JPH02126772A - Picture data encoding system - Google Patents

Abstract

PURPOSE:To eliminate the influence of the characteristic of a picture input device by providing a peak detecting circuit and a dynamic range altering circuit, and attaining expression in gradation expressible with a picture display device by means of a picture element value at the white or black peak of a density histogram. CONSTITUTION:A dynamic range altering circuit 3 is provided which converts the picture element value between the picture element value at the white peak and the picture element value at the black peak obtained by a peak detecting circuit 2 picture element value expressing between one edge and the other edge in the gradation expressible by a picture display device 11 in a prescribed procedure, and outputs it to an encoding circuit 8. The deflection of the gradation of input picture data is corrected so as to be distributed in the whole range of the gradation expressible by the picture display device 11 by means of a preprocessing, and the influence of the characteristic of the input device is eliminated. Thus, the picture data can be coded at high compression ratio without being affected by the characteristic of the picture input device 1 and the stain of the picture, etc.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of Application The present invention relates to an image data encoding method for the purpose of data compression.

<Prior art> As an image data encoding method for encoding image data obtained from black and white images such as documents, graphs, tables, etc., input image data is binarized using a certain threshold value.
In many cases, value image information is encoded by MH (Modify Huffman) or MR (Modify Read).

The mode method is well known as a binarization process for input image data taken in from an image input device.

This mode method performs binarization processing as follows.

That is, it is known that the histogram of shading in the black-and-white shading target image described above becomes bimodal in correspondence with white and black, as shown in FIG. Therefore, a threshold value is set between the two peaks of white and black, and the image data is binarized.

In this way, M
Processing such as R encoding or MH encoding is performed.

In the above MH encoding, a short code is assigned to a run length that appears frequently (the number of whites or blacks that continue in the scanning line direction), and a long code is assigned to a run length that appears less frequently. By doing this, image data is compressed for each scanning line. That is, the run length (RL)
is expressed as RL-(64M)+(T), and data is compressed by assigning a Huffman code to each of the terms (64M) and (T) in the above equation. Here, the term (64M) is a make-up code term, and the (T) term is a terminating code term.

On the other hand, the above MR encoding compresses image data by utilizing the fact that images have correlation in both the vertical and horizontal directions. In other words, the relative distance ( number of pixels)
Data is compressed by encoding it.
However, when an image based on binary image data obtained by compressing image data as described above is output to an image display device (or image output device) such as a display or a printer, an image as shown in FIG. 15(a) is output. However, the appearance is very poor due to jagged edges.

Therefore, recently, the edges of the target image are often represented as black-and-white shading images.

In other words, by inserting an intermediate color between the white and black of the edge part, the sudden change in black and white is eliminated, and the result shown in Fig. 15 (b
), it performs anti-aliasing processing to prevent the appearance of jagged edges. At this time, performing anti-aliasing processing on binary image data requires a very large amount of calculation. Therefore, a method has been proposed in which the pixel values of multi-gradation image data read by an image input device are directly encoded without being binarized, and the target image is displayed in multi-gradation.

<Problems to be Solved by the Invention> However, in the conventional method of directly encoding the pixel values of read image data, the pixel values read from the image input device are directly used, so the resulting image is It will depend largely on the characteristics of Another problem is that it is easily affected by dirt on the original.

The former problem is mainly related to the dynamic range of the image. That is, for example, O to 7 on an image display device.
When displaying in 8 gradations, if the minimum value of the pixel value read by the image input device is 2 and the maximum value is 5, then 4
This becomes a gradation expression, and the remaining four levels are useless. Furthermore, the image displayed on the image display device also becomes an image with unclear shading. Also, the latter problem is
Dirt on the document prevents the generation of a long run length during compression, resulting in a reduction in the compression ratio of image data.

SUMMARY OF THE INVENTION Accordingly, an object of the present invention is to provide an image data encoding method that can encode image data at a high compression rate without being affected by the characteristics of an image input device or image stains.

<Means for Solving the Problems> In order to achieve the above object, the present invention encodes black and white image data of documents, graphs, tables, etc. input from an image input device using a predetermined gradation using an encoding circuit. In an encoding method for image data, a density histogram is obtained from pixel values of input image data, and a peak detection circuit obtains pixel values at two peaks corresponding to white and black in the histogram, and the peak detection circuit The pixel value at the white peak is converted to a pixel value representing one end of the gradation that can be expressed by the image display device, while the pixel value at the black peak is converted to the other end of the gradation that can be expressed by the image display device. Convert to pixel value representing the edge,
Furthermore, the pixel value between the pixel value at the white peak and the pixel value at the black peak determined by the peak detection circuit is determined by a predetermined procedure as the one end and the other end of the gradation that can be expressed by the image display device. It is equipped with a dynamic range changing circuit that converts into pixel values representing the range between
The present invention is characterized in that preprocessing is performed so that the image is distributed over the entire range of gradations that can be expressed by the image display device, thereby eliminating the influence of the characteristics of the input device.

Furthermore, the image data encoding method binarizes the pixel values converted by the dynamic range changing circuit based on a set threshold, and indicates different pixel values according to the obtained binary image information. an edge detection circuit that detects adjacent pixel locations as edge sections; and a pixel value converted by the dynamic range changing circuit that converts the pixel value at the edge detected by the edge detection circuit out of the binary image information into a pixel The present invention is characterized in that it includes a noise removal circuit that replaces the value with a value and outputs it to the encoding circuit, thereby removing noise caused by the influence of the characteristics of the image input device, dirt on the document, and the like.

Function> When image data of a predetermined gradation of a black and white gray image is input from an image input device, a density histogram is determined by the pixel values of the input image data by the peak detection circuit, and the density histogram corresponding to white and black of the histogram is determined. The pixel values at the two peaks are determined.

Then, the dynamic range changing circuit converts the pixel value at the white peak determined by the peak detection circuit into a pixel value representing one end of the gradation that can be expressed by the image display device, while converting the pixel value at the black peak. is converted into a pixel value representing the other end of the gradation that can be expressed by the image display device, and further, the pixel value between the pixel value at the white peak and the pixel value at the black peak determined by the peak detection circuit is The value is converted into a pixel value expressing between the above-mentioned - end and the other end of the gradation that can be expressed by the image display device according to a predetermined procedure, and is output to the encoding circuit. Therefore, the gradation bias of the input image data is corrected so as to be distributed over the entire range of gradations that can be expressed by the image display device.

Thereafter, the pixel information output from the dynamic range changing circuit is encoded by the encoding circuit.

In addition, in the image data encoding method, when the pixel values of the input image data are converted by the dynamic range changing circuit into pixel values distributed over the entire range of gradations that can be expressed by the image display device, the edge The portion detection circuit binarizes the converted pixel value based on a set threshold value, and detects a portion of adjacent pixels exhibiting different pixel values as an edge portion according to the obtained binary image information. The noise removal circuit then replaces the pixel values of the edge portions of the binary image information with the pixel values converted by the dynamic range changing circuit, and outputs the pixel values to the encoding circuit. Therefore, the influence of the characteristics of the image input device is removed, as well as noise caused by dirt on the original, etc.

Thereafter, the binary image information in which the edge portions are replaced with multi-gradation pixel values is encoded by the encoding circuit.

<Example> Hereinafter, the present invention will be explained in detail with reference to illustrated embodiments.

FIG. 1 shows a block diagram of an image data encoding/decoding apparatus according to the high efficiency encoding method for image data of the present invention.

A document is captured as N-bit gradation digital data from an image input device 1 such as a scanner, and the captured N-bit gradation digital data is input to a peak detection circuit 2. The peak detection circuit 2 performs the following processing. here,
Black is represented by “0” and white is represented by “2N”. In other words,
As shown in Fig. 2, a histogram hi (hi is the frequency of appearance in a certain interval i of image pixel values) of image density (pixel values) is created, and a threshold value T (0 <T<2N-1), the pixel value P at the black peak is calculated from the above histogram hi by equation (1).
Pixel values Pw at the peaks of B and white are calculated.

Next, the dynamic range adjustment circuit 3 performs the processing expressed by equation (2) on the calculated PB and P%, and as shown in FIG. 2, the 2M gradation ( Convert to a value (N>M).

Here, d' Quantization width Qi' of the image data taken in from the image input device l: Each quantization value in section i d: Insect drying width Qi after conversion to 2M gradation image data: 2MPa gradation image By performing a conversion process of each quantization value or higher after conversion to data, it becomes possible to correct pixel bias caused by the influence of characteristics of the image input device l.

The threshold determination circuit 4 uses the output result from the dynamic range adjustment circuit 3 after the conversion process as described above has been performed.
Then, the binarization threshold T H necessary for edge detection is determined by equation (3).

TH-(PB+P1)/2-(3) In this case, the threshold TH is the minimum line thickness when displaying characters, graphs, tables, etc. on the image display device 11, etc.
(minimum display width). Therefore, instead of determining the threshold value TH using equation (3), if the user determines the threshold value TH while checking the thickness of the line on the screen, etc., the threshold value TH can be determined easily and conveniently. It is.

In order to realize such a method of determining the threshold value TH, a lookup table 7 is used. Then, as shown in FIG. 3, the contents of the lookup table 7 are rewritten. That is, the user sets an arbitrary threshold value TH for the lookup table 7. Then, if the input pixel value P is less than the threshold TH, the corresponding binarized pixel value f(P) is set to "0" (black), and the pixel value P
If is greater than or equal to the threshold TH, the contents of the lookup table 7 are rewritten so that the corresponding binarized pixel value f(P) is output as "2N-bi (white)".

Then, based on the lookup table 7 rewritten in this way, pixel values less than the threshold fiaT) are output as black, and pixel values greater than or equal to the threshold TH are output as white to the image display device II. Determine the appropriate T and I while checking the thickness of the line displayed.

The threshold value TH determined in this way is determined by the edge detection circuit 5.
The enon detection circuit 5 uses the threshold value when binarizing the image data.

The binarization process in the edge detection circuit 5 is performed using the conversion equation (4).

Here, X: Pixel value before conversion y; Pixel value after conversion Based on the binary image data obtained in this way, the noise removal circuit 6 selects pixels whose adjacent pixels are different as shown in FIG. The pixel location indicating the value is detected as an edge portion. Then, the pixel value of the pixel near this edge part is set to “
The binarized pixel values of "0" and "2N-1" are replaced with the 2N gradation pixel values output from the dynamic range adjustment circuit 3 before binarization.

In this way, since the noise removal circuit 6 detects only the changing part of the binary image data as an edge part, the edge detection circuit 6
In this case, the influence of the pixel values of the dirt areas that have pixel values equal to or higher than the threshold value of 18 and are converted to "2N-1 (white)" is removed, and only the edges of the image are detected. In other words, only the edge portions of the image are detected. By performing preprocessing such as this, it is possible to create original image data that is not significantly affected by dirt on the original.

The above peak detection circuit 2. Dynamic range adjustment circuit 3
．． Threshold value determination circuit 4. The edge detection circuit 5 and the noise removal circuit 6 constitute a preprocessing circuit 12 of the image data encoding/decoding device.

The pixel values of the image data thus obtained are input to the encoding circuit 8 and encoded at a high compression rate. The obtained encoded data is then written to the recording/reproducing device 9, and when displayed, is decoded by the decoding circuit 10 and displayed on the image display device 11.

Hereinafter, an example of encoding the above-mentioned image data using the high-efficiency encoding method will be specifically shown.

For example bits l&N=4 (i.e. 24
= 16 gradations), a black and white grayscale image is read.

An example of the original at this time is shown in FIG. 5(a), and an example of image data consisting of 8×8 pixels captured from the original by the scanner 1 is shown in FIG. 5(b). Here, the threshold value T obtained based on experiment or experience is assumed to be T=7.

FIG. 6 is a histogram of pixel values of input image data taken in from the scanner I. From this histogram, the peak detection circuit 2 determines the pixel value Pl at the white peak and the pixel value PB at the black peak using the above formula.

PB = 3 P buy = 12 Next, the PB and Pi results are input to the dynamic range adjustment circuit 3 and processed according to the above equation (2),
The pixel values taken into the image input device 1 are transferred to the image display device I.
It is expressed by M=2 (that is, 22=4 gradations) which can be expressed by I. That is, using the above equation (2), the ulceration width and the insect drying value before and after conversion are determined as follows.

d°2(12-3-1)/(4-1)=2.70,'
=3, Q1'=6, Qt'=8, Q3'=11d=(1
6-1)/(4-1)=5.0Q0=0,Q,=5,Q
, = 10, Q3 = 15 Therefore, the pixel value P read from the scanner l is converted into the pixel value P° as shown in FIG. 7(a) using the following equation.

P≦4 → P'=Q, =0 5≦P≦7 → P'=Q, =5 8≦P≦to -p'=Q*=i01 l≦p
-P =03=+5 Here, the boundary value of the division of P is Q,',Q,°,Q, obtained as above,
', Q3' (for example, the average of Qio and Qt+'). The image data obtained by such conversion is shown in FIG. 7(b).

By doing this, the range of pixel values that was 2≦Vp≦13 before the processing was executed by the preprocessing circuit 12 is
By performing preprocessing, the range of pixel value distribution becomes 0≦Vp≦15, and the range of pixel value distribution can be widened and a large dynamic range can be obtained.

Next, the threshold determination circuit 4 determines the pixel value py=12 at the white peak and the pixel value P3=3 at the black peak.
Find the threshold value TH using .

In addition to finding the threshold value TH using the above formula, the threshold value determination circuit 4
It is also possible to find it using the lookup table 7 added to. That is, if the pixel value input to the look-up table 7 is less than the threshold TH, the output of the look-up table 7 becomes "0" (i.e., black display), and when other pixel values are input, the look-up table 7 The output is "15" (white display). The look-up table 7 operating as described above displays the threshold TH as a line thickness on the image display device 11, and while checking the displayed line thickness, selects an appropriate threshold TH.
is set.

In this way, when the threshold value TH is set, the edge detection circuit 5
As shown in FIG. 8, the output pixel value (Po) of the dynamic range adjustment circuit 3 is converted by setting the pixel value below the threshold TH to "0" and the other pixel values to "15", and converting it into binary image data. Create a pixel value (P'').

P"-o (p'≦TH) P"=15 (P'>TH) The edge detection circuit 6 first examines the binary image data consisting of the thus binarized pixel values line by line in the horizontal direction. As shown in FIG. 9, only a portion where two adjacent pixels exhibit different values (“0” and “15” or “15” and “0”) is detected as an edge portion.

Next, as shown in FIG. 10, the noise removal circuit 6 converts only the edge portion detected by the edge detection circuit 5 out of the binary image data into the output value of the dynamic range adjustment circuit 3 (i.e., in four gradations). (represented image data). In that case, the values of the binary image are left as they are except for the edge portions. Furthermore, the edge detection circuit 5 examines the image data thus obtained line by line in the vertical direction and performs similar processing (FIGS. 1i and 12).

Since the edge portions of the image information obtained in this way retain black and white image information, jagged edges are noticeable at the edge portions of the image displayed on the image display device II, as shown in FIG. First, since pixel values in flat areas (white background areas and black background areas) are binarized, granular noise is removed. In Figure 13,
The * portion 21 is a black portion corresponding to the pixel value “0” of the image data in FIG. 12, and the X portion 22 is an edge portion corresponding to the pixel value “5° or “” of the image data in FIG. 10
", and the other part 23 represents a white part corresponding to the pixel value "15' of the image data in FIG. 12.

Therefore, the encoding circuit 8
Therefore, when performing MR encoding or MH encoding, a high compression rate can be obtained without deteriorating image quality.

As described above, according to the image data encoding method of the present invention, deviations in the gradations of input image data due to the characteristics of the image input device 1 can be reduced to the entire range of gradations that can be expressed by the image display device 11. Since the correction is performed so as to be distributed, the influence of the characteristics of the image input device 1 can be removed. Further, according to the image data encoding method of the present invention, the edge portion of the binary image data that has been binarized based on the set threshold value is converted into a multi-gradation pixel output from the dynamic range adjustment circuit 3. Since the data is replaced with a value, noise caused by dirt on the original can be removed and encoding can be performed at a high compression rate.

In this embodiment, the input image data has 16 gradations, and the gradation represented by the image display device has 4 gradations, but it goes without saying that the present invention is not limited to this. .

<Effects of the Invention> As is clear from the above, the image data encoding method of the present invention includes a peak detection circuit and a dynamic range changing circuit, and detects the white peak of the density histogram based on the pixel values of input image data. The pixels of input image data are Since the values are converted, it is possible to correct the bias in the gradation of the input image data due to the characteristics of the image input device so that it is distributed over the entire range of gradations that can be expressed by the image display device. , the influence of the characteristics of the image input device can be removed.

The above-mentioned image data encoding method includes an edge detection circuit and a noise removal circuit, and binarizes the pixel values converted by the above-mentioned dynamic range change circuit.
According to the value image information, adjacent pixels exhibiting different pixel values are detected as edge portions, and the pixel values of the edge portions of the binary image information are replaced with pixel values converted by the dynamic range changing circuit. Therefore, it is possible to eliminate the influence of the characteristics of the image input device, and also to eliminate noise caused by dirt on the original.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of an image data encoding/decoding device according to the present invention, FIG. 2 is an explanatory diagram of changing the dynamic range, and FIG. 3 is an explanatory diagram of rewriting the lookup table.
Figure 4 is an explanation of the edge portion of the image, Figure 5 (a) is a diagram showing an example of a read original, and Figure 5 (b) is an example of image data obtained by reading the original in Figure 5 (a) with a scanner. Figure 6 is a diagram showing the histogram of the pixel values in Figure 5(b), Figure 7(a) is an explanatory diagram of changing the dynamic range of the pixel values in Figure 5(b), Figure 7 (b) is Figure 7(a)
Figure 8 is a diagram showing the image data with the dynamic range changed as shown in Figure 8 is the image data obtained by binarizing the image data in Figure 7 (b), Figure 9 is the horizontal direction obtained from Figure 8. Figure 1θ is a diagram showing the image data obtained by converting the pixel values of the horizontal edge part in Figure 8 to the pixel values in Figure 7(a), and Figure 11 is a diagram based on Figure 8. Figure 12 is a diagram showing the determined vertical edge part, Figure 12 is a diagram showing image data obtained by converting the pixel values of the vertical edge part in Figure 1O to the pixel values in Figure 7(a), and Figure 13 is 14 is a diagram showing an example of a density histogram of a black and white grayscale image; FIG. 15(a) is a diagram showing an example of a binary image; FIG. b) is shown in Figure 15 (
FIG. 3 is an explanatory diagram of an image in which an anti-area song has been applied to image a). I... Image input device, 2... Peak detection circuit,
3...Dynamic range adjustment circuit, 4...Threshold value determination circuit, 5...Edge detection circuit,
6... Noise removal circuit, 7"° torque up table 8... Encoding circuit, 9
...recording/reproducing device, io...decoding circuit, ll
...Image display device, 12...Preprocessing circuit. Patent applicant Shap Co., Ltd.

Claims (2)

[Claims] (1) In an image data encoding method in which black-and-white image data of documents, graphs, tables, etc. with predetermined gradations inputted from an image input device is encoded by an encoding circuit, the density is determined by the pixel value of the input image data. A peak detection circuit that obtains a histogram and obtains pixel values at two peaks corresponding to white and black in the histogram; while converting the pixel value at the black peak into a pixel value representing the other end of the gradation that can be expressed by the image display device, which is further determined by the peak detection circuit. The pixel value between the pixel value at the white peak and the pixel value at the black peak expressed by a predetermined procedure is a pixel value that expresses between the above one end and the other end in the gradation that can be expressed by the image display device. and a dynamic range changing circuit that converts the input image data into a dynamic range and outputs it to the encoding circuit, so that the gradation bias of the input image data is distributed over the entire range of gradations that can be expressed by the image display device through preprocessing. An image data encoding method characterized in that the influence of the characteristics of the input device is removed by modification. (2) In the image data encoding method according to claim 1, the pixel values converted by the dynamic range changing circuit are binarized based on a set threshold, and
an edge part detection circuit that detects, as an edge part, a location of adjacent pixels showing different pixel values according to the obtained binary image information; and a pixel in the edge part detected by the edge part detection circuit among the binary image information. It is equipped with a noise removal circuit that replaces the pixel value with the pixel value converted by the dynamic range changing circuit and outputs it to the encoding circuit, thereby removing the influence of the characteristics of the image input device as well as noise caused by dirt on the original, etc. An image data encoding method characterized by: