JP3045555B2 - Binarization processing method for multi-tone image information - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for binarizing multi-tone image information read by an apparatus having an image input function, such as an image reader or a facsimile apparatus. The present invention relates to a binarization processing method for improving information deterioration.

[0002]

2. Description of the Related Art With the development of multimedia such as OA equipment,
The importance of storing and transmitting image information has been increasing in recent years. In general, the storage capacity and transmission capacity of a system for handling image information are generally not enough compared to the amount of image information, and image information must often be compressed and handled.

There are various methods for compressing image information depending on the purpose and the nature of the image information. One of them is to express the gradation of the original image in two gradations (see FIG. 4). The specific example is shown). Expressing an image in two tones is
There is an advantage that not only the amount of information can be compressed but also an image output device that can handle only two gradations, such as a dot printer or a liquid crystal display device, can easily output.

[0004] In the binary representation of an image, there is an appropriateness depending on the type of the image. An image such as a character or a drawing can be originally expressed in binary, and there is often no problem in expressing it in binary. However, a person image generally contains a lot of important information in a wide range of levels of multiple gradations, and when expressed in binary form, it is usually unnatural.

[0005] In particular, considering the expression of a face, a binarization threshold suitable for expressing an expression within the range of change of the tone value inside the face, and a binarization threshold value suitable for expressing the face and the background within the range of change of the tone value. It is usually different from a threshold suitable for representing the face contour at. For example, if an expression is appropriately expressed, the whole is dark and the outline is unclear, and if the outline is to be clearly expressed, the expression tends to fly brightly.

As described above, when the absolute value of the gradation is binarized using a single threshold value, important information for a human image such as expression of a facial expression and expression of a face outline can be expressed in a compatible manner. Can not do it.

[0007]

According to the present invention, in the conventional binarization processing of a multi-tone image, important contour information is sometimes ignored. An object of the present invention is to allow important outline information to be taken into binary information without being ignored.

[0008]

According to the present invention, there is provided a method for binarizing multi-tone image information according to the present invention. ) And a change in the input multi-tone image information is detected, and a portion where the amount of change is larger than a predetermined value is determined and extracted as an outline of the image, and the floor of the extracted outline is extracted. Binarize the tonal value with the relative magnitude of the magnitude, and according to the direction of the change in the tone value
The outline information consisting of a row of black pixels and white pixels or white
And Guo information extraction processing means for extracting contour information consisting of a sequence of pixels and black pixels (5), is the relatively binarized black screen
Contour information consisting of a row of pixels and white pixels or white pixels and black
Compositing that combines the binarized image information and the binarized outline information by replacing the binarized image information with the binarization processing means using the contour information composed of the arrangement of pixels. Means (6) are provided, and the inputted multi-tone image information is contour-enhanced and binarized.

FIG. 1 shows the basic configuration of the present invention. FIG.
In the figure, reference numeral 1 denotes an image input device, which reads an image using an image pickup device such as a CCD and outputs it as multi-tone image information in a digital signal format.

Reference numeral 2 denotes an image information file which stores multivalued image information and is read out as needed. An image processing device 3 performs image processing, display, printing, storage, transmission, and the like. Reference numeral 4 denotes a binarization processing unit, which binarizes the input multi-tone image information with a set single threshold and outputs the binarized image information.

Reference numeral 5 denotes contour information extraction processing means for comparing the tone values of two adjacent pixels in the input multi-tone image information, and determining the difference between the tone values of the two unmatched pixels, It is determined whether or not the amount of change is greater than a predetermined threshold value. If the amount of change is greater than the threshold value, it is determined to be a contour, and the tone values of the two pixels in the pair are binarized by a relative magnitude. Output.

Reference numeral 6 denotes a synthesizing means, which is a binarization processing means 4
And each output of the contour information extraction processing means 5 is synthesized for each pixel. In the combining method, for example, when both output values are equal, the output value is output, and when the output values are different, the output value of the contour information extraction processing means is selected and output.

[0013]

The operation of the present invention will be described with reference to FIG. FIG.
(A) is input multi-tone image information, and shows an example of the tone value of each pixel arranged on a certain scanning line.

The binarization processing means 4 shown in FIG. 1 binarizes the multi-tone image information shown in FIG. 2 (a) by a threshold value E _S0 , and sets a value larger than E _S0 to “1” and a value smaller than E _S0 to “0”. "Is output. The contour information extraction processing means 5 shown in FIG. 1 checks whether the change amount of the gradation value between adjacent pixels is larger or smaller than E _S1 , and a pair of pixels (P _l , P _{l + 1} ) having a larger change amount is determined. (P _m , P
_{m + 1} ) and ( _Pn , _{Pn + 1} ) are detected.

FIG. 2 (b) shows a binarized version of this.
The value of a pixel closer to white is converted to “1”, and the value of a pixel closer to black is converted to “0”. 2C shows the output of the binarization processing means 4, FIG. 2D shows the output of the contour information extraction processing means, and FIG. 2F shows the binarization result of the synthesis by the synthesis means 6. Indicates image information.

In this way, a black line corresponding to the outline is formed in the white area of the binarized image information, and a white line is formed in the black area, so that loss of the information on the outline can be prevented.

[0017]

FIG. 3 shows the configuration of an embodiment of the present invention. Although this embodiment is shown as a hardware configuration, it goes without saying that the same logic can be easily realized on software.

In FIG. 3, reference numeral 7 denotes an image memory for storing multi-tone image information. Here, the data width of the multi-tone image information is 1 byte per pixel, and 256 gradations can be expressed. The pixel address is represented by Xi in the horizontal direction and by Yj in the vertical direction. The pixel at the address Xi, Yj is represented by Pi _, j, and its pixel value is represented by Di, j. The adjacent pixel in the X direction is Pi + 1, j, and the pixel value is Di + 1, j.

Reference numeral 8 denotes an address generator. 9 is Xi
An X address counter for generating an address. 10
Is a Y address counter for generating a Yi address. Reference numeral 11 denotes a binarization processing unit.

A comparator 12 binarizes Dij with a threshold value E _S0 , and outputs Bij = 1 when Dij ≧ E _S0 as a binary output, and outputs Bij = 0 when Dij <E _S0. . 13
Is a contour extraction processing unit. Reference numeral 14 denotes a Dij register in which a pixel value Dij is set.

Reference numeral 15 denotes Di + for setting a pixel value Di + 1, j.
1, j register. Reference numeral 16 denotes a subtractor for detecting a difference (change amount) Dij-Di + 1, j between Dij and Di + 1, j. 17
The subtraction result (Dij-Di + 1, j ), and compares the threshold value _{E S1, Dij, Di + 1} , j each binary output Mij, Mi + 1, j
Is a binary logic circuit. Where | Dij−Di +
1, j | a ≧ E _S1, if Dij> Di + 1, j, if Mij = 1, Mi + 1, j = 0 Dij <Di + 1, j, Mij = 0, Mi + 1, j = 1 is output.

Further, | Dij-Di + 1, j | < in the case of E _S1 is, Mij, Mi + 1, j output is disabled. 18 is
A Mij register for storing Mij. 19 is Mi + 1, j
Is a Mi + 1, j register.

Reference numeral 20 denotes Bij and Mij, Bi + 1, j and Mi + 1, j
Are respectively combined, and the corresponding Bij, Bi + 1, j are replaced with Mij, Mi + 1, j. As a result, Bij ≠ M
ij or Bi + 1, j ≠ Mi + 1, j, then Bij,
Bi + 1, j is replaced by Mij, Mi + 1, j, and Bij = M
If ij or Bi + 1, j = Mi + 1, j, Bij,
Bi + 1, j is not changed.

The Xi counter 9 of the address generator 8 counts the clock and scans the address in the X direction. When the count reaches a full count, the Yi counter 10 counts up by one and advances the address by one in the Y direction. The addresses Xi and Yj of the pixels Pij generated continuously in this manner.
Is applied to the image memory 7, the Pij is scanned, and the pixel value Dij is read. Dij is calculated by the binarization processing unit 11 as 2
The binarized contour information Mij, Mi + 1, j is obtained by detecting a sudden change in the contour extraction processing unit 13, and the binarized contour information Bij, Bi + 1, j is obtained by the synthesizing unit 20. They are combined and output.

The value of the threshold value E _S1 for determining the extraction level of the rapidly changing portion in the multi-tone image information can be set arbitrarily. The extraction of the outline information from the multi-valued image information is not limited to the X direction, but can be performed two-dimensionally in both the X and Y directions.

[0026]

According to the present invention, important outline information can be left in the binarization of a multi-tone image such as an image that is entirely whitish, an image that is entirely black, or a photograph that includes various tones. The useful information amount in the binarized image information is increased, and the usefulness of the image information is significantly improved.

[Brief description of the drawings]

FIG. 1 is a diagram showing the basic configuration of the present invention.

FIG. 2 is an explanatory diagram showing a specific example of a binarization process for explaining an operation of the present invention.

FIG. 3 is a configuration diagram of an embodiment of the present invention.

FIG. 4 is an explanatory diagram of a specific example of a conventional binarization process.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 image input device 2 image information file 3 image processing device 4 binarization processing device 5 contour information extraction processing means 6 synthesis means

Continuation of the front page (56) References JP-A-60-68774 (JP, A) JP-A-61-133774 (JP, A) JP-A-1-134683 (JP, A) JP-A-1-281575 (JP) , A) JP-A-2-185456 (JP, A) JP-A-62-285078 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G06T 1/00 G06T 5/00 H04N 1/409

Claims (3)

(57) [Claims] 1. A binarization processing method for multi-tone image information, comprising: a binarization processing means (4) for binarizing input multi-tone image information with a single threshold; A change in the multi-tone image information is detected, and a portion where the amount of change is larger than a predetermined value is determined and extracted as a contour portion of the image, and the tone value of the extracted contour portion is relatively large or small. In the direction of the change in the gradation value
Depending on the type of contour information or black pixel and white pixel
A wheel <br/> Guo information extraction processing means for extracting contour information consisting of a sequence of white pixels and black pixels (5), it of a sequence of the relatively binarized, black and white pixels
The binarized image information is replaced with the binarized image information by replacing the binarized image information by the binarization processing means or the outline information composed of the arrangement of white pixels and black pixels. And a synthesizing means (6) for synthesizing the input multi-tone image information, wherein the multi-tone image information inputted is edge-enhanced and binarized. 2. The multi-tone image according to claim 1, wherein a change point in the multi-tone image information is detected one-dimensionally between adjacent pixels along a scan line of the image. Information binarization processing method. 3. The binarization processing of multi-tone image information according to claim 1, wherein the extraction of a change point in the multi-tone image information is performed two-dimensionally between vertically and horizontally adjacent pixels. Method.