JPH1084486A - Coding method for image data and coder - Google Patents

Abstract

PROBLEM TO BE SOLVED: To encode a binarized image at a high efficiency through a run length coding. SOLUTION: A conversion pattern of a block, consisting of 2×2 pixels where black or white pixels are consecutive in a line direction for run length coding, is prepared for a table 8, and a natural still image received from an input means 1 is binarized by a binarization means 4 and blocks each consisting of 2×2 pixels in the binary images obtained are replaced with conversion patterns, stored in the table 8 by a pattern conversion means 4. Then binary image data subject to replacement processing are run-length-coded by the MH coding and sent as G3 facsimile data from communication means 6, 7. That is, a block in the binary image is replaced with the conversion pattern, to extend the run length of the black pixel or white pixel, and the code length as a result of applying run length coding to the binary image is reduced band the data amount is reduced.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and apparatus for encoding image data for run-length encoding a binary image composed of black pixels and white pixels, and more particularly to a run-length encoding method for replacing a pixel. The present invention relates to a technique for reducing the amount of encoded data by increasing the length.

[0002]

2. Description of the Related Art In communication of image data in a facsimile apparatus, a natural still image converted into a binary image composed of black pixels and white pixels is transmitted by run-length encoding. For such facsimile transmission, IT
Various standards are defined in the UT recommendation, for example,
T. In a G3 facsimile apparatus of four standards, a binary image is run-length encoded by extended Huffman encoding (hereinafter, MH encoding).

FIG. 5 shows a conventional apparatus configuration for transmitting a natural still image to a G3 facsimile apparatus. In this apparatus, a natural still image captured from an input means 21 is stored in a frame memory 22, and this natural still image data is binarized by using a gray scale conversion and a dither method, and further subjected to MH coding. That is, bit image data 2 obtained by binarizing natural still image data
4 is held in the RAM 23, and the bit image data 24 is MH-coded and converted into facsimile data 25. Then, the facsimile data 25 is modulated using the modem 26, and the facsimile data 25 is transmitted to the G3 facsimile apparatus via a telephone line or the like.

Here, the above-mentioned binarization processing is described with reference to FIG.
To FIG. 8, and FIG. 9 to FIG.
This will be described in more detail with reference to FIG. As shown in FIG.
The color natural still image is fetched into the frame memory 22 as data of 3 bytes per pixel, and each byte data represents the gray value of the three primary colors of red (R), green (G), and blue (B). In the binarization processing of the color natural still image data, a luminance value (Y) is obtained by grayscale conversion from the RGB grayscale value of each pixel, and the luminance value of each pixel is applied to each pixel using a dither method or the like. It is determined whether to assign a black pixel or a white pixel.

In the above-described binarization processing, for example, when the error diffusion dither method is used, the binarization processing is performed as shown in FIGS. In FIG. 7, the luminance values of pixels on one line in data obtained by performing grayscale conversion on color natural still image data are shown as Y0 to Y10, and the calculated luminance for each pixel in this line is shown. When the value (original luminance value + error) is 128 or more, white pixels are assigned, and when the value is less than 128, black pixels are assigned. That is, as shown in FIG. 8, first, the error is initialized to zero and the processing is started (step S10), and the calculated luminance is calculated for the first pixel of the line (step S11). Note that this error is the calculated luminance value of the immediately preceding pixel when a black pixel is assigned to the immediately preceding pixel, whereas the error is one when a white pixel is assigned to the immediately preceding pixel. This is a value obtained by subtracting 255 from the calculated luminance value of the immediately preceding pixel. Therefore, the error is zero for the first pixel of the line.

Next, it is determined whether the calculated brightness is 128 or more (step S12). If the calculated brightness is less than 128, a black pixel is assigned to the head pixel (step S12).
13) The calculated brightness is set as the error of the next pixel (step S14). On the other hand, the calculated brightness is 1
If it is not less than 28, a white pixel is assigned to the head pixel (step S15), and 255 is calculated from the calculated luminance.
Is set as the error of the next pixel (step S16). The above processing is sequentially repeated for each pixel included in one line (step S17). When the processing for the last pixel in the line is completed, the processing for the line is terminated. Note that, for a color natural still image of one frame, the above-described binarization processing for each line is repeated for all lines, and converted into bit image data for one frame.

In the above MH encoding process, a process of representing how many black pixels or white pixels continue in the line direction (run length) of the bit image data by a code constituted by a Huffman code is performed. In particular, M
In the H coding, a terminating code as shown in FIG. 9 is assigned to the run lengths of black pixels or white pixels from 0 to 63, and the run lengths of 64 or more black or white pixels are shown in FIG. A process for allocating a makeup code as shown in FIG. For example, as shown in FIG.
In a line in which 4 white pixels, 70 black pixels, and 7 white pixels are arranged, a terminating code “1011” is assigned to a white pixel of run length “4” and a black code of run length “64 + 6”. The makeup code “000001111” and the terminating code “0010” are assigned to the pixel, and the terminating code “1111” is assigned to the white pixel of run length “7”, and the code as a whole is “1011000”.
000111100101111 ".

In actual encoding, since the head of a line is set to start from a white pixel, 1
At the beginning of the code as a whole line,
A terminating code “00” for a white pixel of “0”
110101 ". For example, as shown in FIG. 12, two black pixels, three white pixels, one black pixel,.
.., In a line in which one white pixel is arranged, a terminating code “1” is assigned to a black pixel of run length “2”.
1 "is assigned, a terminating code" 1000 "is assigned to a white pixel having a run length of" 3 ", and a terminating code" 010 "is assigned to a black pixel having a run length of" 1 ". .., A terminating code “000111” is assigned to a white pixel with a run length of “1”. Since this line starts with a black pixel, a white pixel with a run length of “0” is inserted at the head of the line. The code as a whole is “0”
01101011111000010000111010
000111010000111 ".

[0009]

As described above, when a natural image is binarized by using a dither method or the like, black pixels and white pixels having short run lengths are often arranged alternately. And the like, and converted to facsimile data, the data amount of the entire image becomes relatively large. That is, as can be seen from FIG. 9, “B” representing the number of bits per pixel of the MH code
“IT / pixel” becomes smaller as the run length becomes longer (in other words, as the run length becomes longer, the code length per pixel becomes shorter), and the coding efficiency becomes higher. However, the run length was short, and high coding efficiency could not be achieved.

FIG. 13 shows 480 pixels vertically and 768 pixels horizontally.
FIG. 14 shows a grayscale-converted natural still image of a pixel. FIG. 14 shows an image obtained by binarizing the natural still image by the error diffusion dither method.
When encoded, the total size is 151 Kbytes,
The data amount is about 3.35 bits per pixel.

The present invention has been made in view of the above-mentioned conventional circumstances, and provides an image data encoding method and an encoding device for encoding a binary image with high efficiency by run-length encoding such as Huffman encoding. The purpose is to do. The present invention also provides image data for generating a binarized image from a natural image by grayscale processing, dithering, or the like, and encoding the binarized image with run-length encoding such as Huffman encoding with high efficiency. It is an object of the present invention to provide an encoding method and an encoding device. In addition, the present invention generates a binarized image from a natural image by grayscale processing, dither method processing, or the like, and encodes the binarized image with run-length coding such as Huffman coding with high efficiency. It is an object of the present invention to provide an encoding method and an encoding device for image data transmitted as facsimile data.

[0012]

In order to achieve the above object, the present invention provides a method for run-length encoding a binary image composed of black pixels and white pixels.
A conversion pattern in which a plurality of black pixels or white pixels in the line direction of the run-length encoding is continuous is prepared in advance, and a block including a plurality of pixels in the binary image in the line direction is converted based on a predetermined reference. After the replacement with the pattern, the binary image is run-length encoded by MH encoding or the like. That is, by replacing the blocks in the binary image with the conversion pattern, the run length of the black pixels or white pixels in the block is increased, and the arrangement of the pixels in the entire binary image is converted into a state in which the run length is long. As a result, the code length as a result of the run-length encoding of the binary image is shortened, and the data amount is reduced.

In the present invention, the input natural still image is binarized by gray scale processing, dither method processing, or the like, and the obtained binary image is subjected to the above-described run-length encoding. Further, in the present invention, the run-length coded data obtained by the above-described processing is transmitted as facsimile data, thereby realizing quick image data transmission. The conversion pattern and the block in the binary image to be replaced are set as, for example, blocks each including 2 × 2 pixels adjacent to each other, but any pattern or block including two or more pixels connected in the line direction. ,
It can be set arbitrarily.

Further, the present invention is realized by an image data encoding device having a facsimile transmission function, for example.
It has a table holding a conversion pattern of blocks of 2 × 2 pixels in which black pixels or white pixels continue in the line direction of the run-length encoding, and binarizes a color natural still image to be processed input from the input means. Binarization is performed by gray scale processing and dithering by means. And
The pattern conversion means replaces the block composed of 2 × 2 pixels in the obtained binary image with a conversion pattern held in a table based on a predetermined criterion, and the binary image subjected to the replacement processing is subjected to extended Huffman coding. , The encoding means performs run-length encoding, and the transmission means transmits the run-length encoded binary image data as G3 facsimile data. Therefore, the transmission of image data to the G3 facsimile apparatus can be performed rapidly with a significantly reduced data amount.

[0015]

DESCRIPTION OF THE PREFERRED EMBODIMENTS A method and apparatus for encoding image data according to the present invention will be described with reference to the drawings. FIG. 1 shows an image data encoding apparatus according to one embodiment of the present invention. The encoding apparatus includes a natural image input unit 1 for capturing a natural still image, an image storage memory 2 for storing captured image data, a ROM 3 storing a control program and a conversion pattern table 8 described later, and a control program. , A RAM 5 forming a work area for processing by the main control unit 4, a communication unit 6 for performing facsimile data transmission processing, and modulating the facsimile data. G3
And a modem 7 for transmitting to the facsimile machine.

The main control unit 4 executes a control program to execute grayscale conversion processing on natural still image data, and performs error diffusion dither processing on the grayscale converted data. Error diffusion dither processing means, pattern conversion processing means for performing replacement processing of the dither-processed binary data with the conversion pattern shown in FIG. 3, and MH encoding processing for the pattern-converted binary data Various kinds of processing means such as an MH coding means for performing, an communication control means for transmitting the MH-coded facsimile data via the communication means 6 and the modem 7, and the like. That is, the process of encoding the natural still image data held in the image storage memory 2 and converting it into facsimile data is performed by the above-described processing means in the procedure shown in FIG.

First, the natural still image data held in the image storage memory 2 is read for each pixel along a horizontal line, grayscale conversion is performed, and the grayscale data is stored in the buffer 9 in the RAM 5 (step). S
1). Next, after performing an error diffusion dither process on the grayscale data (step S2), a pattern conversion process is performed using the conversion pattern held in the conversion pattern table 8 (step S3). Is stored in the buffer 10 in the memory. Then
MH-encoding the pattern-converted data (step S
4) The MH-encoded facsimile data is stored in the RAM 5
Is stored in the buffer 11 in the storage area. Then, the facsimile data stored in the buffer 11 is read and transmitted to the G3 facsimile apparatus via the communication means 6 and the modem 7.

In the present embodiment, the above-described pattern conversion processing is performed using the conversion pattern shown in FIG. FIG.
As shown as “image pattern after conversion” in (a) to (f), this conversion pattern is a block pattern of 2 × 2 pixels, and each pattern is black in a horizontal direction (line direction of run-length encoding). The pixel array pattern has a series of pixels or white pixels. That is, (a) and (b)
In the pattern (1), white pixels are continuous in the upper and lower stages in the line direction, respectively. In the patterns (c) and (d), white pixels and black pixels are continuous in the upper and lower stages in the line direction.
Is a pattern in which black pixels are respectively continuous in upper and lower two stages in the line direction.

These conversion patterns are replaced with 2 × 2 pixel block patterns (“original image patterns” in FIGS. 3A to 3F) in the dither-converted data (bit image data). Each of the conversion patterns (a) to (f) is associated with some “original image patterns” in accordance with a predetermined standard.
That is, when the “original image pattern” is all white pixels or only one white pixel as shown in FIG. 3A or 3B, it is associated with the conversion pattern of all white pixels, and FIG. ) Or (f), in the case of only one black pixel or all black pixels, it is associated with the conversion pattern of all black pixels. Also, “original image pattern” is shown in FIG.
When the number of black pixels and the number of white pixels are the same as shown in (c) or (d), the upper two pixels are set as black or white pixels and the lower two pixels according to the upper left pixel in the block pattern. It is associated with a conversion pattern in which two pixels are white pixels or black pixels.

In the conversion process using the above conversion pattern, a block of 2 × 2 pixels included in the bit image data is sequentially extracted along the encoding line direction, and the pixel pattern of each block (see FIG. 3). "Original image pattern"
Is replaced with a corresponding conversion pattern (corresponding to the “converted image pattern” in FIG. 3). For example, the bit image data before the conversion processing is performed as shown in FIG.
In the case of the pixel array shown in FIG. 4A, the pattern conversion process causes the pixel array in the bit image data to be such that black pixels and white pixels are long and continuous in the line direction as shown in FIG. The run length is converted to long bit image data.

Therefore, compared with the case where the bit image data before the conversion is MH-coded, the amount of the coded data can be greatly reduced by performing the MH coding on the bit image data after the pattern conversion. . FIG.
5 shows an image obtained by subjecting the grayscale converted image of 480 pixels vertically and 768 pixels horizontally shown in FIG. 13 to error diffusion dither processing and further pattern conversion processing. In the image data subjected to the error diffusion dither processing shown in FIG.
When encoded, the total data amount was 151 Kbytes, and the data amount was about 3.35 bits per pixel of the original image.
The data amount is about 0.36 bits per pixel of the original image of K bytes, and the data amount is reduced to about 11%.

It should be noted that various other modes can be adopted for associating the original block pattern with the conversion pattern shown in the above-described embodiment. What is necessary is to keep the black pixels or white pixels as continuous as possible in the line direction of the run-length encoding while keeping as much as possible. In the above embodiment, the run-length encoding is performed by MH encoding. However, the run-length code amount is generally more efficient as the run-length is longer. It is also possible to perform run-length encoding.

In the above embodiment, the gray scale processing and the error diffusion dither processing are used for the binarization processing. However, a dither method of another method can be used. There is no particular limitation on the method as long as the value can be converted.

When the blocks and conversion patterns to be converted are 2 × 2 pixel patterns as described in the above embodiment, the run length can be easily expanded while maintaining the density distribution of the original image. According to the present invention, if a pattern in which two or more pixels are continuous in the line direction, 2
Various pixel patterns other than × 2 pixels can be employed. Also, in the present invention, not only monochrome images,
A color image can also be encoded by binarization.

[0025]

As described above, according to the present invention,
After the block composed of a plurality of pixels connected in the line direction in the binary image is replaced with the conversion pattern in which the black pixels or the white pixels are connected, the binary image is run-length coded. Can be significantly reduced as compared with the related art. Therefore, facsimile data obtained by encoding image data with high efficiency can be obtained, and data transmission can be performed quickly. In particular, application to a G3 facsimile apparatus using MH coding for run-length coding is effective, and it is possible to realize faster data transmission as compared with the related art.

[Brief description of the drawings]

FIG. 1 is a configuration diagram illustrating an image data encoding device according to an embodiment of the present invention.

FIG. 2 is a diagram showing a processing procedure and processing means of an image data encoding device according to an embodiment of the present invention.

FIG. 3 is a diagram showing a conversion pattern according to an embodiment of the present invention.

FIG. 4 is a diagram illustrating a pattern conversion process according to an embodiment of the present invention.

FIG. 5 is a configuration diagram showing a conventional image data encoding device.

FIG. 6 is a diagram illustrating a binarization process of a color image.

FIG. 7 is a diagram for explaining processing by an error diffusion dither method.

FIG. 8 is a flowchart showing a procedure of processing by an error diffusion dither method.

FIG. 9 is a diagram showing a terminating code of an MH code.

FIG. 10 is a diagram illustrating a makeup code of an MH code.

FIG. 11 is a diagram illustrating an example of MH coding.

FIG. 12 is a diagram illustrating another example of MH coding.

FIG. 13 is a diagram illustrating an example of an image that has been subjected to grayscale processing.

FIG. 14 is a diagram illustrating an example of an image subjected to error diffusion dither processing.

FIG. 15 is a diagram illustrating an example of an image subjected to a pattern conversion process.

[Explanation of symbols]

1 ... natural image input means 4 ... main control unit 6
... Communication means, 7 ... Modem, 8 ... Conversion pattern table, 9 ... Gray scale data buffer, 10 ... Pattern conversion data buffer, 11 ...
・ Facsimile data buffer,

────────────────────────────────────────────────── ───

[Procedure amendment]

[Submission date] November 22, 1996

[Procedure amendment 1]

[Document name to be amended] Statement

[Correction target item name] FIG.

[Correction method] Change

[Correction contents]

FIG. 13 is a drawing substitute photograph showing an example of an image subjected to grayscale processing.

[Procedure amendment 2]

[Document name to be amended] Statement

[Correction target item name] FIG.

[Correction method] Change

[Correction contents]

FIG. 14 is a drawing substitute photograph showing an example of an image subjected to error diffusion dither processing.

[Procedure amendment 3]

[Document name to be amended] Statement

[Correction target item name] FIG.

[Correction method] Change

[Correction contents]

FIG. 15 shows an example of an image subjected to a pattern conversion process.
It is a drawing substitute photograph .

Claims (9)

[Claims] 1. An image data encoding method for run-length encoding a binary image composed of black pixels and white pixels, wherein a conversion pattern in which a plurality of black pixels or white pixels in a line direction of the run-length encoding is connected in advance. It is characterized in that after preparing, a block composed of a plurality of pixels connected in the line direction in the binary image is replaced with the conversion pattern based on a predetermined criterion, and then the binary image is run-length encoded. Image data encoding method. 2. An image data encoding method for run-length encoding a binary image composed of black pixels and white pixels, wherein a conversion pattern in which a plurality of black pixels or white pixels in a line direction of the run-length encoding is continuous is determined in advance. On the other hand, the natural still image is binarized, and a block composed of a plurality of pixels connected in the line direction in the obtained binary image is replaced with the conversion pattern based on a predetermined criterion. A method for encoding image data, comprising: performing run-length encoding on an image. 3. An image data encoding method for transmitting a binary image composed of black pixels and white pixels by run-length encoding and transmitting the binary image, wherein a plurality of black pixels or white pixels in a line direction of the run-length encoding are continuous. Is prepared in advance, the natural still image is binarized, and a block composed of a plurality of pixels connected in the line direction in the obtained binary image is replaced with the conversion pattern based on a predetermined criterion. A method for encoding image data, wherein a binary image is run-length encoded and transmitted. 4. The image data encoding method according to claim 2, wherein the binarization processing further includes a dither method. 5. The image data encoding method according to claim 1, wherein the run-length encoding is performed by extended Huffman encoding. Method. 6. The image data encoding method according to claim 1, wherein the conversion pattern and the block in the binary image to be replaced include 2 × 2 pixels adjacent to each other. A method for encoding image data, characterized in that the image data is a block comprising: 7. An image data encoding apparatus for run-length encoding a binary image composed of black pixels and white pixels, wherein a conversion pattern in which a plurality of black pixels or white pixels in a line direction of the run-length encoding is continuous. Table, a pattern conversion unit that replaces a block composed of a plurality of pixels connected in the line direction in the binary image with a conversion pattern held in the table based on a predetermined reference, and a replacement process performed by the pattern conversion unit. Encoding means for performing run-length encoding of the value image; and an encoding device for image data, comprising: 8. An image data encoding device for run-length encoding a binary image composed of black pixels and white pixels, wherein a conversion pattern in which a plurality of black pixels or white pixels in a line direction of run-length encoding is continuous. Means for inputting a natural still image to be processed, binarizing means for binarizing the input natural still image, and a plurality of pixels connected in the line direction in the obtained binary image. Pattern conversion means for replacing the blocks formed by the conversion pattern stored in the table on the basis of a predetermined criterion, and coding means for run-length coding the binary image subjected to the replacement processing by the pattern conversion means. An image data encoding device characterized by the above-mentioned. 9. An image data encoding apparatus for transmitting a binary image composed of black pixels and white pixels by run-length encoding and transmitting the binary image, wherein 2 × 2 pixels in which black pixels or white pixels continue in a line direction of the run-length encoding. A table holding a conversion pattern of blocks consisting of the following: input means for inputting a color natural still image to be processed; binarization for binarizing the input color natural still image using grayscale processing and dithering Means, pattern conversion means for replacing a block consisting of 2 × 2 pixels in the obtained binary image with a conversion pattern held in a table on the basis of a predetermined criterion, and binary data subjected to a replacement process by the pattern conversion means Coding means for performing run-length coding on the image by extended Huffman coding; and G3 coding for the run-length-coded binary image data. Encoding apparatus of the image data, characterized by comprising a transmitting means for transmitting a comb millimeter data.