KR20130099642A - Apparatus and method for recongnizing face using adaptive illumination - Google Patents

Abstract

PURPOSE: A face recognition device using adaptive lighting control and a method thereof are provided to increase the face recognition rate by using adaptive lighting control based on an analysis result which is provided by analyzing images for face recognition. CONSTITUTION: An image acquisition unit (10) acquires images by taking pictures of a photographed person, and a face region detecting unit (120) which detects a face region from the images. An image analysis unit (122) analyzes a histogram of a face region, and discriminates the validity of the images. A lighting control unit (124) controls the brightness of lights based on the result of the discrimination, and the image acquisition unit to acquire images again by using the lights whose brightness is controlled. The image analysis unit includes a histogram generation unit, which generates a histogram showing the pixel value and the brightness of the light and shade of the face region, and a first histogram analysis unit. [Reference numerals] (10) Image acquisition unit; (120) Face region detecting unit; (122) Image analysis unit; (124) Lighting control unit; (12a) Control unit; (14) User input unit; (16) Image output unit

Description

Apparatus and method for recongnizing face using adaptive illumination

One aspect of the present invention relates to a security authentication technology, and more particularly to a face recognition technology.

In biometrics, access control, time and attendance management, building integrated systems, financial automation equipment, computer security, electronic commerce authentication, airport information systems, and other biometric technologies have been proposed. Among them, face recognition is a technology for recognizing a person using face information having unique characteristics for each person.

A general face recognition apparatus receives an image of a subject and outputs it on a screen, and the user is configured to authenticate the subject while checking the face of the output image. However, if the illuminator irradiated for the image acquisition is too close to the subject, strong light may be emitted and saturation may occur, and if too far, the light may be weak enough to express facial features so that the face is properly exposed. I can't recognize it.

According to an aspect, a face recognition apparatus using adaptive lighting control and a method thereof are provided to obtain an optimal image capable of face recognition.

According to an aspect, an apparatus for recognizing a face may include: an image acquirer configured to capture an image of a subject under illumination, and to acquire an image, a face region detector to detect a face region from the acquired image, and a histogram of the detected face region The image analyzing unit for determining the validity of the image and the image analysis unit to adjust the brightness of the illumination according to the image acquisition unit includes a lighting control unit for controlling to re-acquire the image using the light with the brightness adjusted.

According to an embodiment, the face recognition rate may be increased by analyzing an image acquired for face recognition and adjusting the lighting adaptively to the environment based on the analysis result. In other words, the surrounding environment for the image acquisition, for example, the illuminator irradiated for the image acquisition is too close to the subject, so that strong light is emitted and saturation occurs or the distance is too weak to express the facial features. It is possible to obtain an optimal image for face recognition regardless of the case where illumination is illuminated.

1 is a reference diagram for explaining an environment requiring adaptive lighting control of the present invention,
2 is a block diagram of a face recognition apparatus according to an embodiment of the present invention;
3 is a detailed configuration diagram of the image analyzer of FIG. 2;
4 is a block diagram of a face recognition apparatus according to a further embodiment of the present invention;
5A to 5C are reference views illustrating examples of histograms for image analysis of the present invention;
6 is a reference diagram illustrating an example of an input image requiring adaptive lighting adjustment according to an embodiment of the present invention;
7 is a reference diagram illustrating a histogram for explaining an histogram analysis example of an image analyzer;
8 is a reference diagram for explaining an adaptive lighting control process when feature information is not extracted according to an embodiment of the present invention;
9 is a flowchart illustrating a face recognition method using adaptive lighting control according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the following description of the present invention, a detailed description of known functions and configurations incorporated herein will be omitted when it may make the subject matter of the present invention rather unclear. In addition, the terms described below are defined in consideration of the functions of the present invention, and this may vary depending on the intention of the user, the operator, or the like. Therefore, the definition should be based on the contents throughout this specification.

1 is a reference diagram for explaining an environment requiring adaptive lighting control of the present invention.

The present invention is a technique for acquiring an image by photographing a subject with illumination, and recognizing the subject of the subject from the acquired image. However, due to the nature of the lighting, the amount of light emitted varies depending on the distance between the subject and the lighting. Accordingly, saturation may occur when the illuminator is too close to the subject as shown in FIG. 1 (a), and when the illuminator is too far as in FIG. Can be illuminated. In the former case, a shorter person is often used. In the latter case, a shorter person is often used. Accordingly, the present invention proposes to prevent the phenomenon that the above-mentioned saturation occurs or the light is weak enough that facial features cannot be expressed by analyzing the acquired image and adaptively adjusting the illumination according to the analysis result. It became.

2 is a block diagram of a face recognition apparatus 1a according to an embodiment of the present invention.

The face recognition apparatus 1a includes an image acquisition unit 10, a control unit 12a, a user input unit 14, and an image output unit 16.

The face recognition device 1a is a device for identifying a face between people, and has various application ranges such as access control, time and attendance management, building integration system, financial automation device, computer security field, electronic commerce authentication, and airport information system. The configuration of the face recognizing apparatus 1a shown in FIG. 2 schematically shows only the components necessary for explaining the embodiment of the present invention. Therefore, components for performing other functions essential for the operation of the face recognizing apparatus 1a may be additionally included in the face recognizing apparatus 1a. In this case, the additional components may vary according to the type or function of the face recognizing apparatus 1a.

The face recognition apparatus 1a according to the present invention adaptively adjusts the brightness of the illumination through image analysis. Accordingly, even in various environments for face recognition, for example, when the illuminator irradiated for image acquisition is too close to the subject, saturation occurs or the light is so weak that the facial features cannot be expressed. Adaptive lighting brightness control can improve face recognition rate by acquiring images optimized for the environment. In particular, in the face recognition using active illumination, the present invention can be applied to adjusting the illumination adaptively to the environment. Active lighting is designed to produce ambient lighting instead of spot lighting or directional lighting. Ambient light is not directional, and under that light all objects are illuminated with uniform brightness, regardless of location. The configuration of the face recognizing apparatus 1a will be described later in detail.

The image acquisition unit 10 acquires an image of a face of a subject, for example, it may be a camera. The image acquisition unit 10 may acquire an infrared image by photographing a subject in a state where the infrared light is irradiated. An infrared illuminator may be included to irradiate infrared light, and the infrared illuminator may be located outside or inside the image acquisition unit 10. In addition, the infrared illuminator can adjust the wavelength of the infrared light, for example, a near-infrared light of about 700 to 1400nm wavelength, or a long wavelength infrared light of about 1400nm wavelength can be selected as necessary. When the image acquisition unit 10 using the infrared illuminator is difficult to recognize the face from the visible light image when acquiring the face image using the visible light illuminator, the image acquisition unit 10 uses the infrared image to extract the face region from the visible light image. It may be used as an auxiliary means.

The controller 12a performs overall management and control functions necessary for the operation of the face recognition apparatus 1a. That is, the controller 12a controls the image acquisition unit 10 to acquire an image, processes the user input signal received through the user input unit 14, and outputs the processing result through the control unit 12a to the image output unit ( Output via 16).

According to an exemplary embodiment, the controller 12a may include a face region detector 120, an image analyzer 122, and an illumination controller 124.

The face area detector 120 detects a face area from an image acquired by the image acquirer 10. Facial region detection can use the disclosed general techniques. For example, the face region detector 120 may detect an approximate face from the input image, extract eye, nose, mouth, etc., which are certain components in the face, and extract the face region based on the extracted face region. In this case, if the positions of the two eyes are detected, the distances of the two eyes can be obtained. The face region detector 120 may extract the face region from the input image based on the distance between the two eyes.

The image analyzer 122 analyzes the histogram of the face region detected by the face region detector 120 to determine the validity of the image acquired by the image acquirer 10. The detailed configuration of the image analyzer 122 will be described later with reference to FIG. 3.

The lighting controller 124 adjusts the brightness of the light according to the result of the validity determination of the image analyzer 122, and controls the image acquisition unit 10 to reacquire the image using the light whose brightness is adjusted. For example, the lighting may be brightened by increasing the intensity of light irradiated through the illuminator of the image acquisition unit 10, or the lighting may be darkened by lowering the intensity of light. Alternatively, the number of illuminators for irradiating light may be increased to brighten the light, or the number of illuminators may be reduced to dim the light. Alternatively, the brightness of the light may be adjusted by changing the wavelength range of the illuminator.

The user input unit 14 receives an input signal from the user and transfers it to the controller 12a to process an event corresponding to the input signal. The image output unit 16 outputs an image acquired through the image acquisition unit 10. In this case, the image output unit 16 may output the image reacquired through the image acquisition unit 10 in a state in which the brightness of the illumination is reset through the lighting control unit 124. The user may recognize the face of the subject while viewing the image output from the image output unit 16.

3 is a detailed block diagram of the image analyzer 122 of FIG. 2.

Referring to FIG. 3, the image analyzer 122 includes a histogram generator 1220, a first histogram analyzer 1222, and a second histogram analyzer 1224. In the following, the configuration of FIG. 3 will be described with reference drawings shown in FIGS. 5A to 5C for better understanding of the description.

The histogram generator 1220 generates a histogram indicating a number of picture elements for each contrast of the face region detected by the face region detector 120. For example, as shown in FIG. 5A, one face area, such as A, B, C, and D, is divided into different areas, such as A, B, C, and D, which are called light and dark. The histogram is a graph showing the number of pixels belonging to each contrast based on the brightness of each contrast. Brightness can be expressed numerically on a scale of 0 to 255. For example, the histogram may be graphed as a histogram a in which the region A occupies 3/8 of the entire image, as shown in FIG. 5B. In the same way, histograms b, c and d can also be graphed. At this time, as shown in Fig. 5c, when the histograms (a, b, c, d) of each of the contrasts (A, B, C, D) are accumulated and shown in a cumulative distribution diagram, a slope of 0 to 1 has a slope of 1 to 1. You can see that.

According to a further embodiment, the histogram generator 1220 may perform an operation of equalizing the histogram. For example, the histogram generator 1220 may perform the task of equalizing the histogram. When B, C, and D) are represented up to 255 points with the maximum brightness, the histogram d corresponding to the contrast D having the maximum brightness of the acquired image is positioned at 255 points, and the remaining histograms a, b, and c are It can be shifted in proportion to the histogram d.

The first histogram analyzer 1222 calculates a contrast value from the histogram generated by the histogram generator 1220 and determines that the image is invalid when the calculated contrast value is smaller than a predetermined threshold. The contrast can be obtained from the difference between the maximum value and the minimum value of the pixel value. In this case, the lighting controller 124 may increase the brightness of the lighting when the contrast value is smaller than the predetermined threshold as a result of the determination of the first histogram analyzer 1222. An embodiment thereof will be described later with reference to FIGS. 6A and 7A.

The second histogram analyzer 1224 calculates a pixel value distribution from the histogram generated by the histogram generator 1220, and determines that an image is invalid when the distribution of pixel values is concentrated. At this time, the lighting controller 124 lowers the brightness of the illumination when a pixel having a brightness greater than a predetermined threshold is concentrated as a result of the determination of the second histogram analyzer 1224, and the pixel having a brightness smaller than the predetermined threshold is reduced. If it is concentrated, it can increase the brightness of the light. An embodiment thereof will be described later with reference to FIGS. 6B and 7B.

4 is a block diagram of a face recognition device 1b according to a further embodiment of the present invention.

Referring to the difference between the face recognizing apparatus 1b of FIG. 4 and the face recognizing apparatus 1a of FIG. 2, the feature recognition extracting unit 126 is further included in the face recognizing apparatus 1b of FIG. 4.

The feature information extractor 126 extracts feature information from the face region extracted by the face region detector 120. The feature information refers to information representing unique characteristics of the face included in the face region. For feature information extraction, LBP (Local Binary Pattern) can be used to express texture information using local image information. LBP can extract feature information that is robust to changes in lighting and is widely used to express facial features. As another example, Principal Component Analysis (PCA), Linear Discriminate Analysis (LDA), or the like may be selected and used. As another example, 2-Dimensional Principal Component Analysis (2D-PCA) may be used to extract feature information of a face region based on a two-dimensional image matrix. Since the size of the projection matrix is smaller than that of the general PCA method, the 2D-PCA technique can be easily applied to portable devices, low-end embedded systems, etc., because the amount of computation and required memory can be relatively low. Therefore, face recognition can be performed within a short time, so that face recognition can be performed in real time.

6 is a reference diagram illustrating an example of an input image requiring adaptive lighting adjustment according to an embodiment of the present invention.

As shown in FIG. 6A, in a dark lighting environment, sufficient contrast is not secured, so that a face cannot be properly recognized. The contrast is obtained from the difference between the maximum value and the minimum value of the pixel value. In addition, even in the case of saturation in which pixels having an image brightness value of a specific value or more are concentrated as shown in FIG. 6B, a face cannot be properly recognized.

FIG. 7 is a reference diagram illustrating a histogram for explaining an histogram analysis example of the image analyzer 122.

The first histogram analyzer 1222 of the image analyzer 122 calculates a contrast value from the histogram generated by the histogram generator 1220, and if the calculated contrast value is smaller than a predetermined threshold, the input image is invalid. Determine that it is not. An example of the input image at this time is as shown in (a) of FIG. 6, the histogram is as shown in (a) of FIG. This is because the dark lighting did not ensure sufficient contrast. Therefore, the lighting controller 124 increases the brightness of the light when the contrast value is smaller than the predetermined threshold value according to the determination result of the first histogram analyzer 1222, so that the image acquirer 10 may adjust the light whose brightness is adjusted. Control to reacquire the image.

On the other hand, the second histogram analyzer 1224 of the image analyzer 122 calculates a pixel value distribution from the histogram generated by the histogram generator 1220, and if the distribution of pixel values is concentrated, the image is invalid. Determine. When the distribution of the pixels is concentrated in the high brightness region, it is as shown in FIG. 6B and the histogram is as shown in FIG. 7B. This is because the brightness of the image is too bright. Therefore, the lighting controller 124 lowers the brightness of the illumination when pixels having a brightness greater than a preset threshold are concentrated according to the determination result of the second histogram analyzer 1224. In contrast, when the pixels having a brightness smaller than the preset threshold are concentrated, the brightness of the illumination may be increased.

8 is a reference diagram for describing an adaptive lighting control process when feature information is not extracted according to an embodiment of the present invention.

When there is a sudden feature change in the face region or when there are many values having a planar feature in the feature distribution, sufficient feature information is not extracted. In this case, the lighting controller 124 adjusts the brightness of the lighting. That is, in the input image illustrated in FIG. 8A, the pixel value distribution is uniform in the histogram as shown in FIG. 8B, but the distribution of the feature information is not good. In this case, the light control unit 124 brightly adjusts the brightness of the light to extract the feature information.

9 is a flowchart illustrating a face recognition method using adaptive lighting control according to an embodiment of the present invention.

According to an embodiment, the face recognition apparatuses 1a and 1b acquire an image by capturing a subject under illumination in operation 900 and detect a face region from the acquired image in operation 910. Subsequently, the histogram of the detected face region is analyzed to determine the validity of the acquired image (920), and the brightness of the illumination is adjusted (940) according to the determination result. Reacquire

According to an exemplary embodiment, the face recognizing apparatuses 1a and 1b generate a histogram representing the pixel value and the brightness of each contrast of the detected face region, calculate a contrast value from the generated histogram, and calculate the contrast value. If less than the predetermined threshold, it is determined that the image is invalid. In this case, when the contrast value is smaller than the predetermined threshold, the brightness of the illumination may be increased.

According to another exemplary embodiment, the face recognizing apparatuses 1a and 1b generate histograms of the detected face regions, calculate a pixel value distribution from the generated histogram, and determine that the image is invalid when the distribution of pixel values is concentrated. Determine. In this case, when the pixels having a brightness greater than the preset threshold are concentrated, the brightness of the illumination may be lowered, and when the pixels having the brightness smaller than the preset threshold are concentrated, the brightness of the illumination may be increased.

According to another embodiment, the face recognizing apparatus 1b extracts feature information from the detected face area (930). If the feature information cannot be extracted according to the feature information extraction result, the brightness of the lighting is adjusted.

The embodiments of the present invention have been described above. Those skilled in the art will appreciate that the present invention can be implemented in a modified form without departing from the essential features of the present invention. Therefore, the disclosed embodiments should be considered in an illustrative rather than a restrictive sense. The scope of the present invention is defined by the appended claims rather than by the foregoing description, and all differences within the scope of equivalents thereof should be construed as being included in the present invention.

1a, 1b: face recognition device 10: image acquisition unit
12a, 12b: control unit 14: user input unit
16: image output unit 120: face area detection unit
122: image analysis unit 124: lighting control unit
126: feature information extraction unit 1220: histogram generator
1222: first histogram analyzer 1224: second histogram analyzer

Claims (8)

An image acquisition unit which acquires an image by photographing a subject under illumination;
A face region detector for detecting a face region from the obtained image;
An image analyzer which analyzes the histogram of the detected face region to determine the validity of the acquired image; And
An illumination control unit for controlling the image acquisition unit to re-acquire the image using the light whose brightness is adjusted by adjusting the brightness of the illumination according to the determination result of the image analyzer;
And a face recognition unit for recognizing the face. The method of claim 1, wherein the image analyzer,
A histogram generator for generating a histogram representing pixel values and brightness of each contrast of the detected face region; And
A first histogram analyzer which calculates a contrast value from the generated histogram and determines that an image is invalid when the calculated contrast value is smaller than a predetermined threshold;
And a face recognition unit for recognizing the face. The method of claim 2, wherein the lighting control unit,
And a brightness of the illumination is increased when the calculated contrast value is smaller than a predetermined threshold value according to the determination result of the first histogram analyzer. 3. The method of claim 2,
A histogram generator for generating a histogram representing pixel values and brightness of each contrast of the detected face region; And
A second histogram analyzer which calculates a pixel value distribution from the generated histogram and determines that an image is invalid when the distribution of pixel values is concentrated;
And a face recognition unit for recognizing the face. The method of claim 4, wherein the lighting control unit,
According to the determination result of the second histogram analyzer, when the pixels having a brightness greater than a predetermined threshold are concentrated, the brightness of the illumination is lowered, and when the pixels having a brightness smaller than the preset threshold are concentrated, the brightness of the illumination is increased. Face recognition device. The method of claim 1,
A feature information extracting unit extracting feature information from the detected face region; More,
And the lighting controller adjusts the brightness of the lighting when the feature information cannot be extracted according to the feature information extraction result of the feature information extractor. The method of claim 1,
And wherein the illumination is active illumination forming an ambient illumination. The method of claim 1,
An image output unit which outputs an image acquired from the image acquisition unit;
Face recognition apparatus further comprises a.

Images