KR101165415B1 - Method for recognizing human face and recognizing apparatus - Google Patents

Abstract

The present invention relates to a biometric face recognition method and a recognition apparatus in an image, and an object of the present invention is to provide a biometric face recognition method and a recognition device in an image that accurately recognizes only the coordinates of the biometric face in an image such as a panoramic picture.
According to the present invention, a first step of acquiring visible light image data by capturing a specific point by using a visible light photographing apparatus, and the same point and the same angle of view at the same point as the photographing point of the first step by using an infrared photographing apparatus A second step of acquiring infrared image data by photographing using a face pattern algorithm, pattern recognition of a biological face included in the visible light image data, and extracting coordinate information of a point where the corresponding biological face is located A third step of storing in a first table, a fourth step of extracting coordinate information of a biological face included in the infrared image data using a temperature color value, and storing the coordinate information in a second table; Left stored in the first table using the stored coordinate information and the coordinate information stored in the second table. An in-image biometric recognition method is provided, comprising a fifth step of identifying coordinates in the table that are erroneously recognized as a non-living face.

Description

Biometric Face Recognition Method and Recognition Device in Image {METHOD FOR RECOGNIZING HUMAN FACE AND RECOGNIZING APPARATUS}

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a facial recognition method and a recognition device, and more particularly, to a biometric face recognition method and a recognition device in an image for accurately and automatically identifying only a human face in an image obtained through a photographing device.

With the development of wired and wireless communication, the development and expansion of electronic information terminals such as computers, navigators and mobile phones are increasing. Accordingly, users who want to obtain geographic information of a specific region can easily access the terminal and receive an electronic map service containing all information related to the region, such as roads, terrain, and features. However, since the conventional geographic information providing service is implemented and displayed in 2D, it was impossible for the user to visually accurately recognize the terrain and features related to the corresponding area.

In order to overcome the above problems, recently, a panoramic photo geographic information providing service for displaying a real picture of a corresponding area on an electronic map has been developed and attempted to visually recognize roads, terrain, and features.

In order to implement the geographic information providing service using the above-mentioned panorama picture, basically, a picture taking work, a picture connecting work, and a mapping work are required.

Photographing refers to a task of acquiring image frames through segmented photographing of terrain, features, and roads by moving each region using a plurality of cameras installed in a vehicle.

Photo linking refers to a process of editing a plurality of image frames acquired through a photo shooting process while connecting them as panorama photos in the order of being located along the roadside. Each panorama picture configured through the picture connection operation can be mapped to a digital map that can be connected with a GPS code, thereby providing location information of an actual image.

The photo mapping task refers to a task of placing a corresponding photo on an accurate position on an electronic map by using GPS (Global Positioning System) coordinates of each image frame obtained through the photo shooting task.

Commercial sites that provide such panoramic photo geographic information service include 'Road View' provided by Daum (www.daum.net) site, and 'Street View' provided by 'Google Site'. .

However, each panorama picture collected by the vehicle includes personal information such as the vehicle number and the face of the person. In order to prevent human face from being exposed through the panorama picture, the face pattern recognition algorithm is used. When the photographed image includes an image corresponding to the eyes, nose, and mouth, the pattern is examined to recognize that the face and blur the corresponding portion.

However, the face pattern recognition algorithm does not recognize the face when the side of the face is exposed or misrecognizes the face of the movie post, billboard, etc., which is similar to the face shape, as a biometric face. After application, considerable time and manpower is spent to correct this misunderstanding included in the image through manual operation, and it is difficult to identify non-living facial images without recognizing the face of the human body if it depends on the manual operation of the operator. There was this.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and an object of the present invention is to provide a biometric face recognition method and a recognition device in an image that accurately recognizes only the coordinates of the biometric face in an image such as a panoramic photograph.

The object of the present invention is to obtain visible light image data by capturing a specific point by using a visible light photographing apparatus, and to use the infrared photographing apparatus to view the same point as the photographing point of the first step with the same direction and same angle of view. A second step of acquiring infrared image data by photographing using a face pattern algorithm, pattern recognition of a biological face included in the visible light image data, and extracting coordinate information of a point where the corresponding biological face is located A third step of storing in a first table, a fourth step of extracting coordinate information of a biological face included in the infrared image data using a temperature color value, and storing the coordinate information in a second table; Coordinates stored in the first table by using the stored coordinate information and the coordinate information stored in the second table And a fifth step of identifying coordinates in which the non-living face is incorrectly recognized as the living face.

According to the biological face recognition apparatus and the recognition method in the image according to the present invention, at the same time as photographing a specific point through the visible light imaging device, and configured to shoot the same point in the same direction and the same angle of view through the infrared imaging device, By using two pieces of data acquired at the same point, that is, visible light image data and infrared image data, only a human face can be automatically identified without errors in the captured image, thereby reducing the time and manpower required for photo correction. There is a remarkable effect of drastically reducing and avoiding omissions that occur when relying on the manual work of the operator.

1 is a block diagram of an apparatus for recognizing a biometric face in an image according to an exemplary embodiment of the present invention.
2 is a flowchart illustrating a flow of a method for recognizing a biometric face in an image according to a first exemplary embodiment of the present invention.
3 is a flowchart illustrating a flow of a method for recognizing a biometric face in an image according to a second exemplary embodiment of the present invention.
4 is a flowchart illustrating a flow of a method for recognizing a biometric face in an image according to a third exemplary embodiment of the present invention.

The biometric face recognition method and the recognition device in the image according to the present invention are configured to capture a specific point through a visible light photographing apparatus and to simultaneously photograph the same spot with the same direction and the same angle of view through an infrared photographing apparatus. By using two pieces of data obtained for a point, that is, visible light image data and infrared image data, a technical feature of automatically identifying and correcting only a human face in a captured image is provided.

Hereinafter, with reference to the accompanying drawings, the spirit and preferred embodiments of the present invention will be described in more detail.

1 is a block diagram of an apparatus for recognizing a biometric face in an image according to an exemplary embodiment. Referring to FIG. 1, the in-image biometric face recognition apparatus according to the present invention includes a visible light photographing apparatus 11 for obtaining visible light image data by photographing a specific point, and an infrared light for obtaining infrared image data by photographing a specific point. The visible light image using the photographing apparatus 13, the first memory 10 storing the visible light / infrared image data collected by the visible light / infrared photographing apparatuses 11 and 13, and a face pattern algorithm. Pattern recognition of a biological face included in the data is performed to extract coordinate information of a point at which the corresponding biological face is located, extract coordinate information of the biological face included in the infrared image data using a temperature color value, and the extracted The non-living face among the coordinates stored in the visible light image data using two kinds of coordinate information is incorrectly used as the living face. It is configured to include at least one processing unit 23 to perform an operation to identify the sikhan coordinates.

In particular, the biometric face recognition apparatus according to the present invention compares the coordinate information extracted from the two types of image data (that is, visible light / infrared image data) by the arithmetic processing unit 23 to correct face coordinates. It can be configured to have the following features.

That is, when the visible light photographing apparatus 11 of the present invention performs a photographing operation on a specific point (for example, the first point), the infrared photographing apparatus 13 simultaneously moves the same first point to the same direction and the same angle of view. It is configured to perform a photographing operation so that the operator acquires the same picture represented in two kinds (ie, visible light image and infrared image) for the first point.

The biometric face recognition apparatus 20 of the present invention is generally provided in the form of a computer including a personal computer, a notebook computer, etc. together with the visible light / infrared photographing apparatuses 11 and 13, and the arithmetic processing apparatus 23 described above. ), An input / output device (21, 25) including a mouse, a keyboard, and a monitor, and a second memory (27) installed inside or outside the computer.

An example of the visible light image data stored in the first memory 10 may include 'road view' provided by 'next (www.daum.net) site' or 'street view' provided by 'google site'. Panoramic image data. The infrared image data refers to image data photographed with infrared rays at the same angle of view and in the same direction at the same point as the corresponding visible light image data.

The visible light image data and the infrared image data use a method of automatically photographing four cameras whenever a certain distance is moved using four cameras mounted on a vehicle. In the case of using such an automatic vehicle imaging system, four infrared ray photographing apparatuses having four visible ray photographing apparatuses so as to maintain a right angle to each of the concentric axes, and four infrared ray photographing apparatuses maintaining the same angle as the visible ray photographing apparatus 11 on the same concentric axis ( 13) when the image is taken with the infrared image data can be obtained by the infrared image data taken with the same angle of view in the same direction at the same point and the corresponding visible light image data. More precisely, in this way, since the visible light imaging device 11 is located at a lower position than the infrared imaging device 13, it is not possible to photograph exactly the same point with each other, but this error can be overcome by software processing. Therefore, it can be seen that the same angle of view was taken in the same direction at the same point.

Using the biological face recognition apparatus of the present invention including the above-described visible light photographing apparatus 11, infrared photographing apparatus 13, first memory 10, and arithmetic processing apparatus 23, A method for identifying only a human face in an image without errors may include a first step of acquiring visible image data by photographing a specific point using a visible light photographing apparatus, and using the infrared photographing apparatus. A second step of acquiring infrared image data by photographing the same point as the photographing point with the same direction and the same angle of view; pattern recognition of the biological face included in the visible light image data using a face pattern algorithm, and the corresponding biological face being positioned Extracting coordinate information of a point to be stored in a first table, and storing the infrared information by using a temperature color value; Extracting coordinate information of the biological face included in the image data and storing the coordinate information in the second table, and using the coordinate information stored in the first table and the coordinate information stored in the second table. And a fifth step of identifying coordinates erroneously recognized as non-living faces among the coordinates stored in the table.

Hereinafter, in the fifth step, a method of identifying coordinates in which a non-living face is incorrectly recognized as a living face among coordinates stored in the first table using coordinate information stored in the first table and coordinate information stored in the second table. It will be described for the preferred embodiment of.

2 is a flowchart illustrating a flow of a method for recognizing a biometric face in an image according to a first exemplary embodiment of the present invention.

First, the visible light image data is acquired by photographing a target point (for example, a first point) using a visible light photographing apparatus, and at the same time, the same point and the same angle of view as the first point using the infrared photographing apparatus are obtained. Infrared image data is acquired and stored in the first memory.

The coordinates corresponding to the face are extracted from the visible light image data stored in the first memory using the face pattern algorithm and stored in the first table (ST 201).

However, the face pattern recognition algorithm performed on the visible light image also recognizes the face displayed on the poster or the advertisement leaflet of the theater, and the first table of ST 201 includes incorrect face coordinates (ie, non-living face coordinates). It may be. Therefore, it is necessary to identify and remove the non-living face coordinates among the face coordinates stored in the first table using the face pattern algorithm as described above, which can be achieved by using the coordinate information extracted from the infrared image data of the present invention. Do. Details will be described later.

For reference, the present invention is not an invention in which the face pattern algorithm is a main technical idea, and various methods are known in the face pattern algorithm, and thus detailed description thereof will be omitted.

In addition, since several faces may exist on one visible light image, a plurality of coordinates may be stored. As an example, suppose that five faces are included in the first visible light image, and extracted them, the first table includes {(x1, y1), (x2, y2), (x3, y3), (x4, y4), (x5, y5)}. Where (x1, y1), (x2, y2), ..., (x5, y5) represent the positions of the first, second, ..., fifth faces present on the first visible light image. This is expressed by the x and y coordinate values on the visible light image.

Next, the temperature color value of the coordinates on the infrared image data corresponding to each coordinate point (that is, the coordinates on the visible light image) extracted by the ST201 is extracted (ST 203), and at least two different temperature color values from the extracted temperature color values. The temperature color average value is calculated using the remaining temperature color values except for the temperature color value having the difference between the temperature color value and the predetermined threshold value.

That is, the temperature color value on the infrared image corresponding to each coordinate value of the first table is obtained from the first infrared image matched with the first visible light image, and at this time, another temperature color value among the temperature color values obtained according to each coordinate value. The temperature color average value is calculated using the temperature color value corresponding to the remaining coordinate values without including the difference between the first threshold value and the difference. For example, on the first infrared image corresponding to {(x1, y1), (x2, y2), (x3, y3), (x4, y4), (x5, y5)} on the first visible light image, respectively. When the temperature color value is obtained as {100, 105, 140, 99, 95}, if the first threshold value is '30', the remaining temperature color value except for the temperature color value '140' corresponding to (x3, y3) Find the average. The average of the temperature colors obtained in the above example is calculated as '99 .75 '.

However, unlike the above example, when only one or two biological faces are included in one visible light image, the average of these temperature color values may become meaningless. Even in this case, the temperature color average value can be determined by using the biological face included in the continuously visible image. For example, if there is one biometric face recognized in the first visible light image, two biometric faces recognized in the second visible light image, and one biometric face recognized in the third visible light image, the first visible face is present. The biometric face coordinate values included in the three visible light rays of the light ray, the second visible light, and the third visible light may be obtained, and the average temperature color value may be obtained using the corresponding temperature color values. Since the average value of temperature color used in the present invention has a meaning of a numerical value representing an average temperature corresponding to a living body face, the temperature color average value is obtained from a single visible light image or a coordinate of a biological face included in a plurality of visible light images. Both sides using the corresponding color value are available.

Next, a predetermined temperature color value range is set based on the aforementioned temperature color average value, and coordinates of an area included in the temperature color value range are extracted from the infrared image data and stored in the second table (ST 205).

That is, the coordinates having color values near the temperature color average value included in the first infrared image are obtained and stored in the second table. The color values near the temperature color average value are based on the temperature color average value calculated in ST 203. As a result, a pattern having an area included in a color value range within a predetermined range is found and expressed as a coordinate value. For example, if the average temperature color value is '99 .75 'as shown in the above example, a face pattern within ± 5 range is retrieved from the first infrared image and stored as a coordinate value in the second table.

Next, the coordinates stored in the first table and the coordinates stored in the second table are compared with each other (ST 207). In other words, the number and coordinates of the coordinates stored in each table are compared, and the face pattern recognition algorithm performed on the visible light image and the infrared image are performed on the position comparison between the coordinate values of the first table and the coordinate values of the second table. Since the face pattern recognition algorithm performed cannot be identical, the face pattern recognition algorithm is separated by a predetermined threshold distance, and the corresponding coordinates are recognized as representing the same position.

As a result of the comparison in step ST 207, if coordinates stored only in the first table and not stored in the second table are detected, the detected coordinates are corrected to be deleted on the first table (ST 209). That is, when the position and the number of coordinate values of the first table and the coordinate values of the second table do not match, the coordinates on the visible light image are corrected based on the coordinate values of the second table to adjust the position of the living body face on the visible light image. To confirm.

In addition, if the comparison result of the step ST 207, the coordinates stored only in the second table and not stored in the first table is detected, the method further includes the step of additionally inputting the detected coordinates on the first table. This is because the face pattern recognition algorithm performed on the visible light image may be the result of not grasping the side face. For reference, deleting a coordinate stored only in the first table and not in the second table on the first table may include a facial pattern recognition algorithm performed on a visible light image included in a poster of a theater or a promotional leaflet of an advertisement leaflet. This is because the face may be determined as an error recognized as a living face.

Next, the biometric face position determined in step 209 is blurred, or when the position and the number of coordinates of the first table and the coordinates of the second table coincide with each other as a result of the determination of step ST 207, the corresponding biometric face position If the blur is processed immediately, the operation is completed (ST 211).

3 is a flowchart illustrating a flow of a method for recognizing a biometric face in an image according to a second exemplary embodiment of the present invention.

First, the visible light image data is acquired by photographing a target point (for example, a first point) using a visible light photographing apparatus, and at the same time, the same point and the same angle of view as the first point using the infrared photographing apparatus are obtained. Infrared image data is acquired and stored in the first memory.

Then, the coordinates corresponding to the face are extracted from the visible light image data stored in the first memory using the face pattern algorithm and stored in the first table (ST 301).

For example, suppose that five living body faces are included in the first visible light image, and extracted and extracted {(x1, y1), (x2, y2), (x3, y3), (x4, y4) into the first table. ), (x5, y5)}. In a similar manner, it is determined whether a pattern within a temperature color value range corresponding to the living body face is included from the infrared image data stored in the first memory, and the corresponding coordinates are extracted and stored in the second table (ST303).

An infrared image is an image representing a temperature emitted from a photographed object in color, for example, a high temperature is displayed in red and a low temperature is displayed in blue. When the living body is photographed with an infrared ray photographing device, the color is expressed according to the temperature of the living body. By the way, the living body temperature is slightly different depending on the weather and depending on whether or not exercise, there is a characteristic that has a temperature range within a certain range due to the characteristics of the constant temperature animal. These characteristics can be used to accumulate the temperature range of the biological face expressed by season and weather in several infrared images, and then numerically express the biological face temperature range and its corresponding color range by arithmetic mean or other statistical methods. Will be.

Next, a temperature color value corresponding to each coordinate extracted in ST303 is extracted, and an average value of the extracted temperature color values is calculated (ST305). As described in the flowchart of FIG. 2, when the number of biological faces included in one infrared image data is small, the average temperature color value may be calculated using the number of biological faces included in several consecutive sheets.

Next, the temperature color value of the coordinates on the infrared image data corresponding to each coordinate point extracted in ST301 is extracted to obtain a difference value from the average value of the temperature color calculated in ST305, and whether the difference value is larger than a predetermined threshold value. It is determined (ST306).

That is, the temperature color value of the coordinate value on the infrared image data corresponding to each coordinate value stored in the first table is extracted, the difference value between the extracted temperature color value and the temperature color average value is obtained, and the difference value is the third threshold value. To see if it is greater.

As a result of the determination of ST306, the temperature color value having a difference larger than the third threshold value is very likely to be not a living body, and thus the coordinate value on the visible light image data (ie, the first table) corresponding to the temperature color value is determined by the biological. If not, the corresponding coordinate is deleted from the first table (ST309). This is to remove an error caused by the case in which the biometric face printed on the promotional leaflet is included as described above.

Next, the temperature color value on the infrared image data corresponding to each coordinate point stored in the first table having passed through ST309 is extracted, and the average value of the extracted temperature color value is calculated (ST311).

The difference value between the temperature color value of each coordinate stored in the second table of the ST303 and the average value of the temperature color calculated in the ST311 is obtained, and it is determined whether the difference is greater than the fourth threshold (ST313).

As a result of the determination of ST313, when a temperature color value having a difference greater than the fourth threshold value is detected, the coordinate on the second table corresponding to the detected temperature color value is determined to be a cat or other living organism, although it is a living body. (ST 315).

It is determined that each of the coordinates stored in the first table and the second table, which have all the above-described processes, represent the living face, and the coordinate region included in either of the tables is determined as the living face (ST317), and the blurring is performed. Process (ST319).

For more accurate determination, in step ST317, the coordinates included in both the first table and the second table may be determined as a living face, and the blurring may be performed.

4 is a flowchart illustrating a flow of a method for recognizing a biometric face in an image according to a third exemplary embodiment of the present invention.

First, the visible light image data is acquired by photographing a target point (for example, a first point) using a visible light photographing apparatus, and at the same time, the same point and the same angle of view as the first point using the infrared photographing apparatus are obtained. Infrared image data is acquired and stored in the first memory.

The coordinates corresponding to the face are extracted from the visible light image data stored in the first memory using the face pattern algorithm and stored in the first table (ST401). For example, suppose that five living body faces are included in the first visible light image, and extracted and extracted {(x1, y1), (x2, y2), (x3, y3), (x4, y4) into the first table. ), (x5, y5)}.

Next, a temperature color value range corresponding to the living face is set, a pattern included in the set temperature color value range is extracted from the infrared image data stored in the first memory, and corresponding coordinates are calculated and stored in the second table ( ST403). For example, suppose that six biological faces are included in the first infrared image corresponding to the first visible light image, and extracted from the {(a1, b1), (a2, b2), (a4, b4), (a5, b5), and (a6, b6)}.

Next, the coordinates on the visible light image stored in the first table and the coordinates on the infrared image stored in the second table are compared with each other to determine whether they match (ST405). In the above example, the biological face extracted from the first visible light image is {(x1, y1), (x2, y2), (x3, y3), (x4, y4), (x5, y5)} in the first table. The biometric face coordinates, which are detected with five coordinates of and extracted from the first infrared image, are represented by {(a1, b1), (a2, b2), (a4, b4), (a5, b5), (a6) in the second table. , b6)}, of which (x1, y1), (x2, y2), (x4, y4), and (x5, y5) are (a1, b1), (a2, b2), Assume that the same coordinates corresponding to (a4, b4) and (a5, b5) are displayed, (x3, y3) is a coordinate detected only in the first table, and (a6, b6) is detected only in the second table. Assume that the coordinates are

As a result of the determination in step ST405, if mutually inconsistent coordinates are found, the coordinates are displayed on the visible light image and provided to the image processor through the display device (ST407). At this time, the coordinates (x3, y3) detected only in the first table and the coordinates (a6, b6) detected only in the second table are displayed to the image processor using different expression formats (for example, markers having different colors or shapes). Do it.

The image processor checks the coordinates corresponding to the living body by looking at the provided mismatched coordinates, inputs them through the input device, and updates the coordinate values of the first table according to the received coordinates (ST 409). Thereafter, the visible light image area is blurred according to the verified coordinates of the first table (ST 411).

Although specific embodiments of the present invention have been described and illustrated above, it will be apparent that various modifications may be made by those skilled in the art without departing from the technical spirit of the present invention. Such modified embodiments should not be understood individually from the spirit and scope of the present invention, but should fall within the claims appended to the present invention.

10: first memory 11: visible light imaging device
13: infrared imaging device 20: biometric face recognition device
21: output device 23: processing unit
25: input device 27: second memory

Claims (8)

An in-image biometric face recognition method for automatically identifying only a face of a person in an image obtained by a photographing apparatus for a blurring process of a living face in a panoramic photo geographic information providing service,
A first step of obtaining visible light image data by photographing a specific point using a visible light photographing apparatus;
A second step of acquiring infrared image data by capturing the same spot as the photographing point of the first step with the same direction and the same angle of view by using the infrared photographing apparatus;
A third step of recognizing a biometric face included in the visible light image data, extracting coordinate information of a point at which the biometric face is located, and storing the coordinate information in a first table;
A fourth step of extracting coordinate information of a living face included in the infrared image data using a temperature color value and storing the coordinate information in a second table;
A fifth step of identifying coordinates erroneously recognized as non-living faces among the coordinates stored in the first table by using the coordinate information stored in the first table and the coordinate information stored in the second table; And
And a sixth step of blurring the visible light image area corresponding to the coordinates stored in the first table in which the coordinates identified in the fifth step are excluded.
An in-image biometric face recognition method for automatically identifying only a face of a person in an image obtained by a photographing apparatus for a blurring process of a living face in a panoramic photo geographic information providing service,
A first step of obtaining visible light image data by photographing a specific point using a visible light photographing apparatus;
A second step of acquiring infrared image data by capturing the same spot as the photographing point of the first step with the same direction and the same angle of view by using the infrared photographing apparatus;
A third step of extracting coordinates corresponding to a face from the visible light image data obtained in the first step and storing the coordinates corresponding to the face in a first table;
Extracting a temperature color value of the coordinates on the infrared image data corresponding to each coordinate point extracted in the third step, and a temperature color having at least two different temperature color values from the extracted temperature color values and a difference greater than a predetermined threshold A fourth step of calculating a temperature color average value using the remaining temperature color values except for the values;
A predetermined temperature color value range is set based on the temperature color average value of the fourth step, and the coordinates of an area included in the temperature color value range are extracted from the infrared image data obtained in the second step, and the second color table is extracted. A fifth step of storing;
A sixth step of comparing the coordinates stored in the first table with the coordinates stored in the second table;
A seventh step of deleting the detected coordinates on the first table when the coordinates stored in the first table and not stored in the second table are detected as a result of the comparison in the sixth step; And
And an eighth step of blurring the visible light image area corresponding to the coordinates remaining on the first table.
The method of claim 2,
A ninth step of adding the detected coordinates on the first table when the sixth step comparison result detects coordinates stored only in the second table and not stored in the first table; And
And a tenth step of recognizing a coordinate area stored in the first table, which has passed through the ninth step, as a living body face.
The method according to claim 2 or 3,
If the coordinates of the first table and the coordinates of the second table to be compared with each other in the sixth step is a corresponding coordinate within a predetermined threshold range, it is processed as the coordinates representing the same position.
An in-image biometric face recognition method for automatically identifying only a face of a person in an image obtained by a photographing apparatus for a blurring process of a living face in a panoramic photo geographic information providing service,
A first step of obtaining visible light image data by photographing a specific point using a visible light photographing apparatus;
A second step of acquiring infrared image data by capturing the same spot as the photographing point of the first step with the same direction and the same angle of view by using the infrared photographing apparatus;
A third step of extracting coordinates corresponding to a face from the visible light image data obtained in the first step and storing the coordinates corresponding to the face in a first table;
A fourth step of setting a temperature color value range corresponding to a living body face, extracting a pattern included in the temperature color value range from the infrared image data obtained in the second step, calculating a corresponding coordinate, and storing the coordinate in a second table; ;
A fifth step of calculating a mean value of the extracted temperature color values by extracting temperature color values corresponding to the respective coordinates extracted in the fourth step;
Extracting a temperature color value of the coordinates on the infrared image data corresponding to each coordinate point extracted in the third step to obtain a difference value from the average value of the temperature color in the fifth step, and whether the difference is greater than a predetermined threshold Determining a sixth step;
A seventh step of deleting, on the first table, coordinates corresponding to the detected temperature color value when a temperature color value having a difference greater than the threshold is detected as a result of the sixth step determination; And
And an eighth step of blurring the visible light image area corresponding to the coordinates remaining on the first table.
6. The method of claim 5,
A ninth step of extracting a temperature color value on the infrared image data corresponding to each coordinate point stored in the first table through the seventh step, and calculating an average value of the extracted temperature color values;
A tenth step of obtaining a difference value between a temperature color value of each coordinate stored in the second table of the fourth step and a temperature color average value calculated in the ninth step, and determining whether the difference is greater than a predetermined threshold; ;
An eleventh step of deleting coordinates corresponding to the detected temperature color value on the second table when a temperature color value having a difference greater than the threshold is detected as a result of the determination of the tenth step; And
And a twelfth step of recognizing a coordinate area included in any one or more of the first table through the seventh step and the second table through the eleventh step as a human face. Way.
An in-image biometric face recognition method for automatically identifying only a face of a person in an image obtained by a photographing apparatus for a blurring process of a living face in a panoramic photo geographic information providing service,
A first step of obtaining visible light image data by photographing a specific point using a visible light photographing apparatus;
A second step of acquiring infrared image data by capturing the same spot as the photographing point of the first step with the same direction and the same angle of view by using the infrared photographing apparatus;
A third step of extracting coordinates corresponding to a face from the visible light image data obtained in the first step and storing the coordinates corresponding to the face in a first table;
A fourth step of setting a temperature color value range corresponding to a living body face, extracting a pattern included in the temperature color value range from the infrared image data obtained in the second step, calculating a corresponding coordinate, and storing the coordinate in a second table; ;
A fifth step of comparing the coordinates stored in the first table with the coordinates stored in the second table to determine whether they match;
A sixth step of providing the image processor by displaying the inconsistent coordinates on the visible light image when the mismatched coordinates are detected as a result of the determination of the fifth step;
A seventh step of receiving coordinate information determined to correspond to a face by the image processor among the mismatched coordinates;
An eighth step of updating the coordinates stored in the first table based on the coordinate information received in the seventh step; And
And a ninth step of blurring the visible light image area corresponding to the coordinates stored in the first table.
The method of claim 7, wherein
The method of claim 6, wherein the mismatch coordinates detected only in the first table and the mismatch coordinates detected only in the second table are provided to the image processor through different representation formats.

Images