JPH0983806A - Image coding/decoding device and image processing unit - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image coder in which a high coding efficiency is obtained in the case of converting multi-value image data into binary image data string for coding. SOLUTION: Multi-value image data are converted into binary data string at a multi-value/binary conversion circuit 101 and the binary data string is binary-coded by a binary data coding/decoding circuit 102 in this image coder. The multi-value/binary conversion circuit 101 selects one binary data string from plural binary data strings so that same data are much consecutive in the adjacent binary data strings with respect to at least one data in the multi- value image data and provides an output of it.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image encoding / decoding apparatus for handling images such as images and texts, and an image processing system for applying them to an image recording system such as a printer or a copying machine or an image transmission system such as a facsimile. Related to the device,
In particular, the present invention relates to an image encoding / decoding device and an image processing device that handle an image as multi-valued image data.

[0002]

2. Description of the Related Art When an image is output by a printer or copier or transmitted by facsimile or the like, an original image is converted into a binary data string and compression-encoded in order to effectively use an image memory and a transmission path. Value image coding techniques have been used for some time. On the other hand, in recent years, output devices such as printers having the ability to express several gradations have been supplied at low cost, and at the same time, original images have been handled as multi-valued and recorded in multi-valued. The demand for coding technology has increased.

In order to meet such a demand, for example, as disclosed in Japanese Patent Application Laid-Open No. 60-100880, a technique for converting multi-valued image data into weighted data of a plurality of bits has been considered. . Specifically, for example, when the multi-valued image data is octal data, as shown in FIG. 49, the multi-valued image data is 7-bit binary data defined in one-to-one correspondence with each data value of the octal data. Convert to a column,
In this method, multi-valued image data is treated as a one-plane high-resolution binary data sequence, and multi-valued image data is encoded by applying a binary image encoding technique to the binary data sequence.

[0004]

In the above-described conventional image coding method, multi-valued image data is represented by (number of pixels per original image) × (multi-valued number of multi-valued image data−1) pixels. Since it is converted into a value data string, the number of pixels of the binary data string, that is, the number of pixels to be coded in the binary coding becomes very large, and the coding efficiency is low. For this reason, when storing the coded data obtained by the binary coding, a large-capacity memory is required, or the number of pieces of document information that can be stored under a limited memory capacity is reduced. When transmitting coded data, there is a problem that the transmission time becomes long.

The present invention has been made in order to solve such problems of the prior art, and converts multivalued image data into a binary image data sequence and encodes the binary image data using a binary encoding technique. It is an object of the present invention to provide an image encoding device capable of obtaining high encoding efficiency in such a case.

Another object of the present invention is to provide an image decoding apparatus capable of decoding original multi-valued image data from data encoded by the above image encoding apparatus.

Further, the present invention provides an image processing apparatus for outputting multi-valued image data via an image memory or a transmission line using these image encoding / decoding devices.

[0008]

In order to solve the above problems, the present invention converts multi-valued image data having three or more data values into a binary data string having binary data values.
In an image coding apparatus including binary conversion means and coding means for coding the binary data string converted by the multi-value / binary conversion means, the multi-value / binary conversion means is multi-valued. For at least one data value of the image data,
It is characterized in that one binary data string is selected from a plurality of binary data strings and outputted so that the same data values are more consecutive in adjacent binary data strings.

That is, in the past, a binary data string was defined in a one-to-one correspondence with all the data values of multi-valued image data.
In the present invention, a plurality of binary data strings are made to correspond to some data values of multi-valued image data, and for example, pixel values of multi-valued image data around the conversion target pixel or images of multi-valued image data Depending on the nature of the above, by selecting and outputting one binary data string from the plurality of binary data strings, the same data value (black pixel data value or white pixel data value) in adjacent binary data strings is selected. Data values of) are more continuous.

More specifically, in the multi-value / binary conversion means, for at least one predetermined data value of the multi-valued image data, (a) the same value is sequentially applied from the left in the one-dimensional main scanning direction. A first one-dimensional binary data string in which a predetermined number of data values are continuously arranged and a second one-dimensional binary data string in which the same number of data values are continuously arranged in a predetermined number from the right, or
(b) The first one-dimensional binary data string in which a predetermined number of the same data values are continuously arranged in order from the top in the one-dimensional sub-scanning direction and the second one in which the same number of data values are continuously arranged in the order from the bottom One-dimensional binary data string, or (c) Predetermined number of identical data values in order from the upper left in the two-dimensional main scanning direction. One binary data string is selected from each of the second two-dimensional binary data strings arranged continuously and is output.

MH (modified Huffman) and MR (Modified RE), which are known as compression encoding techniques for binary image data.
AD), MMR (Modified MR), JBIG (joint bi-le
Binary coding methods such as velimage experts group)
Both are 1 in the main scanning direction and the sub scanning direction of binary image data.
The compression encoding is performed by utilizing the correlation between pixels in the dimensional direction or the two-dimensional main scanning direction. According to the present invention, since the same data value continues more in the binary data string obtained by converting the multi-valued image data, the correlation between the pixels of the binary data string becomes large. The coding efficiency when the coding is performed by the coding method is dramatically improved.

Further, the present invention provides decoding means for decoding a binary data string from encoded data obtained by encoding a binary data string having a binary data value, and decoding by the decoding means. In a decoding apparatus including a binary / multilevel conversion unit for converting a binary data sequence into multilevel image data having three or more data values, the binary / multilevel conversion unit includes a binary / multilevel conversion unit. For at least two of these predetermined,
It is characterized by outputting multi-valued image data having the same data value.

More specifically, in the binary / multivalue conversion means
(a) A first one-dimensional binary data sequence in which a predetermined number of the same data values are continuously arranged in order from the left in the one-dimensional main scanning direction, and a second one in which a predetermined number of the same data values are continuously arranged sequentially from the right. A one-dimensional binary data sequence, or (b) a predetermined number of the same data values in the one-dimensional sub-scanning direction in order from the top and a first one-dimensional binary data sequence in which the same data values are sequentially arranged in the predetermined number from the bottom A second one-dimensional binary data string continuously arranged, or (c) a first two-dimensional binary data string in which a predetermined number of the same data values are successively arranged in order from the upper left in the two-dimensional main scanning direction and the upper right The multi-valued image data having the same data value is output to a second two-dimensional binary data sequence in which the same data values are sequentially arranged in a predetermined number in order from.

According to this image decoding device, the original multi-valued image data can be decoded from the coded data obtained by the above image coding device.

Further, in the image processing apparatus according to the present invention,
After storing the encoded data encoded by the above-described image encoding device in, for example, a memory or transmitting it to a transmission system,
The image decoding apparatus decodes the original multi-valued image data from the encoded data. According to this image processing apparatus, since the encoding efficiency of the encoded data is high, the memory capacity can be reduced or the number of original images that can be stored in the memory can be increased.
When applied to a transmission system, the transmission time can be reduced.

[0016]

DETAILED DESCRIPTION OF THE INVENTION Embodiments of the present invention will be described in detail below with reference to the drawings.

In the block diagrams used in the following embodiments, those having the same name in the block have basically the same function even if the reference number is different, and duplicate explanation will be omitted.

(Embodiment 1-1) FIG. 1 is a block diagram showing an overall configuration when an image processing apparatus according to an embodiment of the present invention is applied to a digital copying machine. This digital copying machine includes a scanner 1010 for reading a document image as multi-valued image data, an image correction device 1011 for performing a correction process on input multi-valued image data, an image processing device 1012, and a device for temporarily storing data encoded by the image processing device 1012. A memory 1014 for temporarily storing data and a printer engine 1015 for recording multi-valued images.

An outline of the processing of this digital copying machine will be described. The normal basic operation of a digital copier is to read a document image to be copied with a scanner 1010,
The image data is converted into multi-valued image data and output by the printer engine 1015 to obtain a copy of the original image.

The operation of the digital copying machine when a function called an electronic sorter is used will be described. The electronic sorter is a function necessary for performing the following purpose, for example.

(1) A plurality of originals, for example, original No.
The required number of copies of the three images 1, 2, and 3 are copied. 1,
2,3, No. Copy originals are output in the order of 1, 2, 3,...

(2) A double-sided original image is reduced to one side to create a copy original.

Next, the operation of the digital copying machine using the electronic sorter will be described with respect to the processing shown in (1). After the original image to be copied is read by the scanner 1010,
Convert to multi-valued image data. After noise is removed from the multivalued image data by the image correction device 1011, the image data is encoded by the image processing device 1012 and stored in the memory 1014. This operation is performed by manuscript image No. The processing is sequentially performed for 1, 2, and 3. Next, the encoded data is read from the memory 1014 and output by the printer engine 1015 decoded by the image processing device 1012. This operation is performed by manuscript image No. 1,
The processing is performed in the order of steps 2 and 3, and is repeated for the required number of copies.

Next, the image correction device 1011 will be described. FIG. 2 is a block diagram of the image correction device 1011. The input image data is first processed by a threshold processing circuit 104.
Is thresholded. When 0 in the octal image data is white and 7 is black, output image data obtained by converting the input image data as shown in the correspondence table shown in FIG. 3 is output.

The image data subjected to the threshold processing is input to an intermediate value conversion circuit 105, and the intermediate value (2 or more and 5 or less) surrounded by black pixels is converted into black pixels according to the distribution state of peripheral pixels as shown in FIG. Is done. From the converted image data, the arc point of the intermediate value pixel is removed by the arc point removal circuit 106. Specifically, when the pixel to be processed has an intermediate value, FIG.
In the case where the intermediate value pixels are not continuous with the target pixel sandwiched in either of the directions 1 and 3 in the directions shown in FIG. 3A, the target pixel is replaced with a white pixel. Further, a process is performed in which an intermediate value adjacent to the black pixel 7 is left. FIG. 5B shows an example of the removal of an arc standing point.
(C) and (d). In FIG. 5B, the target pixel 注目 has an intermediate value of (2 → 4 → 5) in the direction of 1 so that the target pixel value 4 is left. However, in FIG. Since the direction (0 → 3 → 0) is not continuous in the direction of 3 (5 → 3 → 0), the target pixel is a white pixel. FIG.
In (d), the direction of one of the target pixel (5 → 4 → 0) is also 3
The intermediate value is not continuous in the direction (7 → 4 → 0). However, since the black pixels are adjacent in the direction of 3, the target pixel value 4 is left.

According to the image correcting apparatus 1011, noise and blur due to the contamination of the original are corrected to visually improve the image, and the arc standing point is removed. It is possible to obtain multi-valued image data that can increase the conversion efficiency.

Next, the image processing apparatus 10 to which the present invention is applied.
12 will be described in detail. FIGS. 6 (0) to 6 (7) show various binary patterns when the multi-level image data from the image correction device 1011 is converted into a binary data sequence by the multi-level / binary conversion circuit 101. is there. Multi-value / binary conversion circuit 101
In the case where the data values of the multivalued image data are 0 and 7, FIG.
(0) and (7) are selected, and in the case of 2 to 5, (2L) and (2) in FIG.
One of the patterns (L pattern, R pattern) of R) to (5L) and (5R) is selected.

The basic condition for selecting one of the L and R patterns is that the L pattern is selected when the target pixel to the left is 7 and the R pattern is selected when the right pixel is 7 to the right. . With this as a basic rule, one of the L and R patterns is selected under the conditions shown in FIG. That is, reference is made sequentially from the one pattern having the highest priority, and the L and R of the matched reference pattern are selected. If the original image is a character / line image, the intermediate value is likely to appear as an edge adjacent to the black pixel. Therefore, the intermediate value is converted into a binary data sequence using the pattern selection condition in the present embodiment, so that the black value in the binary image data sequence is obtained. Alternatively, since the same white data value continues, the coding efficiency is improved.

FIG. 8 shows an example in which the multivalued / binary conversion circuit 101 converts multivalued image data into a binary data string. A binary data string corresponding to the data value of one pixel of the multi-valued image data is a matrix drawn by a solid line. In the present embodiment, since the 8-level multi-valued image data is converted into a binary data sequence having a size of 7 × 1 pixels, the 8-level 8 × 8 pixels are converted into a binary 56 ×
It has been converted to eight pixels.

The binary signal generated by the multi-value / binary conversion circuit 101
The value data sequence is encoded by the binary data sequence encoding / decoding circuit 102 and stored in the memory 1014. Memory 10
Encoded data 1013 is read from 14 as needed, and decoded by the binary data string encoding / decoding circuit 102. As a method of encoding / decoding a binary data string,
For example, it is preferable to use a binary image encoding method with a high compression ratio such as MMR or JBIG.

The decoded binary data string is multivalued by the binary / multivalue conversion circuit 103. Binary / multi-value conversion circuit 1
In FIG. 03, as shown in FIG.
The binary image is divided into blocks of, for example, 7 × 1 pixel size in a unit (matrix written by a solid line) processed in step (1), and the pattern in each block is corresponded to the table of FIG. .

The multi-valued image data is sent to the printer engine 101.
5 and recorded as an image. In the above-described embodiment, the encoded data 1013 is stored in the memory 1014. However, any medium that stores data electronically, such as a hard disk or a magneto-optical disk, may be used. Is also good.

As described above, according to this embodiment, when the input image is a character / line drawing, the multi-value / binary conversion circuit 101 converts the multi-valued image data of the character or line drawing to the same data value. Since the data is converted into a continuous binary data sequence, the coding efficiency of the binary coding / decoding circuit 102 can be improved. Specifically, CCITT No. 7
In the case where 8-value image data of (size 4672 × 3304) is multi-valued / binary-converted according to the present embodiment and then binary-coded by the MMR method, the generated code amount is reduced by about 30% as compared with the related art. Can be reduced. Therefore, the memory 10
If the capacities of the fourteen are the same, it is possible to store more document image information than in the prior art.

(Embodiment 1-2) FIG.
-1 is a block diagram showing an embodiment in which a binary data string encoding circuit 112 and a binary data string decoding circuit 1 are modified.
13 is an example where they are connected by a transmission line 1022.

(Embodiment 1-3) FIG. 11 shows the first embodiment.
1 is a block diagram showing a modified example of −1, which is an example of a case where a printer engine 1035 converts multi-valued image data into a pulse width modulation signal in the main scanning direction and outputs it.
In this case, as shown in FIG.
In 2, the binary data string can be directly output to the printer engine 1035 without using the binary / multi-value conversion circuit. Further, in the embodiment 1-1, an example in which one dimension is developed in the main scanning direction is shown, but it is also possible to develop it in the sub scanning direction.

(Embodiment 1-4) As an embodiment obtained by modifying Embodiment 1, as shown in the correspondence table of FIG.
When the number of output multi-values is reduced with respect to the number of input multi-values in 04 (in FIG. 3, the number of output multi-values is reduced to 6 with respect to the number of input multi-values), the output multi-values The binary pattern shown in FIG. 12 may be used by using a binary pattern size to be converted, where the number is a multivalued number of multivalued image data.

That is, in FIG. 6, the input multi-valued number is 8 and the pattern size is 7 pixels in binary, whereas in FIG. 12, the output multi-valued number is 6 and the pattern size is 5 pixels in binary. . According to the present embodiment, the number of pixels to be encoded in the binary data string is reduced, and thus the encoding time is reduced.

(Embodiment 1-5) FIG. 13 shows Embodiment 1.
FIG. 11 is a block diagram showing an embodiment in which the multi-value / multi-value conversion circuits 141 and 142 are used in the first stage and the last stage in the image processing device 1042. The multi-value / multi-value conversion circuit 141 performs multi-value (1) → multi-value (2) conversion on input multi-value image data as shown in FIG. The multivalued image data from the value / multivalued conversion circuit 145 is subjected to multivalued (2) → multivalued (1) conversion, which is the reverse of the multivalue / multivalued conversion circuit 141.

By performing such multi-value / multi-value conversion,
When the background and characters are expressed in halftone as in the document image shown in FIG. 15 (middle value: background 1, character 6), multi-value /
In the binary data string converted by the binary conversion circuit 143, solid white or black data tends to appear, so that the coding efficiency can be improved.

(Embodiment 2-1) The block diagram according to the present embodiment is the same as that of the first embodiment, and the point that the multi-level / binary conversion circuit 101 converts the data into two-dimensional binary data strings instead of one-dimensional data. Is different from the first embodiment.

There are 2 × 4 and 3 × 3 two-dimensional matrix sizes which can represent 8-valued multivalued numbers. If the number of black pixels in a matrix is associated with a multivalued value, the number of matrices + 1 or more can be represented. The number of states is larger than the number of values (8 values). Therefore, the matrix is filled with black pixels according to the multi-valued number.
6 (a) and FIG. 16 (b) in which the partial correspondence is changed. If the correspondence of the number of black pixels to the number of multivalues is uniquely determined, a correspondence in which the number of black values and the number of black pixels do not match as shown in FIG. Hereinafter, the present embodiment will be described with reference to FIG. 16B showing the correspondence and the matrix patterns of FIGS. 17A and 17B having a matrix size of 3 × 3.

FIG. 18 shows an example of converting multivalued image data into a two-dimensional binary data sequence. FIG. 18 (b) (size 15 × 15) in which the multi-valued image data (size 5 × 5) of FIG. 18 (a) is replaced with the pattern of FIG. 17 (a).
And the encoding is performed by the binary encoding circuit as in the first embodiment. Further, since it is only necessary to satisfy the correspondence table of FIG. 16B, as shown in FIG.
The binary coding may be performed by converting the pattern to the pattern of FIG. 17B for each pixel, or by replacing the pattern with reference to the values of surrounding pixels as in the first embodiment. May go.

According to the present embodiment, the multi-valued image data is
By converting to a two-dimensional binary data string, the same data value of the binary data string continues in an image with continuous gradation, so that the coding efficiency is improved.

(Embodiment 2-2) FIG.
1 is a diagram showing a modified example of −1, in which an image processing device 2001 includes a multi-value / binary conversion circuit 201 and a division circuit 20.
2, a memory 203, 205, a binary data string encoding / decoding circuit 204, a synthesizing circuit 206, and a multi-level / binary conversion circuit 207.

The operation of this embodiment will be described with reference to FIG. FIG. 18B shows that a binary data string (3 × 3) corresponding to one pixel of the multi-valued image data is divided in units of horizontal lines of each matrix as shown in FIG. The binary data string is divided for each line, and
The image is divided into three 15 × 5 images composed of Lines A to C in (e).

The operation of the image processing apparatus 2001 is performed in the following procedure.

[Step 1] Multi-value / binary conversion circuit 201
Then, the multi-valued image data of FIG.
Convert to a value data string.

[Step 2] The binary data string is divided by the dividing circuit 2
In step 02, the image is divided into line A, B, and C image units of FIG. 18 e) and stored in the memory 203.

[Step 3] A binary data sequence is read out from the memory 203 in units of Line A, B, and C images, and is encoded by the binary data sequence encoding / decoding circuit 204.

[Step 4] Store the encoded data in the memory 20
10 is stored.

[Step 5] If necessary, the memory 201
Read the encoded data from 0.

[Step 6] The binary data string is decoded by the encoding / decoding circuit 204 and stored in the memory 205.

[Step 7] The combining circuit 206 reads out the binary data strings of Lines A, B, and C corresponding to one multi-valued pixel from the memory 205 and combines them to generate one binary data string. The binary / multi-level conversion circuit 207 converts the data into multi-level data.

According to the present embodiment, the same data value of the binary data string continues mainly in the sub-scanning direction.
The coding efficiency is improved.

(Embodiment 2-3) FIG.
FIG. 11 is a block diagram illustrating an embodiment in which −1 is modified. An image processing apparatus 2002 includes a multi-level / binary conversion circuit 210, a division circuit 211, and a binary data string encoding / decoding circuit 212−.
1, 212-2, 212-3, synthesis circuits 213 and 2
It comprises a value / multi-value conversion circuit 214.

The operation of the image processing apparatus 2002 is performed according to the following procedure.

[Step 1] Multi-value / binary conversion circuit 210
Then, the multi-valued image data of FIG.
Convert to a value data string.

[Step 2] The binary data string is divided by the dividing circuit 2
At 11, the image data is divided into image units of Lines A, B, and C shown in FIG.
The encoding is performed by 12-1, 212-2, and 212-3.

[Step 3] Store the encoded data in the memory 20
Store in 12.

[Step 4] If necessary, the memory 201
2 to read the encoded data.

[Step 5] Encoding / Decoding Circuit 212
-1, 212-2 and 212-3.

[Step 6] The combining circuit 213 extracts and combines the binary data strings of Line A, B, and C corresponding to one multi-valued pixel, generates one binary data string, and outputs the binary / multiple values. The value is converted into a multi-value by a value conversion circuit 214.

According to the present embodiment, A4 / 400 dpi
The memory 203, 205 (3450 × 4680) in FIG.
(× 9 pixels × 2 planes = 34 MB) becomes unnecessary, so that the overall memory capacity is reduced and the encoding speed is improved.

The memories 203 and 20 shown in FIG.
5, 2010 and the method of storing data in the memory 2012 in FIG. 20 may be such that the memory is divided into three as shown in FIG. 21 and a binary data string or a code word is stored from the beginning of each area.

(Embodiment 2-4) As an embodiment obtained by modifying the embodiment 2-1, the correspondence diagram of FIG.
In some cases, a binary data string is formed using the pattern of FIG. The multi-valued image data of FIG.
FIGS. 18D and 18G show a pattern formed by using the pattern shown in FIG. 18C and divided for each line. In this case, Li
Since only one pixel out of three pixels is significant in neC, in FIG. 18D and FIG.
The value is composed of one pixel. So a 15x5 image 2
The image is divided into one image and one 5 × 5 image.

According to this embodiment, since the redundancy of the pixel to be coded can be reduced, the coding efficiency is improved and the coding speed is also improved.

(Embodiment 3-1) FIG. 22 is a block diagram according to the present embodiment. A scanner 3010 for reading a document image as multi-valued image data, an image correction device 3011 for processing input multi-valued image data, It comprises a circuit 3012, an image processing device 3014, a memory 3015 for temporarily storing data encoded by image processing, and a printer engine 3016 for recording multi-valued image data.

As shown in FIG. 23, the identification circuit 3012 identifies an image as a line / character image or an image image in units of one pixel or a plurality of pixels, and outputs an image identification signal 3013 (character:
0, image: 1) is output.

The image correction device 3011 includes an identification circuit 301
The above-described threshold processing, intermediate value conversion processing, and arc standing point removal are performed on the image identification signal 3013 from 2 which is 0, that is, a pixel indicating a character.

The image processing device 3014 comprises a multi-level / binary conversion circuit 301, a binary data string encoding / decoding circuit 302, and a binary / multi-level conversion circuit 303. Multivalue / 2
The value conversion circuit 301 includes the multivalued image data and the image identification signal 30.
13, different patterns are converted between the character and the image as shown in FIG. That is, as shown in FIG. 24A, the pattern is converted so that the character image is linked to black (7), and the multi-valued pixel 4 pixels are converted to the image image as shown in FIG. As one unit,
As shown by (*) in FIG. 24B, an arrangement is performed in which the increasing direction of the binary black pixels is alternately changed.

The 2 value converted by the multi-value / binary conversion circuit 301
The value data sequence is encoded by the binary data sequence encoding / decoding circuit 302 and stored in the memory 3015. Memory 30
The encoded data read out as necessary from 15 is 2
The value data string is decoded by the encoding / decoding circuit 302. If the decoded binary data string satisfies the correspondence table between the multi-valued number and the number of binary black pixels in the conversion by the multi-valued / binary conversion circuit 301, the binary / multi-valued data is not distinguished between characters and images. Conversion circuit 30
In step 3, the signal is converted into a multi-level signal and output from the printer engine 3016.

As described above, according to the present embodiment, the coding efficiency is improved by performing the conversion so as to increase the black / white run length of the binary data string.

(Embodiment 3-2) FIG.
-1 is an embodiment in which the image processing device 3023 is modified.
Denotes a dividing circuit 311, a multi-value / binary conversion circuit 312-1, 3
12-2, Binary Data String Encoding / Decoding Circuit 313-
1, 313-2, multi-value / binary conversion circuit 314-1, 31
4-2 and the combining circuit 315.

The operation of the image processing apparatus 3023 will be described below with reference to FIGS. 7, 25, 26, and 28.

[Step 1] The dividing circuit 311 is used in accordance with the identification signal from the identifying circuit 3022, and FIG.
As shown in, the identification information is inserted into the image region in the character / line drawing and the character / line drawing region in the image image so as to form one image each with the character / line drawing and the image, and divided into two images .

[Step 2] The character / line drawing is converted into a one-dimensional binary bit map of FIG. 26 by a multi-value / binary conversion circuit 312-1, and the image is converted by a multi-value / binary conversion circuit 312-2 of FIG. Is converted to a two-dimensional binary bitmap. When the multi-value of the character / line image is an intermediate value, a conversion pattern is selected according to the values of the peripheral pixels. For example, based on the selection condition of FIG. 7, a one-dimensional pattern in which black pixels are packed from the left or a pattern in which black pixels are packed from the right is selected.

[Step 3] The binary data strings of the characters, line drawings, and images are encoded by the binary data string encoding / decoding circuits 313-1 and 313-2 and stored in the memory 3024.

[Step 4] If necessary, the memory 302
4 to read the encoded data,
Decoding is performed by the decoding circuits 313-1 and 313-2.

[Step 5] The decoded binary data string is converted into multivalued and identification information by the multivalued / binary conversion circuits 314-1 and 314-2.

[Step 6] The combining circuit 315 combines the two pieces of multi-valued image data into one piece of multi-valued image data based on the identification information, and outputs the combined data to the printer engine 3025.

According to the present embodiment, it is possible to perform encoding using the correspondence table of the multi-valued number and the binary black pixel number which are different between the character / line drawing and the image, and further, the binary data string corresponding to the property of each image is generated. Since it can be generated, the coding efficiency is improved.

(Embodiment 3-3) FIG.
The image processing apparatus 3033 includes a dividing circuit 321, a multi-level / binary conversion circuit 322, a binary data string encoding / decoding circuit 323, and a multi-level / binary conversion. It comprises a circuit 324, a binary data string encoding / decoding circuit 325 and a combining circuit 326.

The operation of the image processing apparatus 3033 will now be described with reference to FIG.
6, and will be described with reference to FIGS.

[Step 1] In accordance with the identification signal from the identification circuit 3022, the division circuit 321 is used as shown in FIG.
As shown in (1), the character / line drawing generates one image in which the identification pattern is embedded in the image area, and the image is divided so as to generate an image in which a blank is inserted in the character / line drawing area.

[Step 2] The character / line drawing is converted into the bit map shown in FIG. 26 by the multi-value / binary conversion circuit 322, and is encoded by the binary data string encoding / decoding circuit 323, and is stored in the memory 3
034. Images are multi-valued image data encoding
The data is encoded by the decoding circuit 325 and stored in the memory 3034. When the multi-value of the character / line image is an intermediate value, a conversion pattern is selected according to the values of the peripheral pixels.

[Step 3] Memory 303 as required
4, the encoded data is read out, and the codes of the binary bitmap and the multilevel image data are decoded by the binary data string encoding / decoding circuit 323 and the multilevel image data encoding / decoding circuit 325, respectively.

[Step 4] The decoded binary data string is converted into multivalued and identification information by a binary / multivalued conversion circuit.

[Step 5] The two multi-valued image data are combined into one multi-valued image data by the combining circuit 326 based on the identification information, and then output to the printer engine 3035.

In the embodiments 3-2 and 3-3, as the format of the decoded image used by the image processing devices 3023 and 3033, as shown in FIG. 28D, a format in which a blank is inserted in the image area of the character image and an identification pattern is inserted in the character area of the image image as shown in FIG. And a format (not shown) in which the image image is composed only of the image area, and an identification pattern is inserted in the character area of the image image, and the character image is composed only of the character area (not shown).
And so on.

Further, as an example of the configuration of the identification pattern,
As shown in the correspondence table shown in FIG. 26, not only the correspondence between the multi-valued number and the binary bitmap black pixel number but also the correspondence table with the multi-valued number and the binary bitmap black pixel pattern can be created. is there. The one-dimensional identification pattern in FIG. 26 is five in terms of the number of black pixels, and corresponds to the multi-valued number 5. However, the one-dimensional identification pattern is listed in the correspondence table as a special pattern other than the pattern converted by the multi-valued number 5. Conversion from to multi-value is also used as an identification pattern by following this correspondence table. Further, it is also possible to assign the number of black pixels which is not in the multi-valued number like the two-dimensional identification pattern to the identification code.

According to the present embodiment, it is possible to use different encoding methods depending on the characteristics of images such as characters, line drawings, images, and the like, so that encoding efficiency is improved. Further, since it is not necessary to manage the identification information separately from the image data, data management becomes easy.

(Embodiment 3-4) FIG.
FIG. 13 is a block diagram illustrating an embodiment in which −1 is modified. An image processing device 3043 includes a multi-value / binary conversion circuit 331, a binary data string encoding / decoding circuit 332, and a binary / multi-value conversion circuit 333. Is done.

Hereinafter, the operation of the image processing apparatus 3043 will be described with reference to FIG.
This will be described with reference to FIGS. As for the difference in image quality between text and line drawings and images, high resolution images are required for text and line drawings. As for the differences in image quality, text and line drawings require high resolution images, and images have more gradation than resolution. Image is required. Therefore, as shown in FIG.
Handles data formats with different resolutions and gradations for line drawings and images. As shown in FIG. 31A, an input multi-valued image is input at a resolution of 400 dpi, a character / line image is converted to 400 dpi, and an image is converted to 200 dpi, and output by a printer engine 3055. In the present embodiment, as shown in FIG. 32, the number of gradations of a character / line drawing is 8 with 7 pixels, and the image is 7 pixels × 2 × 2 Line =
28 pixels have 29 gradations.

[Step 1] For the pixels identified by the identification circuit 3042 as characters and line drawings in matrix units, the pattern shown in FIG. 32A) is selected according to the distribution state of the peripheral pixels.
Convert to a binary data string.

[Step 2] Similarly, for the pixel selected as the image, the average pixel value in the matrix is obtained, the average value is converted into 29 values, the pattern shown in FIG. 32 (b) is selected, and converted into a binary data string. I do.

[Step 3] From character / line drawing to image,
The matrix that changes from an image to a line drawing is shown in FIG.
(B) The identification pattern of (a) is embedded, and the multivalued image data of the change point is stored in the memory 3044.

[Step 4] A binary data string composed of a character / line drawing, an image and an identification pattern is encoded by the binary data string encoding / decoding circuit 332, and the encoded data is stored in the memory 3044. The storage method in the memory 3044 may be performed, for example, by dividing the storage areas of the encoded data and the multi-valued image data as shown in FIG.

[Step 5] Memory 304 if necessary
4 to read the encoded data,
The decoding is performed by the decoding circuit 332.

[Step 6] The decoded binary data sequence is converted by the binary / multi-level conversion circuit 333 into multi-level and identification information.

[Step 7] Based on the identification information, the character / line image has a 400 dpi / 8 value, and the image has a 200 dpi / value.
Each of the 29 values is output by the printer engine 3044. With respect to the matrix in which the identification information is embedded, the multivalued image data is read from the memory 3044 and replaced with the multivalued image data.

According to the present embodiment, it is possible to encode and record a character / line image and an image at different resolutions / gradations and to improve the image quality of an output image. Further, as shown in FIG. 31B, data in formats having different resolutions can be collectively handled as one piece of data, which facilitates data management.

(Embodiment 3-5) FIG. 30 is a block diagram according to the present embodiment. In the image processing apparatus 3053, the printer engine 3055 converts multivalued image data into a pulse width modulation signal in the main scanning direction. When outputting by multi-value / 2
A value conversion circuit 341 and a binary data string encoding / decoding circuit 342 are used.

The operation of the image processing device 3053 will be described below.

[Step 1] The identification circuit 3052
For a pixel identified as a character / line image in a 2 × 2 matrix unit, the pattern shown in FIG. 31A is selected according to the distribution state of peripheral pixels, and converted into a binary data string.

[Step 2] Similarly, for the pixel selected as the image, the average pixel value in the matrix is obtained, the average value is converted into 29 values, and the pattern shown in FIG. 31B is selected. Convert to

[Step 3] The binary data string is encoded by the binary data string encoding / decoding circuit 342, and the encoded data is stored in the memory 3054.

[Step 4] Memory 305 as required
4 to read the encoded data,
Decoding is performed by a decoding circuit.

[Step 5] The decoded binary data string is output by the printer engine 3055.

According to the present embodiment, it is possible to encode and record a character / line image and an image at different resolutions / tones with different resolutions, so that the image quality of an output image is improved, and management of identification data and the like are labor-saving. Therefore, the device configuration can be simplified. In this embodiment, the binary data sequence is developed in the main scanning direction, but may be developed in the sub-scanning direction. Further, a configuration in which the identification pattern is used for both characters and images, but the change point is used as a change start point and an end point and the identification pattern is used only for characters or images is also possible.

FIG. 34 shows a method of forming the identification pattern.
It is also possible to adopt a configuration in which an identification code is used for each multi-valued image data as in the pattern shown in FIG. In the configuration of FIG. 34, the identification code is used only for the image,
The identification code can be determined only by determining the upper 14 matrices in the × 2 matrix unit. Further, when images are continuous, the identification code 1 is used. When only one image area is provided, the identification code 2 is used to encode the character / image identification result even if the image areas do not appear continuously. it can. In addition, when a multi-value number is described in addition to the identification code, it can be dealt with by giving a special pattern to 0 and 28 as shown in FIG. 34. However, 0 and 28 are solid white and black. Therefore, a configuration in which the identification code is not provided at 0 or 28 by moving the identification code position to an image area having a value other than 0 or 28 before and after it as a character can be adopted.

In this way, since the identification code and the multi-valued image data of the area in which the identification code is embedded can be encoded together, the data management becomes easy and the device cost can be reduced, and the identification code area can be reduced. Since multi-valued image data of can be encoded, the encoding efficiency is improved.

Further, if the identification pattern is constituted by the multivalued number of each change point at the change point, it is not necessary to separately manage the multivalued image data of the change point, and the apparatus is simplified. In the case of forming a change point identification pattern, in the example of FIG. 25, 0s and 7s are all white or black, and it is not possible to distinguish normal multi-valued image data from identification data. Therefore, when a binary pattern is uniquely determined for a multi-valued number as described above, a pattern not used in other multi-valued image data or identification data is assigned or a change point is set to 0 or What is necessary is just to shift to the point where a multivalued number other than 7 appears. If the change point is 0 or 7, even if the change point is moved, the image is solid in black or white, so that the effect on the image quality and the compression ratio is small.

(Embodiment 3-6) FIG.
FIG. 13 is a block diagram according to the present embodiment in which −1 is modified. The image processing apparatus 3063 includes a multi-level / binary conversion circuit 351, a binary data string encoding / decoding circuit 352, a binary / multi-level conversion circuit 353, Pixel counting circuits 354-1 and 354-
It consists of two.

Hereinafter, the operation of the image processing apparatus 3063 will be described with reference to FIG.
This will be described with reference to FIGS. Now, when the target image is an image having a white and gray background as shown in FIG. 36 (a), if the gray multivalued number is 3, this is converted into the binary pattern of FIG. In this case, a large number of black and white discontinuous portions occur in the gray background portion, and the coding efficiency decreases. Therefore, the following processing is performed on the area of the image determined to be a character / line drawing by the identification circuit 3062.

[Step 1] Pixel Count Circuit 354-
At 1, the histogram for the multi-valued number of the image is sequentially generated as shown in FIG.

[Step 2] When a certain amount of histogram is generated, for example, when 10% of the processing of the entire image is completed, the multi-value / binary conversion circuit 351 refers to the histogram output from the pixel count circuit 351. To switch the background part. In this case, the background default is 0
(White) is set.

[Step 3] Specifically, FIG.
When the background is switched when the multi-valued number shown in FIG. 37B is input, the histogram shown in FIG.
The multivalued numbers 0 and 3 determined as the background using the pattern
Data is replaced with the pattern of 0 in FIG.
Other multi-valued image data is selected from the binary pattern of FIG. 32A except that the pattern of FIG. 32A is used as the background data and the pattern of FIG. 37C is used when the background is switched. .

[Step 4] The converted binary data sequence is encoded by the binary data sequence encoding / decoding circuit 352.

[Step 5] Store the encoded data in the memory 30
Store in 64.

[Step 6] The memory 306 if necessary
4 to read the encoded data.

[Step 7] The binary data string encoding / decoding circuit 352 decodes the data.

[Step 8] Binary / multi-value conversion circuit 353
Thus, the pixel count circuit 354-2 is converted while converting it into multi-valued data.
Generates a histogram.

[Step 9] When the same number of pixels as in the multi-value / binary conversion is processed, referring to the histogram output from the pixel count circuit 354-2, as shown in FIG. Each time the pattern of c) appears, the output multi-value number is switched.

According to the present embodiment, an image having a halftone background can be encoded with little reduction in encoding efficiency.

(Embodiment 3-7) FIG. 38 shows Embodiment 3
FIG. 11 is a block diagram according to the present embodiment obtained by modifying -11,
The image processing device 3073 includes a multi-value / binary conversion circuit 361,
Value data string encoding / decoding circuit 362, binary / multi-value conversion circuit 363, pixel count circuits 364-1, 364-2
And memories 365-1 and 365-2.

Hereinafter, the operation of the image processing apparatus 3053 will be described with reference to FIG.
This will be described with reference to FIGS. 2, 36, 37, and 39.
Now, when the target image is an image having a white background and a plurality of grays as shown in FIG. 36B, if the gray multi-value numbers are 1 and 4, the binary pattern shown in FIG. In the case of conversion, many discontinuous portions having a gray background portion in black and white are generated, and coding efficiency is reduced. Therefore, the following processing is performed on the area of the image determined to be a character / line image by the identification circuit 3072.

[Step 1] Pixel Count Circuit 364-
1, a histogram for the multi-valued number of the entire image (global) as shown in FIG. 39A and a local region of the image as shown in FIG. 39B (for example, surrounded by a dotted line in FIG. 36B). Histograms are sequentially generated.
In order to generate a histogram of the local area, a range surrounded by a dotted line is held in the memory 365-1.

[Step 2] At the time when the histogram is generated to some extent (for example, when the processing of 10% of the entire image is completed), the multi-value / binary conversion circuit 361 is the output of the pixel count circuit 364-1. The background part is switched with reference to the global histogram and the local histogram. In this case, 0 (white) is set as the default of the background. At the time of switching, as shown in FIG.
The pattern of (c) is used. FIG. 32A shows the background data.
37 when the background is switched using the 0 pattern of FIG.
Other than using the pattern of (c), other multi-valued image data is selected from the binary pattern of FIG.

[Step 3] The converted binary data sequence is encoded by the binary data sequence encoding / decoding circuit 362.

[Step 4] The encoded data is stored in the memory 30.
74.

[Step 5] Memory 307 if necessary
4 to read the encoded data.

[Step 6] The binary data string encoding / decoding circuit 362 decodes the data.

[Step 7] Binary / multi-value conversion circuit 363
, The pixel count circuit 364-2 sequentially generates a global histogram and a local histogram. In order to generate a histogram of the local area, a range surrounded by a dotted line is stored in the memory 365-2.

[Step 8] When the same number of pixels as in the multi-value / binary conversion is processed, the global histogram and the local histogram output from the pixel count circuit 365-2 are referred to and FIG. As shown in b), the output multi-value number is switched every time the pattern of FIG. 37C appears.

Next, a method of switching the background portion using a global histogram and a local histogram will be described with reference to FIGS.
7 and FIG. 39.

[Step 1] The global histogram and the local histogram at the time of processing the line A) in FIG. 36 (b) are shown in FIGS. 39 (a) and 39 (c). And 4.

[Step 2] From the local histogram, 0, 3 and 4 are listed as candidates for the background. Since 3 listed in the local histogram is not listed as a candidate in the global histogram, 0 and 4 are finally selected as the background in the line of A) in FIG.

[Step 3] When the background data 0 and 4 are switched as shown in FIG.
Insert the switching data of (c).

[Step 4] The global histogram and the local histogram at the time of processing the line B) in FIG. 36 (b) are shown in FIGS. 39 (b) and (d). And 1 and 4.

[Step 5] There are 1 and 4 as background candidates from the local histogram. Since 0 described in the global histogram is not a candidate in the local histogram, 1 and 4 are finally selected as the background in the line B) of FIG. 36B.

[Step 6] When the background data 1 and 4 are switched as shown in FIG.
Insert the switching data of (c). In the line B), 0 deviates from the background. In such a case, the pattern of the multivalued number used as the background, 1 in the present embodiment.
Use the pattern

According to the present embodiment, encoding can be performed efficiently even if the background is composed of a plurality of colors of three or more.

In the embodiments 3-6 and 3-7, only the area determined to be a character or a line drawing is used. However, in another embodiment, it can be used regardless of the image area or regardless of the identification result. However, it is also possible to adopt a configuration in which a histogram is created independently of a character / line drawing and an image.

Further, to obtain a local histogram, only the histogram of the previous line is used, and the memory 365-1,
It is also possible to adopt a configuration that does not use 365-2.

(Embodiment 4) FIG. 40 shows a block diagram according to this embodiment in FIG. As shown in the figure, the size N × M of the input multi-valued image data is set to N ′ × M ′ (N ′ ≦
N, M ′ ≦ M), a reduction processing device 401 for generating the reduced data and difference information, a multi-valued image data coding / decoding circuit 405 for coding / decoding the reduced data, and setting the difference information to 2 Multi-value / binary conversion circuit 4 for converting to a value data string
02, a binary data string encoding / decoding circuit 403 for encoding / decoding a binary data string, a binary / multi-value conversion circuit 404 for converting a binary data string into multi-valued difference information, A memory 406 for temporarily storing reduced data and difference information,
And an enlargement / reconstruction device 407 for restoring an image based on the decoded reduction information and difference information.

The operation of this embodiment will be described below with reference to FIGS.
This will be described with reference to FIG.

[Step 1] The reduction processing unit 401 creates one reduced pixel from four pixels of the original image.

[Step 2] As shown in FIG. 41, a value having a high appearance frequency among the four pixels is set as a reduced pixel, and if the frequencies are the same, an average value (integer) is set as a reduced pixel, and a difference between the original image and the reduced image is obtained. .

[Step 3] The difference information is converted into five states by the conversion table shown in FIG.
Convert to a value data string. When there are a plurality of conversion patterns of the binary data string, the selection is made based on the values around the conversion target pixel.

[Step 4] The reduced data is encoded by the multi-level image data encoding / decoding circuit 405, and the difference information is encoded by the binary data string encoding / decoding circuit 403 and stored in the memory 406.

[Step 5] If necessary, the memory 40
6, a multi-level code and a differential code are read out, and a multi-level image data encoding / decoding circuit 405 and a binary data string encoding /
Decoding is performed by the decoding circuit 403.

[Step 6] An image is restored by the enlargement / reconstruction device 407 from the decoded reduced data and difference information.

In the case of generating the reduced image and the difference information while transmitting the error amount due to the difference conversion value to the adjacent pixels as shown in FIG. 43, the density value of the local area is preserved. Even if irreversible encoding is used for encoding a reduced image or a reduced image, visually good images can be reproduced. further,
As a modification of the present embodiment, the difference value 1
A conversion pattern may be assigned to each one.

According to the present embodiment, 2
^For data that cannot be represented by ⁿ , the binary data string encoding method can be used without wasting code words,
The coding efficiency is improved.

(Embodiment 5) FIG. 44 is a block diagram according to the present embodiment, in which input image data is divided into blocks (N × M pixels) of a predetermined size, and a DC
A DCT transforming device 5001 for performing T transform, a quantizing device 5002 for quantizing the DCT coefficient obtained by the DCT transforming device 5001, an encoding / decoding device 5003 for encoding / decoding the quantized DCT coefficient 5006, A scaling unit 5004 for scaling a decoded quantized DCT coefficient, and an inverse DCT for performing an inverse DCT transform based on the scaled DCT coefficient
It comprises a T conversion device 5005.

The original image represented by the 8 × 8 matrix in FIG. 45A is DCT-transformed by the DCT converter 5001,
FIG. 45B shows the obtained DCT coefficient quantized by the quantization device 5002, and the DCT quantization coefficient is 0, ±
1, ± 2...

FIG. 46 is a block diagram of the encoding / decoding device 5003. This encoding / decoding device 5003 includes:
A multi-level / binary conversion circuit 501 for multi-level / binary conversion of the DCT quantized coefficient, a binary data string encoding / decoding circuit 502,
It comprises a Huffman encoding / decoding circuit 503, a memory 504 for temporarily storing encoded data, and a binary / multi-level conversion circuit 505.

Next, the operation of the encoding / decoding device 5003 will be described.

[Step 1] The input DCT quantization coefficient is converted into a one-dimensional binary data string of five pixels for each multi-valued pixel by using the correspondence table shown in FIG. 47 by the multi-valued / binary conversion circuit 501. I do. When there are a plurality of conversion patterns of the binary data string, the selection is made based on the coefficients around the DCT coefficients to be converted.

[Step 2] The binary data sequence is sent to the binary data sequence encoding / decoding device 502 to perform the binary data sequence encoding.

[Step 3] In the correspondence table of FIG.
The DCT quantized coefficient that has become rs is sent to the Huffman encoding / decoding circuit 503 to perform Huffman encoding.

[Step 4] The encoded data is stored in the memory 50.
48, the area of the memory 504 is divided into two as shown in FIG. 48, one of which stores a binary data string code and the other stores a Huffman code.

[Step 5] The encoded data is read from the memory 504 as necessary, and the binary data string encoding / decoding circuit 502 and the Huffman encoding / decoding circuit 503 are read.
To decrypt.

[Step 6] Binary / multi-value conversion circuit 505
Then, the binary data sequence decoded by the binary data sequence encoding / decoding device 502 is converted into a DCT quantized coefficient, and oth
Huffman coding /
The DCT quantized coefficient decoded by the decoding circuit 503 is stored.

[Step 7] The decoded DCT quantized coefficients are sequentially output.

The DCT converter 5001 in FIG.
Quantizer 5002, scaling unit 5004, and inverse DCT
The conversion device 5005 uses JPEG (Joint Photographic Experts G), which is an international standard for color still image data compression.
roup), and when used with a quantization coefficient of 1, the SCID
In the monochrome image of 1, the DCT quantization coefficient values are 0,
Since ± 1 and ± 2 occupy 97% or more of the whole, encoding efficiency can be improved by using the encoding / decoding method of the present embodiment.

The embodiments of the present invention have been described above, and the preferred embodiments of the present invention will be listed below.

(1) The multi-value / binary conversion means selects one binary data string from a plurality of binary data strings according to the pixel value around the conversion target pixel or the image property of the multi-valued image data. And output.

(2) For multi-valued image data input to the multi-value / binary conversion means, (a) black-and-white pixel enhancement processing, (b) intermediate value conversion processing, and (c) intermediate value isolated point removal There is an image correction unit for performing at least one pre-process of the process.

By performing such preprocessing, unnecessary noise components are removed, and blurring is corrected, so that the coding efficiency is further improved.

(3) In the black-and-white pixel emphasizing process in (2), the multivalued number of the multivalued image data is set to N (N ≧ 3).
(a) When the data value P (0 ≦ P ≦ N−1) of the multi-valued image data is larger than the threshold value s1 (0 <s1 ≦ N−1), the process of setting the data value P to N−1; When the data value P of the multi-valued image data is smaller than the threshold value s2 (0 ≦ s2 <N−1), this is achieved by selectively performing one of the processes of setting the data value P to 0.

(4) In the intermediate value conversion processing in (2),
The multivalued number of multivalued image data is N (N ≧ 3), and N−1
Is a character or a line, and 0 represents the background.
When the data value P of (x, y) is in the range of (0 <P <N−1), a pixel adjacent to p and p0 (x−1, y−
1), p1 (x, y-1), p2 (x + 1, y-1),
p3 (x-1, y), p4 (x + 1, y), p5 (x-
1, y + 1), p6 (x, y + 1), p7 (x + 1, y
In the case where the number of pixels taking the value of N-1 in +1) is equal to or greater than a predetermined value, the processing is performed by performing the process of setting the value of the pixel p to N-1.

(5) In the intermediate value isolated point removal processing in (2), the multivalued number of the multivalued image data is set to N (N ≧ 3).
When N-1 represents a character or a line, and 0 represents a background,
(a) When the data value P of the pixel p is in the range of (0 <P <N−1) and there is a pixel having the data value N−1 in the peripheral pixels of the pixel p, the value of the pixel p is Leaving process, (b)
The data value P of the pixel p is in the range of (0 <P <N-1), and [p0, p7], [p1, p6], [p2, p
5] and [p3, p4], the pixel value [P ′, P ″] is (0 <P ′ <N−1: 0 <P ″ <N−
When there is no set satisfying 1), it is achieved by selectively performing one of the processing of setting the value of the pixel p to 0.

(6) Multi-valued image data is only gradation data.
Alternatively, the image processing apparatus includes an identification unit configured to include identification data of a gradation image and a character / line drawing image and gradation data, and output the identification data by distinguishing the gradation data into the gradation image and the character / line drawing image.

By providing such identification means,
In at least one of the binary / multi-level conversion unit and the image correction unit, an optimum process according to the type of the image can be performed according to the identification data, and the encoding efficiency can be further improved. In addition, by using the identification data, it is possible to easily manage the encoded data, and the cost of the apparatus is reduced.

The identification data is assigned to a multi-valued number other than the multi-valued numbers used in the character / line-drawing image and the image image, or a pattern other than the binary data string used in the character / line-drawing image and the image image is assigned. Just do it. By doing so, waste of codewords is reduced, and a decrease in coding efficiency can be avoided.

(7) Multi-valued image data is converted into a gradation image and a character /
The line image is handled with a different resolution and gradation number. By doing so, it becomes possible to handle high-quality images, and the encoding efficiency becomes higher than encoding a gradation image and a character / line image at the same resolution and the same number of gradations.

(8) An image conversion apparatus characterized in that the background color of an image is replaced with a code having high coding efficiency. By using such an image conversion device, encoding efficiency when encoding is performed is improved.

(9) The histogram of the image is used for replacing the code of the background color in (8). As a result, a background color that occupies a large proportion of the image regardless of the image is selected, and the selected background color is replaced with a code having a high coding efficiency, so that the coding efficiency when coding is performed is increased.

(10) Global and local histograms of the image are used as the histogram of the image in (9). As a result, the background color to be processed is selected from a plurality of background colors that occupy a large proportion of the image regardless of the image, and the selected background color is replaced with a code having high encoding efficiency. Encoding efficiency is increased.

(11) As the multivalued image data input to the multivalue / binary conversion means, data converted by the image conversion device of (8) is used. By assigning a background color to a code having a high coding efficiency in the image conversion device, the coding efficiency is further improved.

(12) As the multivalued image data input to the multivalue / binary conversion means, data converted by the image conversion device of (9) is used. The encoding efficiency is further improved by selecting a background color using the histogram of the image in the image conversion device and assigning the background color to a code having a high encoding efficiency.

(13) As the multivalued image data input to the multivalued / binary conversion means, data converted by the image conversion device of (10) is used. By selecting a background color using the global and local histograms of the image in the image converter and assigning the background color to a code with high coding efficiency,
The coding efficiency is further improved.

[0184]

As described above, according to the present invention,
When converting multi-valued image data into a binary image data string and encoding using a binary encoding technique, in at least one data value of the multi-valued image data, in an adjacent binary data string, Select one binary data string from a plurality of binary data strings so that the same data value will be more consecutive.
By adopting a configuration for performing value coding, it is possible to provide an image coding apparatus with high coding efficiency.

Further, according to the present invention, it is possible to provide an image decoding device capable of decoding original multi-valued image data from data encoded by the above-mentioned image encoding device.

Further, according to the present invention, in an image processing apparatus such as a printer, a copying machine or a facsimile apparatus which outputs multi-valued image data through an image memory or a transmission line by using these image encoding / decoding apparatuses. By improving the coding efficiency, it is possible to save the memory capacity, increase the amount of image information that can be stored under the limited memory capacity, and improve the transmission speed.

[Brief description of drawings]

FIG. 1 is a block diagram illustrating a configuration of an apparatus according to an embodiment 1-1.

FIG. 2 is a block diagram illustrating a configuration of an image correction device in FIG. 1;

3 is a diagram showing an input / output correspondence table of the threshold processing circuit in FIG. 2;

FIG. 4 is a diagram for explaining the operation of the intermediate value conversion circuit in FIG. 2;

FIG. 5 is a diagram for explaining the operation of the isolated point removing circuit in FIG. 2;

6 is a diagram showing various binary patterns for explaining the operation of the multi-value / binary conversion circuit in FIG.

FIG. 7 is a diagram for explaining selection conditions of L and R patterns in FIG.

FIG. 8 is a diagram showing an example of multivalued data → binary bitmap conversion by the multilevel / binary conversion circuit in FIG. 1;

FIG. 9 is a diagram showing an example of conversion from a binary bitmap to a multilevel data by the binary / multilevel conversion circuit in FIG. 1;

FIG. 10 is a block diagram illustrating a configuration of an apparatus according to an embodiment 1-2.

FIG. 11 is a block diagram illustrating a configuration of an apparatus according to Embodiment 1-3.

FIG. 12 is a diagram showing various binary patterns for explaining the operation of the multi-level / binary conversion circuit according to the embodiment 1-4.

FIG. 13 is a block diagram illustrating a configuration of an apparatus according to Embodiment 1-5.

14 is a diagram for explaining the operation of the multi-level / multi-level conversion circuit in FIG.

FIG. 15 is a view showing an example of a document image in which multi-value / multi-value conversion is effective according to the embodiment;

FIG. 16 is a diagram illustrating various correspondences between the multi-valued number of input / output multi-valued image data and the number of converted black pixels of a binary data sequence for describing Embodiment 2-1.

FIG. 17 is a diagram showing a matrix pattern of a matrix size 3 × 3 for explaining the embodiment 2-1.

FIG. 18 is a diagram illustrating an example of converting multivalued image data into a two-dimensional binary data sequence for describing the embodiments 2-1 and 2-2.

FIG. 19 is a block diagram illustrating a configuration of an apparatus according to an embodiment 2-2.

FIG. 20 is a block diagram illustrating a configuration of an apparatus according to an embodiment 2-3.

FIG. 21 is a view for explaining a method of storing data in a memory in FIGS. 19 and 20;

FIG. 22 is a block diagram illustrating a configuration of an apparatus according to an embodiment 3-1.

FIG. 23 is a diagram for explaining the operation of the identification circuit in FIG. 22;

FIG. 24 is a diagram for explaining the operation of the multi-level / binary conversion circuit in FIG. 22;

FIG. 25 is a view showing the arrangement of an apparatus according to Embodiment 3-2.

FIG. 26 is a view for explaining the operation of the image processing apparatus in FIG. 25;

FIG. 27 is a block diagram showing a configuration of an apparatus according to Embodiment 3-3.

FIG. 28 is a view for explaining the operation of the image processing apparatus in FIG. 27;

FIG. 29 is a block diagram showing a configuration of an apparatus according to Embodiment 3-4.

FIG. 30 is a block diagram showing a configuration of an apparatus according to Embodiment 3-5.

FIG. 31 is a view for explaining the operation of the image processing apparatus in FIGS. 29 and 30;

FIG. 32 is a view for explaining the operation of the image processing apparatus in FIGS. 29 and 30;

FIG. 33 is a view for explaining the operation of the image processing apparatus in FIGS. 29 and 30;

FIG. 34 is a view for explaining a method of forming an identification pattern;

FIG. 35 is a block diagram showing a configuration of an apparatus according to Embodiment 3-6.

FIG. 36 is a view for explaining the operation of the image processing apparatus in FIG. 35;

FIG. 37 is a view for explaining the operation of the image processing apparatus in FIG. 35;

FIG. 38 is a block diagram showing a configuration of an apparatus according to Embodiment 3-7.

FIG. 39 is a view for explaining the operation of the image processing apparatus in FIG. 38;

FIG. 40 is a block diagram illustrating a configuration of an apparatus according to a fourth embodiment.

FIG. 41 is a view for explaining the operation of the apparatus of FIG. 40;

FIG. 42 is a view for explaining the operation of the apparatus shown in FIG. 40;

FIG. 43 is a view for explaining the operation of the apparatus in FIG. 40;

FIG. 44 is a block diagram showing a configuration of an apparatus according to a fifth embodiment.

FIG. 45 is a view for explaining the operation of the DCT transformer and the quantizer in FIG. 44;

46 is a block diagram showing the configuration of the encoding / decoding device in FIG.

FIG. 47 is a view for explaining the operation of the encoding / decoding device in FIG. 44;

FIG. 48 is an exemplary view for explaining a method of storing data in the memory in FIG. 46;

FIG. 49 is a diagram showing an example of conversion from multi-valued image data to a binary data sequence according to a conventional technique.

[Explanation of symbols]

 101: Multi-value / binary conversion circuit 102: Binary data encoding / decoding circuit 103: Binary / multi-value conversion circuit

Claims (6)

[Claims] 1. A multi-valued / multi-valued image data conversion method for converting multi-valued image data having three or more data values into a binary data string having binary data values.
In an image coding apparatus comprising binary conversion means and coding means for coding the binary data string converted by the multi-value / binary conversion means, the multi-value / binary conversion means is multi-valued. One binary data string is selected from a plurality of binary data strings so that the same data value is more continuous in adjacent binary data strings with respect to at least one predetermined data value of the value image data. An image encoding device which is selected and output. 2. A multi-valued / multi-valued image data having three or more data values is converted into a binary data string having a binary data value.
In an image coding apparatus comprising binary conversion means and coding means for coding the binary data string converted by the multi-value / binary conversion means, the multi-value / binary conversion means is multi-valued. For at least one predetermined data value of the value image data, (a)
A first one-dimensional binary data string in which a predetermined number of identical data values are consecutively arranged from the left in the one-dimensional main scanning direction and a second one-dimensional two in which a predetermined number of identical data values are consecutively arranged from the right. Value data string, or (b) One-dimensional sub-scanning direction, a predetermined number of identical data values are arranged consecutively from the top, and a predetermined number of identical data values are arranged consecutively from the bottom. Arranged second one-dimensional binary data string, or (c) Same as the first two-dimensional binary data string in which a predetermined number of consecutive same data values are arranged consecutively from the upper left in the two-dimensional main scanning direction An image coding apparatus, characterized in that one binary data string is selected from a second two-dimensional binary data string in which a predetermined number of data values are continuously arranged and output. 3. Decoding means for decoding a binary data string from coded data obtained by coding a binary data string having binary data values, and binary data decoded by this decoding means. In a decoding device provided with a binary / multivalued conversion means for converting a string into multivalued image data having three or more data values, the binary / multivalued conversion means is a binary data string. An image decoding apparatus which outputs multi-valued image data having the same data value to at least two predetermined ones. 4. Decoding means for decoding a binary data string from coded data obtained by coding a binary data string having binary data values, and binary data decoded by this decoding means. In a decoding device provided with a binary / multivalued conversion means for converting a column into multivalued image data having three or more data values, the binary / multivalued conversion means comprises: (a) one-dimensional main scanning A first number in which a predetermined number of the same data values are consecutively arranged in order from the left of the direction
The second one-dimensional binary data sequence in which a predetermined number of the same data values are sequentially arranged from the right with the one-dimensional binary data sequence of, or (b) the same data value is sequentially determined from the top in the one-dimensional sub-scanning direction. The first one-dimensional binary data string arranged consecutively in a number, and the second one-dimensional binary data string consecutively arranged in the predetermined number of the same data values from the bottom, or (c) in the two-dimensional main scanning direction For a first two-dimensional binary data string in which a predetermined number of identical data values are consecutively arranged from the upper left, and for a second two-dimensional binary data string in which a predetermined number of identical data values are consecutively arranged from the upper right. An image decoding apparatus, which outputs multi-valued image data having the same data value. 5. Multi-valued / converting multi-valued image data having three or more data values into a binary data string having binary data values.
Binary conversion means, encoding means for encoding the binary data sequence converted by the multi-value / binary conversion means, and decoding of the binary data sequence from the encoded data obtained by the encoding means. In the image processing apparatus, the decoding means and the binary / multivalue conversion means for converting the binary data string decoded by the decoding means into multivalued image data having data values of three or more values are provided. The multi-value / binary conversion means includes a plurality of plural multi-valued image data such that a plurality of the same data values are continuous in adjacent binary data strings with respect to at least one predetermined data value. One binary data string is selected from the binary data strings and output, and the binary / multi-value conversion means has multi-valued image data of the same data value for at least two of the binary data strings. An image characterized by outputting Management apparatus. 6. A multi-valued / multi-valued image data conversion device for converting multi-valued image data having three or more data values into a binary data string having binary data values.
Binary conversion means, encoding means for encoding the binary data sequence converted by the multi-value / binary conversion means, and decoding of the binary data sequence from the encoded data obtained by the encoding means. In the image processing apparatus, the decoding means and the binary / multivalue conversion means for converting the binary data string decoded by the decoding means into multivalued image data having data values of three or more values are provided. The multi-value / binary conversion means (a) for at least one predetermined data value of the multi-valued image data
A first one-dimensional binary data string in which a predetermined number of identical data values are consecutively arranged from the left in the one-dimensional main scanning direction and a second one-dimensional two in which a predetermined number of identical data values are consecutively arranged from the right. Value data string, or (b) One-dimensional sub-scanning direction, a predetermined number of identical data values are arranged consecutively from the top, and a predetermined number of identical data values are arranged consecutively from the bottom. Arranged second one-dimensional binary data string, or (c) Same as the first two-dimensional binary data string in which a predetermined number of consecutive same data values are arranged consecutively from the upper left in the two-dimensional main scanning direction One binary data string is selected and output from the second two-dimensional binary data string in which a predetermined number of data values are continuously arranged, and the binary / multi-value conversion means outputs the binary data string. Of which (a) 1
A first one-dimensional binary data sequence in which a predetermined number of the same data values are continuously arranged in order from the left in the dimensional main scanning direction and a second one-dimensional binary in which a predetermined number of the same data values are continuously arranged sequentially from the right A data string, or (b) a first one in which a predetermined number of the same data values are successively arranged in order from the top in the one-dimensional sub-scanning direction.
A second one-dimensional binary data sequence in which a predetermined number of the same data values are sequentially arranged from the bottom with the two-dimensional binary data sequence, or
(c) a first two-dimensional binary data sequence in which a predetermined number of the same data values are continuously arranged in order from the upper left in the two-dimensional main scanning direction, and a second two-dimensional data sequence in which the same data values are continuously arranged a predetermined number sequentially from the upper right. An image processing apparatus for outputting multi-valued image data having the same data value to a two-dimensional binary data sequence.