JPH0318229B2 - - Google Patents

Description

[Detailed description of the invention]

[Summary] The present invention relates to an area labeling method, in which provisional labels that are directly connected are detected and tracked, and provisional labels that are indirectly connected are detected at high speed. This update aims to speed up the area labeling process. [Industrial Field of Application] The present invention relates to a region labeling scheme. Generally, as shown in FIG. 7A, in a binary image represented by and, the background H portion is represented by , and the areas I ₁ , I ₂ , and I ₃ are represented by . However, it is not possible to distinguish between I ₁ , I ₂ , and I ₃ simply by expressing the area. Therefore, as shown in FIG. 7B, if the three areas that are expressed in the same way are labeled as 1, 2, and 3, a clear distinction can be made between them. For example, if each of the illustrated regions is a chromosome and its area is to be determined, each chromosome is labeled 1, 2, and 3 to distinguish them in advance, as shown in Figure 7B, and a histogram of the constituent pixels is created. If you do the calculations, you can easily find the area of each area. The present invention relates to a labeling method that makes it possible to clearly identify which of the regions each region is represented by a binary image. [Problems to be solved by the prior art and the invention] The conventional labeling method adds a temporary label to each small area, then detects the connection relationship between them, and assigns serial numbers to continuous areas formed by the small areas. The method was to change and add a temporary label consisting of: However, when converting into the serial number, clustering of provisional labels is performed, but the conventional clustering method requires many arithmetic processing steps. For this reason, conventional labeling methods have had the problem of being difficult to execute at high speed. [Means and effects for solving the problems] The present invention aims to solve the above problems and achieve high-speed labeling by utilizing the property that the connection relationships of temporary labels constitute a planar graph. , the means consists of a temporary label attaching section, a label organizing section, a directly connected label detecting section, a label tracking section, and a label updating section, and by this means, the amount of calculation required for clustering processing of temporary labels can be reduced. High-speed labeling processing becomes possible. [Example] Hereinafter, the present invention will be described by way of example with reference to the accompanying drawings. FIG. 1 is a block diagram of the system of the present invention, which is composed of a temporary label attaching section 1 ¹ , a label organizing section 2 ¹ , a directly connected label detecting section 3 ¹ , a label tracking section 4 ¹ and ^a label updating section 5 ¹ . ing. Temporary label attachment part 1 above ¹ is installed by window 2
This is a device that temporarily attaches a label with an initial value of 1 by scanning a value image. The label organizing section ²¹ is a device that organizes overlapping parts among the connections of the temporary labels. The direct label detection unit 3 ¹ is a device that detects all small areas to which each small area is directly connected from the above-described organized connection relationships and creates a direct connection table to be described later. The label tracking unit ⁴¹ is a device that tracks labels based on the contents of the direct connection table and detects all temporary labels that are directly or indirectly connected. Furthermore, the label updating unit 5 ¹ is a device that changes the temporary label based on the label tracking result. Hereinafter, the operation of the apparatus shown in FIG. 1 having the above configuration will be explained based on FIGS. 2 to 6. The temporary label attaching unit ¹¹ scans the image G in the α direction using the window W shown in FIG.
Labeling is performed in the j and i directions. The window W in FIG. 2 has three windows designated by a, b, and c, and the logical operations thereof are as shown in Table 1.

[Table] In Table 1, the "input" column is in windows a, b, and c.
The ``operation'' column represents the temporary label attached to each part of the area, that is, the small area, by the window c. Among the inputs a, b, and c above, * is undefined, NZ
indicates a value other than zero. The temporary label attaching unit 1 ¹ attaches a temporary label to the small area using the logical operations shown in Table 1 above.
If b = *, c = 0 (left diagram of Figure 3 A), c
=0 (right side of FIG. 3A), ie, no label is attached. If a=0, b=0, c≠0 (left side of FIG. 3B), a label of c=L, that is, initial value L=1 is provisionally attached (right side of FIG. 3B). This is where the label appears for the first time. If a=0, b≠0, c≠0 (left side of FIG. 3C), c=b=L, that is, the same label L=1 as before is given to c and b. This is the same in the case of a≠0, b=0, and c≠0, and c=a=L, that is, the same label L=1 as last time is attached to c and a. Next, if a, b, and c are all non-zero and the window W completely enters the area I (left side of Figure 3 D), change the attached temporary labels L and L' by a and b. c = Min (a, b), that is, the smaller one, for example, L, is attached as a temporary label (Fig. 3D
right side). Next, the temporary label attaching unit 1 ¹ detects the connection relationship of the temporary labels attached to each small area. The arithmetic logic in this case is as shown in Table 2.
Detection of this connection relationship is carried out by, for example, as shown in FIG.
Detect the connection relationship when L, b=L' is installed.

〔Effect of the invention〕

As described above, according to the present invention, without performing clustering of provisional labels as in the past, provisional labels added to each small area are detected and tracked, connection relationships are detected, and the above-mentioned By changing the temporary label and adding the same temporary label, it has become possible to perform the temporary labeling process at high speed.

[Brief explanation of the drawing]

Fig. 1 is a block diagram of the method of the present invention, Fig. 2 is a block diagram of a window used in the method of the present invention, and Figs.
D is an explanatory diagram of the operation of the temporary label attaching part, and Fig. 4 A~
C is an explanatory diagram of the operation of the label sorting section, FIGS. 5A and B are explanatory diagrams of the operation of the directly connected label detection section, and FIGS. 6A,
B is an explanatory diagram of the operation of the label tracking unit and update unit, 7th
Figures A and B are explanatory diagrams of industrial application fields. 1 ¹ ... Temporary label attaching section, 2 ¹ ... Label organizing section, 3 ¹ ... Direct connection label detection section, 4 ¹ ... Label tracking section, 5 ¹ ... Label updating section.

Claims (1)

[Claims] 1 In a labeling method in which a label is attached to the above region of an image in which the background and the region can be represented by binary values, the above image is scanned, a video signal is input, and a temporary image is created for each small region in response to the input signal. a temporary label attaching unit that attaches the label and detects connection relationships with existing nearby temporary labels; a label organizing unit that generates connection relationships that organize overlapping parts of the connection relationships; and the organized connection relationships. A direct connection label detection unit that generates a direct connection table based on , a provisional label given to an area in the column (row) direction, and a collection of all provisional labels that are directly connected to the provisional label in the row (column) direction. a direct connection table, a label tracking unit that searches the above direct connection table and generates a group of mutually connected temporary labels, and changes the temporary label of each group and adds the same label to each group. 1. A region labeling method comprising a label updating unit that updates a label.