JPH012178A - Color image region extraction method - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to an image processing method for extracting a specific target area of a color image using color information.

(Prior art) Conventionally, as a method for extracting a specific target area from a color image,
For example, there is a method of calculating the distribution frequency of each of the three primary colors R, G, and B, determining the upper and lower thresholds of the multivalue, and extracting a specific color area.

(Problems to be Solved by the Invention) In the conventional method, for example, the histograms of the three primary colors R, G, and B of a color image are calculated to determine the upper and lower thresholds for each. It was difficult to separate regions with overlapping distributions.

FIG. 4 shows an example of image distribution for two colors on the R and G axes. In FIG. 4, the lower diagram of the drawing on the R-0 plane is the frequency distribution projected onto the G axis, and the left diagram is the frequency distribution projected onto the R axis. In the case of Figure 4, R and G
It is impossible to set a threshold value based on the frequency distribution projected onto the axis and isolate a specific target area of the image, so it is necessary to change the axis of projection or sequentially limit the range of the R or G axis to examine the distribution. Even if a method of determining a threshold value is used, extraction of a specific region cannot be performed accurately for an image that includes many objects and has a complex color distribution, such as a natural image.

SUMMARY OF THE INVENTION In view of the above-mentioned shortcomings of the prior art, an object of the present invention is to interactively specify the distribution range of the extraction target region from the distribution state of a plurality of two-dimensional histograms.
It is an object of the present invention to provide a color image region extraction method that can easily extract a target region even in a color image of a complex scene.

(Means for Solving the Problems) To achieve this object, the color image region extraction method of the present invention provides a method for extracting a region of interest from a color image. The two-dimensional histogram of the displayed color image is displayed, and the corresponding position on the two-dimensional color plane of the representative point in the extraction target area of the displayed color image is determined on the plane by referring to the plurality of displayed two-dimensional histograms. This method is characterized by specifying a threshold value and interactively extracting the specific target area from an image of a complex scene.

(Example) The method of the present invention will be described in detail below with reference to the accompanying drawings.

FIG. 1 shows a block diagram of an embodiment of a hardware configuration suitable for implementing the region extraction method of the present invention.

An image input device 1 inputs a color image to the device and outputs an R component,
Write the G component and B component to the image memory 2.

The two-dimensional histogram creation means 3 sequentially outputs pixel data from the image memory 2 pixel by pixel, calculates the distribution frequency of the combination of two colors of pixels, and writes it into the two-dimensional histogram memory 4. The display device 5 displays the contents of the image memory 2 and the two-dimensional histogram memory 4.

An example of a two-dimensional histogram display is shown in FIG.

In FIG. 2, the frequency in each chart is shown by the density of diagonal lines, but on an actual display device, it is shown by brightness. The pointing device 6 refers to the color image and two-dimensional histogram displayed on the display device 5, and
The representative point of the image and the two-dimensional threshold are specified. In Figure 2, this threshold value can be specified, for example, in the dashed line area 11.12.
．． 13, respectively. The two-dimensional threshold memory 7 stores the two-dimensional threshold specified by the pointing device 6. The color coordinate detection means 8 calculates the position on the two-dimensional color plane corresponding to the representative point of the image designated by the pointing device 6, and extracts an image area according to the contents of the two-dimensional threshold memory 7. The CPU 9 controls various devices according to control programs stored in the program working memory 10.

Next, a flowchart of a region extraction method related to the method of the present invention will be explained using FIG.

In step 21, a two-dimensional histogram is calculated from the color image stored in the image memory 2. Next step 2
2 with reference to the image displayed on the display device 5,
The pointing device 6 is used to indicate the position coordinates of one or more representative points in the extraction target area. In step 23, R and G of the representative points are determined.

An example will be given in which the colors displayed at the corresponding positions on the two-dimensional histogram displayed on the display device 5 are changed based on the value of B. Next, referring to the two-dimensional histogram and the position of the illustrated representative point on the color plane, in step 24 the pointing device 6 indicates a two-dimensional threshold value, and in step 25 the color coordinate detection means 8 detects the indicated range. Extract the included pixels from the image.

If the area extracted by the above operation is unsatisfactory, or if there is a target area to be specified other than the already extracted area, the process returns to step 22 and the extraction is performed again.

In the above example, the three primary colors of the image, R, G, and B, were used, but two-dimensional ≠ coordinate system with better separation on the stogram,
Alternatively, this operation may be performed by converting to a coordinate system that has good correspondence with brightness, saturation, and hue.Also, although the example using three types of two-dimensional histograms was shown, it is also possible to convert this to two types. good.

(Effects of the Invention) As described above, by using the extraction method of the present invention, a two-dimensional histogram is created for the color information of an image, and the distribution on the two-dimensional color surface and the information specified by the operator are The position on the color plane corresponding to the representative point of the extraction target area is displayed and the z area is extracted by interactively specifying a two-dimensional threshold, making it suitable for images with complex scenes. This has the effect that the target object can be extracted effectively and correctly.

[Brief explanation of the drawing]

FIG. 1 shows a block diagram of an embodiment of a hardware configuration suitable for implementing the method of the present invention, FIG. 2 shows an example of a two-dimensional histogram and its threshold value related to the present invention, and FIG. The figure shows a processing flowchart related to the present invention, and FIG. 4 shows an example of the distribution of a color image on the envelope surface according to a conventional example. 1... Image input device 2... Image memory 3...
- Two-dimensional histogram creation means 4... Two-dimensional histogram memory 5... Display device 6... Pointing device 7... Two-dimensional threshold memory 8... Color coordinate detection means 9... CPU 10... Working memory program Figure 4 → G

Claims (1)

[Claims] 1. When extracting a specific target area from a color image, display the color image and a plurality of two-dimensional histograms regarding the color components of the color image, and calculate two representative points within the extraction target area of the displayed color image. The specific target area is interactively extracted from an image of a complex scene by instructing a threshold value on the plane by referring to the corresponding position on the dimensional color plane and the plurality of displayed two-dimensional histograms. A color image region extraction method characterized by the following.