JPH09153135A - Method for judging specified symmetrical object - Google Patents

Abstract

PROBLEM TO BE SOLVED: To easily judge a specified symmetrical object without using complicated filtering processing. SOLUTION: In a step 1, a representative picture for a specified symmetrical object is set up as dictionary picture data and correlation between the representative picture and a picture inputted from a picture input means is found out to extract an area to be a candidate of the object. In a step 2, the area to be the candidate of the object is Fourier-transformed and correlation between the Fourier-transformed area and the Fourier-transformed result of the representative picture of the object is checked about an imaginary number component to judge whether the object is the specified symmetrical object or not. Thus only the specified symmetrical object can be extracted by template matching using the dictionary picture data. Since the specified symmetrical object is judged by checking its correlation by the use of the imaginary number component of Fourier-transformed data expressing the non-symmetry of a picture, the necessary of complicated filtering processing other than Fourier transform can be eliminated.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a specific symmetry suitable for determining a suspicious person or the like in the security field such as an ATM management system of a bank by analyzing image information from a surveillance camera. The present invention relates to an object determination method.

[0002]

2. Description of the Related Art A conventional symmetric object determination method based on this type of image analysis basically extracts a symmetric object by filtering a constraint condition for a symmetric object and performing image processing using this. doing.

[0003]

However, the conventional symmetric object determination method by image analysis has a problem that the filter design for extracting the symmetric object is complicated. Further, although it is possible to extract an unspecified symmetrical object, it is difficult to determine whether or not the object is a specific symmetrical object.

The present invention has been made to solve such a problem, and an object thereof is to provide a specific symmetric object determination method capable of easily determining a specific symmetric object by using a simple filtering process. To provide.

[0005]

In order to achieve the above object, the present invention is a method for extracting a specific symmetric object from an image captured by an image input means, which is representative of the specific symmetric object to be extracted. Image as dictionary image data,
The first step of calculating the correlation value between this and the image captured by the image input means, and extracting the area showing the highest correlation value as the specific symmetric object determination area, and the first step. Determination region Fourier transform data obtained by Fourier transforming the specific symmetric object determination region, and calculate the correlation value in each imaginary component of the dictionary Fourier transform data obtained by Fourier transforming the dictionary image data, A second step of determining whether or not the specific symmetric object is included in the specific symmetric object determination region based on the correlation value.

In the above-mentioned specific symmetric object determination method, in the first step, from the change point of the histogram on the coordinate axis of the difference image between the image captured by the image input means and the background image captured in advance by the image input means, Extracting the specific symmetric object determination area by limiting the area in the image is
This is suitable for accurately determining the specific symmetrical object.

Further, in the above-mentioned specific symmetric object determination method, in the first step, the specific symmetric object determination area is extracted by calculating the correlation with the dictionary image data from the area including half or more of the difference image pixels. However, it is suitable for reducing the calculation cost.

In the present invention, first, in the first stage,
A representative image of the specific symmetric object is used as dictionary image data, and by correlating this with the image from the image input means, a region that is a candidate for the specific symmetric object is extracted. Next, in the second step, the region that is a candidate for the specific symmetric object extracted above is Fourier-transformed, and the Fourier-transformed image of a typical specific symmetric object and the correlation with the imaginary component are examined. It is determined whether or not the object is a specific symmetrical object. By using the template matching using the dictionary image data in this way, only a specific symmetrical object can be extracted. In addition, by using the imaginary number component of the Fourier transform data representing the asymmetry of the image and checking the correlation to determine whether or not the object is a specific symmetric object, the need for complicated filter processing other than Fourier transform is eliminated. . The above first
By the step and the second step, the image obtained from the image input means is automatically analyzed by a computer or the like, and it is possible to determine whether or not the specific symmetrical object is included.

[0009]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described in detail below with reference to the drawings.

Here, a method of detecting a suspicious person in a bank ATM will be described as an example. That is, the specific symmetrical object is the face of the suspicious person. In addition, the suspicious person mentioned here uses tools such as sunglasses, glasses or masks,
It refers to a person who hides part of his / her face by being extremely depressed or whose face is not significantly facing the front with respect to the surveillance camera which is the image input means. Further, here, it is assumed that the surveillance camera is set near the operation panel of the ATM.

FIG. 1 is a flow chart of a suspicious person detection method showing an embodiment of the present invention. This suspicious individual detection method includes: [1] extracting a face area; [2] determining whether the extracted face area is a normal face (suspicious person determination);
Consists of Here, the normal face refers to a face that is photographed with an inclination of 30 ° or less with respect to the surveillance camera and is not wearing sunglasses, glasses, or the like.

First, the step of extracting the face area shown in [1] in FIG. 1 will be described. FIG. 2 is a diagram illustrating a method of extracting the face area.

First, background image data is previously captured by the set monitoring camera. In addition, a normal face image of a person is taken, and the face area is m for each image.
_It is divided into a number of _face × n _face blocks, and the average of the gray values of each block is represented by a representative value. Then, this is stored in the face area search dictionary database as dictionary image data.

Next, the face image of the ATM user is captured from the surveillance camera on a computer, and the image difference between this image and the background image prepared in advance is calculated. Then, a value 100 is put in an area where there is a difference in each pixel, and a value 0 is put in a portion where there is no difference, and the result is stored as binarized background difference image information on the computer.

Next, a projected image is obtained by projecting the background difference image information on each of the x-axis and the y-axis. Here, for the histogram curve of the projected image on the x-axis, the distance XL between the two inflection points appears, and for the histogram curve of the projected image on the y-axis, it first appears from the rising point of the histogram curve. Find the distance YL to the inflection point. These two
Based on one value, the block size of the matching area is XL / m _{face in} the x direction and YL / n in the y direction.
Change before and after _face . Furthermore, the search area is limited from this difference image. For example, the range between two inflection points in the histogram curve on the x-axis and the range from the rising point of the histogram curve to the first inflection point on the y-axis are defined as search areas.

Next, template matching is performed using the dictionary image data to extract the face area. First, in the search area in the image, it is determined whether or not the matching area also includes more than half of the pixels having the value 100 of the background difference image information in the matching area to determine whether or not to match the dictionary image data. To do. This is preferable in that the calculation cost is reduced by avoiding invalid matching that is unlikely to be extracted as a face area. Then, when matching, m _face ×
Block image data is created using the average of the gray values of each of the n _face blocks as a representative value, and this m _face × n _face block image data is considered as a vector expression, and the correlation with the inner product with the dictionary image data is used. Calculate and use it as a similarity measure. Then, the matching area having the highest correlation value is extracted as the face area.

Next, the step of determining whether the extracted face area shown in [2] in FIG. 1 is a normal face will be described. FIG. 3 is a diagram illustrating a method of determining whether or not the face area is a normal face (suspicious person determination).

First, a Fourier transform is applied to the face area obtained in the previous step to create face area Fourier transform data.

Next, the correlation of the imaginary number component between the dictionary Fourier transform data created by Fourier transforming a normal face image prepared in advance and the face area Fourier transform data created above is calculated.

Next, when the calculated correlation value exceeds the range of the set threshold value, it is judged as a suspicious person.

In the above embodiment, the suspicious person determination is described as an example, but the present invention is not limited to this example and can be applied to determination of a specific symmetrical object.

[0022]

As is apparent from the above description, according to the present invention, (1) an image obtained from the image input means is automatically analyzed by a computer or the like to determine whether or not there is a specific symmetrical object. You can judge.

(2) Since template matching is performed using dictionary image data, it is possible to perform extraction limited to a specific symmetrical object.

(3) Since only Fourier-transformed data is used when determining whether the object is a specific symmetrical object, it is not necessary to use a complicated filter.

Further, when the symmetrical object determination area is limited and extracted from the difference image between the captured image and the background image, erroneous determination due to inclusion of the background image is reduced, The specific symmetrical object can be determined more accurately.

Further, when the correlation with the dictionary image data is calculated from the area including more than half of the difference image pixels to extract the specific symmetric object determination area, the collation with the waste dictionary image data is particularly performed. It can be eliminated to reduce the calculation cost.

[Brief description of the drawings]

FIG. 1 is a diagram showing an embodiment of a specific symmetric object determination method according to the present invention, and is a flowchart showing an example of a suspicious person detection method when applied to a suspicious person detection system.

FIG. 2 is a diagram illustrating an example of a face area extraction method of the suspicious person detection method according to the above-described embodiment.

FIG. 3 is a diagram illustrating an example of a suspicious person determination method of the suspicious person detection methods in the above-described embodiment.

Claims (3)

[Claims] 1. A method for extracting a specific symmetric object from an image input by an image input means, wherein a representative image of the specific symmetric object to be extracted is used as dictionary image data, and this is input by the image input means. The correlation value with the extracted image and extracting the region showing the highest correlation value as the specific symmetric object determination region, and the specific symmetric object determination region extracted in the first stage Calculating the correlation value in each imaginary component of the determination region Fourier transform data obtained by conversion and the dictionary Fourier transform data obtained by performing Fourier transform of the dictionary image data, and calculating the correlation value based on the correlation value. A specific symmetric object determination method, comprising: a second step of determining whether or not a specific symmetric object determination area includes a specific symmetric object. 2. In the first step, the area in the image is limited from the change point of the histogram on the coordinate axis of the difference image between the image captured by the image input means and the background image captured in advance by the image input means. The specific symmetric object determination method according to claim 1, wherein the specific symmetric object determination region is extracted. 3. The specific symmetric object determination area is extracted in the first stage by calculating a correlation with dictionary image data from an area including half or more of difference image pixels. Item 2. The specific symmetric object determination method according to Item 2.