JP2570437B2 - Facsimile machine - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is used in a facsimile apparatus, and in particular, encodes a pseudo halftone binary image using an adaptive information source encoding method using an error diffusion method. Related to the device.

〔Overview〕

The present invention relates to a facsimile apparatus which generates binary image data by multiplying digital image data by using an error diffusion method, encodes the data, and transmits / receives the binary image data. The coding efficiency is improved by performing an image data prediction operation that is adapted accordingly, transmitting the image data, transmitting the image data, and receiving the data, performing the image data prediction operation and decoding the image data.

[Conventional technology]

Conventionally, as a pseudo halftone binarization technique in this type of facsimile apparatus, as shown in FIG. 4, a multi-tone halftone original is read by a CCD reading unit (not shown), and the analog image data is converted into an analog image data. Digital image data x _i , which is digitally converted and converted into multi-gradation digital image data,
_j and the error data before it is input to the comparator 1.
The e _i , _j are superimposed by the adder 6 according to the following equation (1) using the a _i , _j as coefficients, and input to the comparator 1 as digital image data x _i , _j .

x _i , _j = x _i , _j + a _i−1 , _j−1 · e _i−1 , _j−1 + a _i−1 , _j · e _i−1 , _j + a _i−1 , _{j + 1} · e _{i -1, j + 1 + a i} , j-1 · e i, i-1 ···· (1) _{However, a i-1, j-} 1 + a i-1, j + a i-1, j + 1 + A _i , _j-1
= 1 The digital image data x _i , _j is compared with an arbitrary threshold value in the comparator 21 to output digital image data y _i , _j expressed by the following equation (2).

However, the x _i , _j MAX value and the x _i , _j MIN value are the multi-gradation digital image data x _i , _j
Means the maximum and minimum values of. Further, the threshold values satisfy x _i , _j MIN value <threshold value <x _i , _j MAX value. The digital image data y _i , _j input to the binarization unit 8 is binarized by the following conditional expression (3).
The image data z _i and _j are output.

Digital image data x _i that is input to the comparator 1, error quantization for superimposing the _j data e _{i, j} is defined by the following equation (4).

e _i , _j = x _i , _j− y _i , _j ... (4) The quantization error data e _i , _j output from the subtractor 7 is
Error data storage memory 2 and one-image data delay (T)
5 is stored in accordance with the control of the memory control unit 3, the digital image data x _i, the quantization error data required for calculating the error component superimposed on the _j e _{i, j} is taken out.

As described above, the multi-tone digital image data x _i , _j
By superimposing the quantization error data e _i , _j on the multi-level digital image data x _i , _j , the quantization error of the entire original screen is minimized.

The binarized binarized image data z _i , _j is information source encoded by the MH coder 13 and stored or transmitted to a communication line.

[Problems to be solved by the invention]

In the facsimile apparatus using the above-described pseudo halftone binarization method using the conventional error diffusion method, appropriate reproducibility is obtained theoretically and experimentally with respect to the gradation level of shading. The pseudo halftone binarization method is used in the CCITT (International Telegraph and Telephone Consultative Committee) recommendation because the correlation with the surrounding pixels is very small and the quantization error in the surrounding area is superimposed when binarizing. In contrast to the information source codes MH code, MR code, MMR code, etc., the code compression efficiency is very poor, and there is a disadvantage that the image data storage efficiency and the communication speed are reduced.

The object of the present invention is to eliminate the disadvantages mentioned above,
It is an object of the present invention to provide a facsimile apparatus in which the coding efficiency is improved, and the image data storage efficiency and the communication speed are improved.

[Means for solving the problem]

The present invention provides a binarized image data generating means for generating binarized image data using multi-tone digital image data by using an error diffusion method, and converts the binarized image data into communication data of a predetermined coding system. A facsimile apparatus including an encoding unit including a communication data generating unit for transmitting and transmitting, and a decoding unit for decoding the received communication data, wherein the encoding unit converts the binarized image data line by line. The image processing apparatus further includes a first prediction operation processing unit that performs an image data prediction operation adapted in accordance with the encoding efficiency and outputs the image data prediction operation to the communication data generation unit. The image processing apparatus further includes a second prediction operation processing unit that decodes the binary image data.

[Operation] The first prediction calculation processing means checks the coding efficiency of the binarized image data line by line, and performs an adaptive image data prediction calculation according to the checked value. For example, for each line, the number of changes of the binary image data from “1” to “0” or “0” to “1” is counted,
If the count is equal to or greater than a certain value, the prediction operation is performed.

As a result, it is possible to increase the correlation between the line having a low coding efficiency and the pixels in the peripheral area, and it is possible to increase the coding efficiency.

〔Example〕

Hereinafter, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram showing a main part of an encoding unit according to an embodiment of the present invention, FIG. 2 is a block diagram showing a main part of a decoding unit, and FIG. FIG.

According to FIG. 1, an encoding unit according to an embodiment of the present invention converts a multi-tone digital image data x _i , _j into binary image data z _i , _j using an error diffusion method. A comparator 1, an error data storage memory 2, a memory controller 3, a coefficient multiplier (M) 4, a one-image data delayer (T) 5, an adder 6, a subtractor 7, and binary data as image data generating means And an MH coder 13 as communication data generating means.

Further, the first feature of the present invention is that the first binarized image data z _i , _j is subjected to image data prediction calculation adapted for each line in accordance with the coding efficiency and output to the MH coder 13. Change point counting section 9, 4-line buffer 10, prediction calculation processing section 11, selector as prediction calculation processing means
12 and a prefix adding unit 14.

Further, according to FIG. 2, in addition to the MH decoder 32, the decoding unit according to one embodiment of the present invention uses the binary communication based on the received communication data v _i , _j which is a feature of the present invention. A prefix analysis unit 31, a selector 33, and a four-line buffer 34 as second prediction calculation means for decoding the image data z _i , _j
And a prediction operation processing unit 35.

 Next, the operation of the present embodiment will be described.

Multi-tone digital image data x _{i, j} is input, the error of the peripheral pixels a _i, the quantization error is weighted by a factor of _j data e _{i, j} is output from the coefficient multiplier 4, the adder 6 digital image data _{x 'i, j} is generated by the digital image data x _i, and _j superimposed by. The digital image data x ′ _i , _j is compared with an arbitrary threshold value of the comparator 1, and digital image data y _i , _j is output. Then, the difference between the digital data x ′ _i , _j and y _i , _j is obtained by the subtracter 7 and input to the error data storage memory 2 and the one-image data delay unit 5 as error data e _i , _j . The error data e _i , _j stored in the error data storage memory 2 are sequentially output to the one-image data delay unit 5 according to the memory control unit 3.

The digital image data y _i , _j is binarized by the binarizing section 8 and output as binarized image data z _i , _j . The parameters and variables shown in FIG.
All of the above expressions (1), (2), (3) and (4)
According to the formula.

The binarized data z _i , _j output from the binarization unit 8 are:
The data is input to the change point counting unit 9 where the binary data
The number of changes in numerical values of z _i and _j from “1” to “0” or “0” to “1” is counted, and when this count number D is a certain value or more, a prediction operation is performed, If it is less than a certain value, the prediction calculation process is not performed. It is assumed that the four-line buffer 10 stores data necessary for the prediction calculation processing unit 11 to perform the prediction calculation. The selector 12 uses the image data r _i , _{j in} accordance with the information from the change point counting unit 9. Alternatively, s _i and _j are switched in line units.

The binary data that has passed through the selector 12 is source-coded by the MH coder 13 and input to the prefix adding unit 14 as coded data u _i , _j , and a 1-bit prefix for the presence or absence of prediction calculation processing is added to each line unit And output from the prefix adding unit 14.

Here, it is assumed that the prediction calculation processing unit 11 performs the following calculation using the image data of the area shown in FIG. In FIG. 3, reference numeral 21 denotes target image data, and hatched portions indicate error feedback areas 22 and 23, which are peripheral prediction areas. In the error feedback area 22, equation (1)
The digital image data x _i , _j is converted into binary image data z _i , _j based on the expressions (2), (3) and (4). Here, as an assumption, the average density of the peripheral prediction region 23,
The average density of the error feedback area 22 is substantially equal. It is also assumed that the peripheral prediction area 23 and the error feedback area 22 have a smooth gradient component. Based on this assumption, the components are defined as follows.

r _i , _j = z _i , _j (5) However, K ₁ is peripheral prediction region the number of data However, K ₂ is hatched region number of data Assuming that the above-mentioned average density of the area 23 and the average density of the area 22 are substantially equal, α ≒ β + γγ ≒ α-β, and the following prediction is performed using this α-β. That is, To predict.

Where z _i , _j is the predicted value of z _i , _j And

As for the decoding, as shown in FIG. 2, the prefix of the code data input in units of lines is analyzed by the analysis unit 31, and it is determined whether or not the line has been subjected to the prediction calculation processing. Communicate as information.
The encoded data u _i after passing through the prefix analysis unit 31,
_j is decoded by the MH decoder 32, passes through the selector 33, is directly input to the 4-line buffer 34 when the prediction operation is not performed, and is input when the prediction operation is performed. In the processing unit 35, equation (6), (7)
By the conditions of the equations (8), (9) and (10),
The binarized image data z _i , _j is restored and a 4-line buffer
Stored in 34.

〔The invention's effect〕

As described above, according to the present invention, a pseudo-halftone binary data using an error diffusion method in a facsimile apparatus is subjected to a change point count at one line position of image data, and a line having low encoding efficiency is obtained. Is subjected to a prediction calculation process so that the correlation with surrounding pixels is high, and MH, MR and
By increasing the coding efficiency for MMR codes,
There is an effect that the storage efficiency of image data and the communication speed can be improved.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a main part of an encoding unit according to one embodiment of the present invention. FIG. 2 is a block diagram showing a main part of a decoding unit according to one embodiment of the present invention. FIG. 3 is an explanatory diagram showing the prediction calculation processing execution area. FIG. 4 is a block diagram showing a main part of a conventional encoding unit. 1 comparator 2 error data storage memory 3 memory control unit 4 coefficient multiplier (M) 5 1-picture data delay unit (T) 6 adder 7 …… Subtractor, 8 ……
Binarization section, 9 Change point counting section, 10, 34 ... 4 line buffer, 11, 35 ... Prediction calculation processing section, 12, 33 ... Selector, 13 ... MH coder, 14 ... Prefix addition , 21... Attention image data, 22... Error feedback area, 23... Predicted peripheral prediction area, 31... Prefix analysis section, 32.

Claims (1)

(57) [Claims] 1. A binary image data generating means for generating binary image data from multi-tone digital image data by using an error diffusion method, and converting the binary image data into communication data of a predetermined coding system. In a facsimile apparatus including an encoding unit including a communication data generating unit for converting and transmitting, and a decoding unit for decoding the received communication data, the encoding unit converts the binarized image data line by line. And a first prediction calculation processing unit that performs an image data prediction calculation adapted in accordance with the encoding efficiency and outputs the predicted data to the communication data generation unit, wherein the decoding unit is configured based on the received communication data. A second prediction operation processing means for decoding the binarized image data into a binary image data, wherein the first prediction operation processing means counts the number of change points of the binary image data for each line. Means,
A facsimile apparatus comprising: a means for executing a prediction code operation when the counted number of change points is equal to or more than a predetermined value, and not executing the prediction code operation when the counted number of change points is less than the predetermined value.