JPS63301374A - Vanishing method for background of image - Google Patents

Abstract

PURPOSE:To obtain inter-gradation conversion coefficients corresponding to an image and its background image and to simplify a background vanishing operation, by separating a partial background image from the image and extracting a proper number of aggregates of image elements out of said partial image to analyze the gradation of the corresponding image element via the extracted aggregates of image elements. CONSTITUTION:An image A where a background is superposed on a graphic image is recorded in an image memory 4 and an image B including a background only is recorded in an image memory 11. A threshold value calculating circuit 14 obtains a threshold value to separate a partial background image from the image A and inputs this threshold value to a comparator 16. At the same time, a random number produced from a random number generating circuit 10 is sent to an address generating circuit 9 for generation of the address of the memory 4. Then the image elements of the image A are extracted at random. The gradation (ga) of the extracted image element is sent to the comparator 16 and a background part if decided is stored in a regression analyzing data buffer 17. At the same time, the gradation (gb) of the image element of the image B is also stored in the buffer 17.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to the processing of graphic images, and particularly to a technique for erasing a background image superimposed on a graphic image.

[Conventional technology]

Generally, CCD (charge coupled de
(vice) When trying to capture the fluorescence emitted by a substance attached to an object with a camera, etc., the pattern on the surface of the attached object is inevitably superimposed as a background.

For example, in a fingerprint detection device using a laser fluorescence method, which irradiates a fingerprint-attached object with a laser beam and detects the fingerprint by the fluorescence emitted by the fingerprint, when trying to take a fingerprint image, the pattern of the fingerprint-attached object inevitably appears. will also be superimposed and photographed as a background image.

An example of a device equipped with means for erasing a background image in such a case is a fingerprint detection device disclosed in Japanese Patent Laid-Open No. 61-240383. The device is equipped with a laser beam emitter, an image receiver, and an image storage processor that adds or subtracts images from the image receiver. After enhancing the image by adding , the laser light irradiation is stopped,
It is now possible to perform subtraction processing and erase the background image.

However, since the illumination state of the target object differs depending on whether or not the laser beam is irradiated, many trials or errors are required to obtain an appropriate number of additions and subtractions.

For example, if an image in which a background image is superimposed on a fingerprint image is added m times, they are At(t = 1°..., m)
Assuming that the background image is subtracted n times from the added image, let BE (t = 1° . . . , n). The average level of the background image in Ai is f (At
), and if the average level of the image in the corresponding part is f (Bt ), then in order to erase as much of the background image as possible through addition and subtraction processing, find the pair of m and n that minimizes e. Is there a need to find out?

Even if m is practically set to about 20 to 30, subtraction must be repeated dozens of times to find the corresponding n, which requires a lot of trial and error. Furthermore, excessive subtraction at that stage may damage the graphic image.

[The problem that the invention attempts to solve]

When a CCD camera or the like attempts to capture the fluorescence emitted by a substance attached to an object as a figure, the pattern on the surface of the attached object is inevitably superimposed as a background.

The present invention aims to eliminate the background in such a case and obtain a clear image.

The conventional method for erasing such a background image is to add an image with the background image superimposed on the figure image multiple times, and then repeatedly subtract the background image only.Is there a way to obtain an appropriate number of additions and subtractions? This required a lot of trial and error, and could sometimes damage the graphic image.

An object of the present invention is to perform optimal background erasure in a single operation without relying on such trial and error.

(Means for Solving the Problems) The present invention solves the above-mentioned problems, and as shown in the flowchart of the procedure in FIG. (denoted as image A) and
A background erasing method for extracting only figures to be extracted by erasing the background from a grayscale image of only the background (hereinafter referred to as image B) consists of the following steps.

(b) Separating a portion of the background image from image A.

(b) Extracting a set C of an appropriate number of pixels from the separated partial image of the background image.

(c) Extracting a set of pixels corresponding to set C in image B.

(d) Performing a regression analysis on the gradations of pixels corresponding to the set C and the set, and determining the variable coefficient between the gradations of the background part of the image A and the corresponding part of the image B from the result.

(e) Subtracting image B from image A while correcting image B using the conversion coefficient.

Next, as an embodiment showing preferable specific means of the present invention, the means for separating a partial image of the background image from the image A sets a threshold at the image level, and the portion below the threshold is regarded as the background. It is preferable to use the difference in texture between the figure part and the background part, and in some cases, it is also possible for the operator to specify a part of the image.

(Operation) FIG. 1 is a flowchart showing the procedure of the present invention. In the present invention, it is necessary to separate a partial image of the background image from an image (hereinafter referred to as image A) in which a background image is superimposed on a graphic image.

When a background is superimposed on a figure, it is difficult to completely separate only the background part, but in the present invention, a partial image of the background image is separated.

Usually, substances attached to an object have a higher gradation and are brighter than the background due to fluorescence, so it is possible to separate partial images of the background image by setting a threshold and treating the area below that value as the background. can. At this time, there may be a portion in the background image that has a high gradation and exceeds the threshold value, and that portion may be excluded.

Next, a set C of an appropriate number of pixels is extracted from the separated partial image of the background image. This is because the effective number of pixels from a CCD camera is usually tens to hundreds of thousands, and the aforementioned partial images also have too many pixels and it is inappropriate to use all of them for regression analysis, so the accuracy This is because it is necessary to extract enough pixels to perform regression analysis well. The set C needs to be extracted so as to include parts of various gradations of the background image as evenly as possible, but this is done by, for example, using random numbers to randomly extract pixels from partial images of the background image. be able to.

Next, a set of pixels at positions corresponding to set C is extracted from a grayscale image of only the background (hereinafter referred to as image B), and regression analysis is performed on the gradation of the corresponding pixels. A pixel C1 belonging to a set C, whose gradation is gc, a pixel d of the set corresponding to C, whose gradation is gd, and the relationship between gc and gd is gc = f (gd) +e ....- (1)
Find in the form of. Here, e is an error, and the parameters of the function f(x) are determined so as to minimize this. function f
For example, the form of (x) is f (x) = px + q
・・Fist・−(2) can be assumed.

While correcting image B using the obtained f (x), image A
and all corresponding pixels of image B. That is, a pixel belonging to image A is a, its gradation is ga, the corresponding pixel of image B is gb, the image after the subtraction process is H, its pigment is h, its gradation is gh, and gh = g.
a − f (gb) ...The operation in (3) is performed for all pairs of a and b.

The background part is from (1) gh = gc - f (gci) = e =O...(4) Thus, the background image can be erased.

In other words, the present invention minimizes the average value of the gradations of the background part of the processed image by matching the average value of the gradations of the background part of screen A and the corresponding part of screen B, and then performing subtraction. The background can be erased in a single operation, which is optimal in the sense that it minimizes the influence on the graphical image that is of interest.

Regarding the separation of the partial image of the background image from image A, the difference between the figure part, the background part, and the texture is used.
It is also possible to perform a known texture analysis and use the results. Texture here refers to a regular pattern on the surface of an object, such as the grain of wood or the stitching of textiles. Furthermore, as shown in Fig. 5, if the proportion of the area occupied by one graphic image in image A is small and there is a part that can be clearly identified as only the background image, the operator can remove the background part from image A. For example, a cursor line,
It is also possible to separate partial images of the background image by specifying them using a known method such as a light pen.

〔Example〕

FIG. 2 is a diagram showing an example of the configuration of a background image erasing device suitable for implementing the present invention.

An embodiment will be described based on FIG.

In FIG. 2, 6 is a CCD camera, 7 is an A/D converter, and 8 is an input switching circuit. First, an image (image A) in which a background is superimposed on a graphic image is recorded in the image memory 4. An image containing only the background (image B) is recorded in the image memory 11. The address generation circuit 9 functions as an incremental counter when performing sequential writing or reading.

The threshold calculation circuit 14 calculates a threshold for separating a partial image of the background image from the image A. As the threshold value, for example, one calculated by Otsu's discriminant threshold method, which is a well-known threshold setting method when binarizing a grayscale image, can be used.

The calculated threshold value th is sent from the output terminal T to the comparator circuit 16.

The random numbers generated by the random number generation circuit 10 are sent to the address generation circuit 9, which randomly generates addresses for the image memory 4, and randomly extracts pixels of the image A.

The extracted pixel level 21gga is sent to the comparison circuit 16, and if ga is smaller than th, it is considered to be a background part, and stored in the regression analysis data buffer 17, and the gradation of the corresponding pixel of image B is gb is also sent to 17 and stored.

Such data extraction is performed until the buffer 17 is filled. After data extraction is completed, regression analysis circuit 1
9, regression analysis is performed.

This example assumes a function of the form y=px+q, where the transformation coefficient p between the background part of image A and the corresponding part of image B is
, q are connected to the multiplier 12 and adder 1 from the output terminals P and Q, respectively.
Then, the gradations of the pixels are read out sequentially from the image memories 4 and 11, and the value from the memory 11 is corrected by the multiplier 12 and adder 13, and then subtracted by the subtracter 5. Erase the background.

The processed value is recorded in the image memory 4 or 11 via the switching circuit 8.

The output switching circuit 1 selects the signal from the CCD camera 6 and the signal from the image memory 4.11, and this signal passes through the D/A converter 2 and is displayed on the monitor 3.

The timing controller 15 is for sending control signals to each circuit in order to perform the above-described processing in an orderly manner.

Reference numeral 18 denotes an operation switch, and after recording images A and B by operating this switch, the background can be erased by operating a background processing switch.

Also, images A and B do not need to be written once, S/
In order to improve N, an adder may be provided to add the output from the A/D converter 7 several times.

Further, the number of times of addition at that time may be arbitrary as long as the image memories 4 and 11 do not overflow;
It doesn't matter if the number of times is different.

Further, FIG. 3 is a diagram showing another embodiment of the present invention.

3, the difference from FIG. 2 is that a texture analysis circuit 20 is used instead of the threshold calculation circuit 14 and comparison circuit 16.
, background image mapping memory 21 , background portion determination circuit 22
The point is that this is provided. In the embodiment shown in FIG. 3, after recording one image A and one image B, a texture analysis circuit 20 performs texture analysis on one image A. As a result, whether or not each pixel of image A belongs to the background portion is stored in the background image mapping memory 21.
Next, the random number generation circuit 10 generates a random number and sends it to the address generation circuit 9, which randomly generates an address for the image memory 4. This address is sent to the background portion determination circuit 22, where the background image mapping memory 21 is referred to, and if the address corresponds to a pixel in the background portion, the gradation ga of that pixel is stored in the data buffer for regression analysis. The gradation gb of the pixel of the corresponding image B is also sent to the data buffer 17 and stored therein. The following operations are completely similar to the embodiment shown in FIG.

Moreover, FIG. 4 is a diagram showing still another embodiment of the present invention. In FIG. 4, the cursor generation circuit 23 and the addition circuit 2
4, the operator can operate the operation switch 18 to generate a cursor line at any position on the vertical and horizontal axes of the screen of the monitor 3, thereby specifying the position on the image. It looks like this.

As shown in FIG. 5, when the image A occupies a small proportion of the figure image and there is a portion that can be clearly identified as the background, the operator operates the cursor line using the operation switch 18, and first As shown in Figure 6, by specifying the upper left coordinate C of the figure image using the intersection of the vertical and horizontal cursor lines, and then specifying the lower right coordinate C of the figure image as shown in Figure 7, line segment CD can be drawn on one screen. It can be specified that the inside of a rectangle with one diagonal is the part that contains the graphic image, and the outside is the part that contains only the background.

Therefore, after recording images A and B, the operator can operate the operation switch 18 and while observing image A on the monitor 13, specify the background portion using the method described above.

Next, a random number is generated by the random number generation circuit 10, and sent to the address generation circuit 9, which randomly generates an address for the image memory 4. This address is sent to the address determination circuit 25, which refers to the data from the cursor generation circuit 23, and if the address corresponds to a pixel in the background part, the gray level ga of that pixel is determined by the regression analysis data buffer 17. The gradation gb of the pixel of the corresponding image B is also sent to the data buffer 17 and stored therein. The following operation is the second
It is completely similar to the embodiment shown in the figure.

The embodiments shown in FIGS. 2, 3, and 4 differ only in the method of separating a partial image of the background image from image A.
Since the other parts are common, two or more of these can be combined to form a single device, and a suitable background image separation method can be selected in response to various conditions of the image A. It is also possible to

〔Effect of the invention〕

As described above, the present invention provides a grayscale image in which the background is superimposed on a figure image (both images A) and a grayscale image with only the background (image B).
), we calculate the conversion coefficients of the background part of image A and the corresponding part of image B, and subtract image B from image A while correcting image B with the coefficients. repeating many mistakes like
Optimum background erasure can be performed in a single operation without damaging the graphic image in the process.

[Brief explanation of drawings]

FIG. 1 is a flowchart showing the procedure of the method of the present invention;
3 is a block diagram of an example of a background image erasing device suitable for implementing the present invention, FIG. 3 is another block diagram of a background image erasing device suitable for implementing the present invention, and FIG. 4 is a block diagram for implementing the present invention. A block diagram of yet another example of a suitable background image erasing device, FIG. 5 is an image diagram showing an example of an image in which a background image is superimposed on a figure image, and FIG. 6 is an image diagram showing an example of an image in which a background image is superimposed on a figure image. , an image diagram showing an example of designating the upper left coordinates of a graphic image, and FIG. 7 is an image diagram showing an example of designating the upper right coordinates of a graphic image for an image in which a background image is superimposed on the graphic image. l...Output switching circuit 2...D/A converter 3...Monitor 4...Image memory (for image A) 5...Subtractor 6...COD camera 7...A/ D converter 8...Manual switching circuit 9...Address generation circuit 10...Random number generation circuit 11...Image memory (image B Jfl) 12...
Multiplier 13... Adder 14... Threshold output circuit 15... Timing controller 16... Comparison circuit 17... Regression analysis data buffer 18... Operation switch 19... Regression analysis circuit 20...texture analysis circuit 21...background image mapping memory 22...background portion judgment circuit 23...cursor generation circuit 24...addition circuit 25...address judgment circuit 26...・Graphic image 27...Background image 28...Horizontal cursor line 29...Vertical cursor line

Claims (1)

[Claims] 1. A grayscale image in which the background is superimposed on the figure to be extracted (hereinafter referred to as image A), and a grayscale image with only the background (hereinafter referred to as image B).
In an image background removal method that removes the background and extracts only the figure to be extracted, (a) separates a partial image of the background image from image A, and (b) extracts an appropriate number of images from the partial image. (c) extract a set D of pixels corresponding to set C in image B; (d) perform regression analysis on the gradation of corresponding pixels in set C and set D; From the results, find a conversion coefficient between the gradations of the background part of image A and the corresponding part of image B, and (e) subtract image B from image A while correcting image B using the conversion coefficient. A featured method for removing backgrounds from graphic images. 2. The figure according to claim 1, wherein the means for separating a partial image of the background image from the image A is to set a threshold at the image level, and set a portion below the threshold as the background. How to erase the background of an image. 3. The background erasing method for a graphic image according to claim 1, wherein the means for separating the partial image of the background image from the image A is to utilize a difference in texture between the graphic part and the background part.