of the dimensional signals x (n) and (0 ≦ n <N).

The bit plane coding unit 109 is further connected to the code string forming unit 1403, and the transform coefficients obtained by the discrete wavelet transform are transmitted to the bit plane coding unit 1.
The encoded data for each tile generated in 09 is output to the code string forming unit 1403 (step S15).
7).

Further, the lowest BP determining unit 1402 includes the lowest BP to be included in the finally generated code string of the bit plane encoded data of each subband S based on the determination result F (Ti) of the area determining unit 1401. Lower bit plane LBP
(S) is determined (step S158). That is, the coded data below the least significant bit plane determined by the least significant BP determination unit 1403 is discarded and is not included in the coded data output to the outside of the device. Note that this determination may be performed at the same time as or before or after the coding of the transform coefficient of the discrete wavelet transform for each tile.

For example, when the least significant bit plane LBP (S) of a certain subband S is 2, the coded data of the two bitplanes of the bitplane number 1 and the bitplane number 0 is transmitted for the subband S. Not done. Thus, discarding N bit-plane coded data sets the coefficient value of the subband to 2
There is the same effect as quantization by the Nth power of. FIG.
FIG. 7 is a diagram showing an example of the least significant bit plane LBP (S) selected when F (Ti) = 0. Also, FIG.
FIG. 7 is a diagram showing an example of the least significant bit plane LBP (S) selected when F (Ti) = 1. Hereinafter, a case of using the least significant bit plane LBP (S) of FIGS. 16 and 17 will be described.

Then, the code string forming unit 1403 stores the coded data for the tile Ti generated by the bit plane coding unit 109 in an internal buffer (not shown).
However, the least significant bit plane LBP of the subband S determined by the least significant BP determination unit 1402 for each tile.
(S) The encoded data below is not stored in the buffer.

In the code string forming unit 1403, after the code strings for all tiles T0 to Tn of the input image data are stored in the internal buffer, the code strings stored in the internal buffer are read in a predetermined order. Then, necessary additional information is inserted, a final code string which is an output of the image coding apparatus is formed, and is output to the code output unit 111 (step S159).

The final code string generated by the code string forming unit 1403 is composed of a header and coded data of each tile of tiles T0 to Tn. The encoded data of each tile includes a tile header storing identification information for identifying the position of the tile in the image data, additional information such as the number of discrete wavelet transforms applied to each tile, and level 0. Two of level 1, or
It is composed of encoded data hierarchically divided into three levels of level 0, level 1 and level 2.

FIG. 18 shows the code string forming section 1403.
Sign of tile STi generated when F (Ti) = 0
It is a schematic diagram which shows the structure of the conversion data. In addition, FIG.
F (Ti) = generated in the code string forming unit 1403
Outline showing the structure of encoded data of tile Ti in the case of 1
It is a figure. As shown in FIG. 18 and FIG.
The encoded data is obtained by encoding the coefficients of the LL subband.
CS (LL, N _BP(LL) -1) to CS (L
L, 0).

Level 1 is LH1, HL1, HH
Code sequence C obtained by encoding the coefficients of each subband of 1
From S (LH1, N _BP (LH1) -1) to CS (LH1,
0) and CS (HL1, N _BP (HL1) -1) to CS
(HL1,0) and CS (HH1, N _BP (HH1)-
1) to CS (HH1,0). Further, level 2 has a code string CS (LH2, NH) obtained by coding the coefficients of the subbands LH2, HL2, and HH2.
_{From BP} (LH2) -1) to CS (LH2,0), CS
From (HL2, N _BP (HL2) -1) to CS (HL2,
0) and CS (HH2, N _BP (HH2) -1) to CS
(HH2,0).

Then, the code output unit 111 outputs the code string generated by the code string forming unit 1403 to the outside of the device (step S160). The code output unit 111 has a first
Similar to the above embodiment, it is realized by a storage device such as a hard disk or a memory, an interface of a network line, or the like.

That is, in the image coding apparatus according to the present invention, the input unit (area determining section 1401) determines, for each pixel in the divided area, whether or not the pixel is an edge pixel, and the inside of the area. A generation unit that calculates the total sum of the number of pixels determined to be edge pixels and a generation unit that compares the calculated number of edge pixels with a predetermined threshold value and generates region information of the region, It is characterized by inputting the region information that has been created.

Further, the image coding apparatus according to the present invention determines the least significant bit of the transform coefficient of the coded subband in each region based on the inputted region information (the least significant BP determining unit). 1402) is further provided, and the assigning means (code string forming unit 1403) uses the bit string of the determined least significant bit or more from the bit string of the coefficient of the sub-band of each region as encoded data. .

As described above, in the second embodiment,
The image data was divided into a plurality of tiles of a predetermined size, and it was determined in tile units whether or not it was a high-resolution encoded area. Then, the tiles that belong to the high-resolution coded area are coded by discrete wavelet transforming the tiles that do not belong to the high-resolution coded area and the reduced tiles that have been reduced at a predetermined reduction rate. As a result, it is possible to generate encoded data that contains information for high-resolution reproduction only in the part that requires high resolution, and that can obtain a suitable low-resolution restored image for parts that do not require high resolution. Becomes

<Third Embodiment> FIG. 20 is a block diagram showing the arrangement of an image encoding apparatus according to the third embodiment of the present invention. Also in FIG. 20, parts common to the parts used in the first and second embodiments are denoted by the same reference numerals, and description thereof will be omitted. 20, the tile division unit 103 includes a discrete wavelet transformation unit 2001.
And the image reduction unit 104. The image reduction unit 104 further includes a discrete wavelet transform unit 200.
2 and the discrete wavelet transform unit 2002 is connected to the transform coefficient quantization unit 108.

Further, the selector 2003 is connected to the area determining section 105, the discrete wavelet transform section 2001 and the transform coefficient quantizing section 108, and based on the area information, the transform coefficient for the tile Ti or the tile STi.
It outputs the quantized transform coefficient for the bit plane coding unit 109.

Also in this embodiment, as in the first and second embodiments, description will be made assuming that the monochrome image data in which the luminance value of each pixel is represented by 8 bits is encoded. However, the present invention is not limited to this, and is also applicable to the case of encoding image data expressing a luminance value with a bit number other than 8 bits such as 4 bits, 10 bits, and 12 bits. It is also applicable when encoding color image data in which each pixel is represented by a plurality of color components such as RGB and CMYK, or luminance and chromaticity / color difference components such as YCbCr. In this case, each component in the color image data may be regarded as monochrome image data.

Further, in the present embodiment, as in the first and second embodiments, the image data to be encoded is encoded in one of two resolutions for each divided tile. After that, the resolution of the image data to be encoded is high
The resolution obtained by halving both in the vertical direction is called low resolution. The size of the encoded image data is fixed, and the size of the input image is X pixels × Y pixels. Further, both X and Y are multiples of 4.

FIG. 21 is a flow chart for explaining the operation procedure of the image coding apparatus having the configuration shown in FIG. First, as in the first embodiment, the image data encoded in the image encoding device according to the present embodiment is
Inputs are made in raster scan order from the image input unit 101 (step S211). The input image data is divided into tiles having a predetermined width TW and height TH in the tile division unit 103, similarly to the above-described embodiment (step S212). Then, the divided tiles from T0 to Tn are processed by the discrete wavelet transform unit 20 in raster scan order.
01 and the image reduction unit 104.

The discrete wavelet transform unit 2001 performs two-dimensional discrete wavelet transform twice while appropriately storing the image data of the tile Ti input from the tile division unit 103 in an internal buffer (not shown). As shown, LL, HL1, LH1, HH1, HL2, LH
2, decomposed into 7 subbands of HH2 (step S
213). After that, the coefficient of each subband is C (S, x,
y). Here, S represents a subband, and LL, H
Any one of L1, LH1, HH1, HL2, LH2, and HH2. Further, (x, y) is each subband S
The coefficient position in the horizontal direction and the vertical direction when the coefficient position in the upper left corner of the inside is (0, 0).

In the image coding apparatus according to this embodiment,
N one-dimensional signals x (n), (n = 0, ..., N-1)
It is assumed that the one-dimensional discrete wavelet transform for is performed by the equations (1) and (2) similarly to the discrete wavelet transform unit 107 in the first embodiment.

On the other hand, the image reduction unit 104 generates a reduced tile STi that is 1/2 the size of the tile Ti output from the tile division unit 103 in both the horizontal and vertical directions (step S214). The image reduction processing in the image reduction unit 104 is the same as that in the first embodiment. In addition, the horizontal x-th pixel value and the vertical y-th pixel value in the reduced tile STi are represented by STi (x, y), (i = 0, ...
, N, 0 ≦ x <TW / 2, 0 ≦ y <TH / 2). Note that STi (0,0) is the pixel value at the upper left corner of the reduced tile STi.

Then, the discrete wavelet transform unit 200
2 is a reduced tile STi input from the image reduction unit 104.
Is appropriately stored in an internal buffer (not shown), the two-dimensional discrete wavelet transform is performed once, and as shown in FIG. 5, it is decomposed into four subbands LL, HL1, LH1, and HH1 (step S215). Hereinafter, the coefficient of each subband is represented as C (S, x, y). Here, S represents a subband, and is any one of LL, HL1, LH1, and HH1. Further, (x, y) represents the coefficient position in the horizontal direction and the vertical direction when the coefficient position at the upper left corner in each subband S is (0, 0).

In this image coding apparatus, the one-dimensional discrete wavelet transform for N one-dimensional signals x (n), (n = 0, ..., N-1) is performed by the discrete wavelet in the second embodiment. Similar to the conversion unit 1404, (5),
It is assumed that this is performed by the equation (6). Then, the LL and HL generated by the discrete wavelet transform unit 2002
The coefficients of the subbands of 1, LH1, and HH1 are quantized by the transform coefficient quantization unit 108 as in the first embodiment (step S216). Incidentally, step S214
The processing from the generation of the reduced image to the quantization in step S216 may be performed simultaneously with or before or after the discrete wavelet transform in step S213.

Next, as in the first embodiment, from the high resolution coded area information that specifies the high resolution coded area input from the area information input unit 102, the area determination unit 105 determines the tile Ti of interest. The determination result F (Ti) is generated (step S217). The high-resolution encoded area information may be input from the area information input unit 102 at the same time as the input of the image data, or before or after the input.
Further, the region determination result F (Ti) may be generated at the same time as or before or after the discrete wavelet transform of each region as long as the tile is divided from the image data.

Next, the selector 2003 determines the tile Ti of interest based on the area determination information F (Ti).
The transform coefficient value generated by the discrete wavelet transform unit 2001 or the quantized coefficient value generated by the transform coefficient quantization unit 108 is selected and output (step S21).
8). That is, the area determination information F (Ti) of the tile Ti
Is 0, the tile Ti selects a quantized coefficient value that does not include a high-resolution encoded area such as a character area,
When F (Ti) is 1, it is determined that the tile Ti includes the high resolution coding area, and the transform coefficient value is output.

The transform coefficient value or the quantized coefficient value selected and output by the selector 2003 is the same as in the first embodiment, that is, the bit plane coding unit 109, the code string forming unit 110, and the code output unit 111. Is encoded and output to the outside (step S219).

That is, the image coding apparatus according to the present invention divides an image into predetermined regions to generate a small image (tile dividing unit 103), and frequency-converts the divided small images into a predetermined number. First decomposing means (discrete wavelet transform section 2001) for decomposing into sub-bands and input means (area information input section 102, area determining section 105) for inputting area information about each divided area, For each region, a generation unit (image reduction unit 104) that generates a reduced image at a predetermined reduction ratio, and frequency conversion of the reduced image of each region based on the region information of each divided region, a predetermined number of sub-images. Second decomposing means (discrete wavelet transform section 2002) for decomposing into bands, and encoding means (bit plane encoding section 1 for encoding decomposing transform coefficients of each subband) Characterized in that it comprises 9) and.

Further, the image coding apparatus according to the present invention is a quantizing means (transform coefficient quantizing unit 108) for quantizing a predetermined coefficient of the transform coefficients of the decomposed subbands of the divided predetermined area. Is further provided.

In the case of the present embodiment, the method of selecting the data obtained by the discrete wavelet transform or the quantization is changed according to the area determination result, and the character area can be reversibly encoded.

<Other Embodiments> The present invention is not limited to the above-described embodiments. For example, in the above-described first to third embodiments, (1), (2)
Although the example of the encoding using the discrete wavelet transform by the equation or the equations (3) and (4) has been shown, the discrete wavelet transform is not limited to the one used in the present embodiment. For example, the type of filter and the adaptation method may be changed. For example, instead of the 9/7 filter or the like, a filter having a long tap number may be used, and the two-dimensional discrete wavelet transform may be repeatedly applied to other than the low frequency subband.

QM-Co is used as a coefficient coding method.
Although the bit plane coding method using der is shown,
The present invention is not limited to the above embodiment. For example,
An arithmetic coding method other than QM-Coder such as MQ-Coder may be applied, or another binary coding method such as MELCODE may be applied. Furthermore, the bit plane may be classified into a plurality of sub-bit planes according to the state of the coefficient near the coefficient of interest, and may be encoded by a plurality of passes. Furthermore, the Golomb code or the like may be applied to entropy-encode the multi-value as it is without dividing the coefficient into binary.

Furthermore, in order to simplify the explanation, in the above-mentioned embodiment, bit-plane coding in sub-band units was explained, but in order to improve random accessibility, each sub-band is further divided into small blocks. do it,
Bit plane coding may be applied in units of this small block. Furthermore, when forming the code string, the codes are arranged so that the image can be restored by gradually increasing the resolution on the receiving side, but the present invention is not limited to this, and codes are arranged in order from the coefficient with the largest value so that the image quality gradually improves. Rows may be formed.

In the above-described embodiment, by inputting the image data in which the character / photo area is mixed in the image, the character area is set as an area requiring high resolution, and the photo area is set as the other area. This area division may be arbitrarily set using other divisions such as characters and photographs.

Even if the present invention is applied to a system composed of a plurality of devices (eg, host computer, interface device, reader, printer, etc.), a device composed of one device (eg, copying machine, facsimile). Device).

Further, an object of the present invention is to supply a recording medium (or storage medium) recording a program code of software for realizing the functions of the above-described embodiment to a system or apparatus, and to supply a computer of the system or apparatus ( Alternatively, the CPU or MPU) reads and executes the program code stored in the recording medium,
It goes without saying that it will be achieved. In this case, the program code itself read from the recording medium realizes the functions of the above-described embodiments, and the recording medium recording the program code constitutes the present invention.
Further, by executing the program code read by the computer, not only the functions of the above-described embodiments are realized, but also based on the instruction of the program code,
An operating system (OS) running on the computer does some or all of the actual processing,
It goes without saying that the processing includes the case where the functions of the above-described embodiments are realized.

Further, after the program code read from the recording medium is written in the memory provided in the function expansion card inserted into the computer or the function expansion unit connected to the computer, based on the instruction of the program code, , The CPU provided in the function expansion card or the function expansion unit performs some or all of the actual processing,
It goes without saying that the processing includes the case where the functions of the above-described embodiments are realized.

When the present invention is applied to the above recording medium, the recording medium stores the program code corresponding to the above-mentioned flowchart.

[0098]

As described above, according to the present invention,
It is possible to efficiently encode a mixed image including regions having different required resolution levels such as a character region and a photo region in a state in which a high-quality image can be restored.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an image encoding device for encoding an image according to a first embodiment of the present invention.

FIG. 2 is a diagram for explaining image data divided into a plurality of tiles.

FIG. 3 is a schematic diagram for explaining image data to be encoded and a high-resolution encoded area according to the present embodiment.

FIG. 4 is a diagram showing an example of a case where two determination results are output for a tile of image data by the method described above.

FIG. 5 is a schematic diagram for explaining a case where a tile STi is decomposed into four subbands.

FIG. 6 is a schematic diagram for explaining a case where tile Ti is decomposed into seven subbands.

FIG. 7 is a schematic diagram for explaining an application example of a two-dimensional discrete wavelet transform to an encoding target image.

FIG. 8 is a diagram showing an example of sub-band division when two-dimensional wavelet transform is repeated twice.

FIG. 9 is a flowchart for explaining an operation procedure of a bit plane encoding unit 109 that encodes a subband S.

10 is a diagram showing the structure of a code string generated by the code string forming unit 110. FIG.

FIG. 11 is a schematic diagram showing a configuration of encoded data of a tile STi when F (Ti) = 0.

FIG. 12 is a schematic diagram showing a configuration of encoded data of tile Ti when F (Ti) = 1.

FIG. 13 is a flowchart for explaining an operation procedure of the image encoding device according to the first embodiment.

FIG. 14 is a block diagram showing a configuration of an image encoding device according to a second embodiment of the present invention.

FIG. 15 is a flowchart for explaining an operation procedure of the image encoding device according to the second embodiment shown in FIG.

FIG. 16 is a diagram showing an example of the least significant bit plane LBP (S) selected when F (Ti) = 0.

FIG. 17 is a diagram showing an example of the least significant bit plane LBP (S) selected when F (Ti) = 1.

FIG. 18 is an F generated in a code string forming unit 1403.
It is a schematic diagram which shows the structure of the encoded data of the tile STi when (Ti) = 0.

FIG. 19 is an F generated in a code string forming unit 1403.
It is a schematic diagram which shows the structure of the encoded data of tile Ti in case of (Ti) = 1.

FIG. 20 is a block diagram showing a configuration of an image encoding device according to a third embodiment of the present invention.

FIG. 21 is a flowchart for explaining an operation procedure of the image encoding device having the configuration shown in FIG.

[Explanation of symbols]

Reference Signs List 101 image input unit 102 region information input unit 103 tile division unit 104 image reduction unit 105, 1401 region determination unit 106, 2003 selectors 107, 1404, 2001, 2002 discrete wavelet transform unit 108 transform coefficient quantization unit 109 bit plane coding unit 110, 1403 code string forming unit 1402 lowest BP determining unit

   ─────────────────────────────────────────────────── ─── Continued front page    F term (reference) 5C059 MA24 MC11 MC38 PP01 PP20                       PP24 PP28 RB02 RB09 SS06                       SS11 TA80 TC34 TD06 TD11                       UA02 UA15                 5C076 AA22 CA02                 5C078 AA04 BA53 CA01 DA01                 5J064 AA02 BA16 BB13 BC14 BC16                       BC22 BC25                 5L096 FA06 FA54 GA51

Claims (16)

[Claims] 1. A dividing unit that divides an image into predetermined regions to generate a small image, an input unit that inputs region information about each divided region, and a predetermined reduction ratio for each divided region. A generation unit that generates a reduced image, a decomposition unit that frequency-converts a small image or a reduced image for each region based on the region information of each divided region, and a decomposition unit that decomposes the image into a predetermined number of subbands. An image coding apparatus, comprising: a coding unit that codes the transform coefficient of each subband. 2. Dividing means for dividing an image into a predetermined area to generate a small image; first dividing means for frequency-converting the divided small image to decompose into a predetermined number of subbands; Input means for inputting area information for each area, generation means for generating a reduced image at a predetermined reduction ratio for each divided area, and reduction for each area based on the area information for each divided area. An image coding apparatus comprising: a second decomposing means for frequency-converting an image to decompose it into a predetermined number of subbands; and an encoding means for encoding the decomposed transform coefficient of each subband. 3. The input unit calculates, for each pixel in the divided area, a determination unit for determining whether or not the pixel is an edge pixel, and a calculation for calculating a sum of the number of pixels determined to be an edge pixel in the area. And a generation unit that compares the calculated number of edge pixels with a predetermined threshold value to generate region information of the region, and inputs the generated region information. The image encoding device according to 1 or 2. 4. The image coding apparatus according to claim 1, wherein the means for frequency-converting the image to decompose it into a predetermined number of subbands uses a discrete wavelet transform. . 5. A quantizing means for quantizing a predetermined coefficient among the transform coefficients of the decomposed subbands for the divided predetermined area is further provided.
The image coding apparatus according to any one of items 1 to 4. 6. The image encoding apparatus according to claim 1, further comprising an attaching unit that attaches predetermined header information to the encoded bit string. 7. A determination means for determining the least significant bit of a transform coefficient of an encoded subband in each area based on the input area information is provided, wherein the assigning means determines the least significant bit of the subband of each area. 7. The image encoding apparatus according to claim 6, wherein encoded data is obtained by using a bit string having a least significant bit determined from the bit string of coefficients. 8. A dividing step of dividing an image into predetermined areas to generate a small image, a generating step of generating a reduced image at a predetermined reduction ratio for each divided area, and a dividing step for each divided area. An acquisition step of acquiring area information, a decomposition step of frequency-converting a small image or reduced image for each area based on the area information of each divided area, and decomposition into a predetermined number of subbands, and decomposition And an encoding step of encoding the transform coefficient of each subband. 9. A dividing step of dividing an image into predetermined areas to generate a small image, an acquisition step of acquiring area information about each of the divided areas, and an area information of each of the divided areas. , Frequency-converting a small image in each region and dividing it into a predetermined number of subbands
And the generation step of generating a reduced image with a predetermined reduction ratio for each divided area, and the reduced image of each area is frequency-converted based on the area information of each divided area, and a predetermined number of An image coding method comprising: a second decomposition step of decomposing into sub-bands; and an encoding step of encoding the decomposed transform coefficients of each sub-band. 10. The determining step of determining whether each pixel in the divided area is an edge pixel, and the calculation step of calculating the sum of the number of pixels determined to be an edge pixel in the area. The image code according to claim 8 or 9, further comprising: a step, and an area generation step of comparing the calculated number of edge pixels with a predetermined threshold value to generate area information of the area. Method. 11. The image according to claim 8, wherein the step of frequency-converting the image to decompose it into a predetermined number of subbands is performed by using a discrete wavelet transform. Encoding method. 12. The method according to claim 8, further comprising a quantization step of quantizing a predetermined coefficient among the transform coefficients of the decomposed subbands of the divided predetermined area. The image coding method according to item 1. 13. The image encoding method according to claim 8, further comprising an attaching step of attaching predetermined header information to the encoded bit string. 14. The method further comprises a determining step of determining the least significant bit of the transform coefficient of the coded subband in each area based on the acquired area information, wherein the assigning step includes the subband of each area. 14. The image coding method according to claim 13, wherein the coded data is formed by using a bit string having a least significant bit determined from the bit string of the coefficient of. 15. A computer program for controlling an image encoding apparatus for encoding an image, comprising a program code of a dividing step for dividing an image into predetermined areas to generate a small image, and each divided For a region, based on the program code of the generation process that generates a reduced image at a predetermined reduction ratio, the program code of the acquisition process that acquires the region information about each divided region, and the region information of each divided region, Program the code of the decomposing process that frequency-converts the small image or reduced image for each region and decomposes it into a predetermined number of subbands, and the program code of the encoding process that encodes the decomposed transform coefficients of each subband. A computer program having. 16. A recording medium on which the computer program according to claim 15 is stored.