KR100901125B1 - Face detecting method and apparatus - Google Patents

Abstract

The present invention relates to a face recognition method and apparatus for preventing contamination of feature information of a face by a lighting environment of an entire subject and a decrease in recognition rate. A process of calculating a distribution range in the image of the face region, a process of measuring brightness of a subject part corresponding to the calculated distribution range of the face region, and setting exposure conditions corresponding to the measured brightness and receiving the next image Identifying the face. According to the face recognition method and apparatus of the present invention, the brightness of the image coming from the camera module in the real environment is set as the face reference to minimize the contamination of the feature information required for face recognition. Therefore, the accuracy and recognition rate of face recognition can be improved, so that related application operations can be performed smoothly.

Face Recognition, Biometrics, Mobile Terminal, Camera, Preview

Description

Face detecting method and apparatus

1 is a block diagram of a face recognition device according to an embodiment of the present invention;

2 is a flowchart illustrating a face recognition method according to a first embodiment of the present invention;

3 is a flowchart illustrating a face recognition process of the face recognition method according to the first embodiment of the present invention;

4 is a conceptual view showing a face recognition method according to a first embodiment of the present invention;

5 is a flowchart of a face registration process in the face recognition method according to the first embodiment of the present invention;

6 is a flowchart illustrating a face recognition method according to a second embodiment of the present invention;

7 is a flowchart illustrating a face recognition method according to a third embodiment of the present invention;

8 is a detailed flowchart of an image input process in a face recognition method according to a third embodiment of the present invention; and

9 is a detailed flowchart of an exposure compensation value setting process in the face recognition method according to the third embodiment of the present invention.

The present invention relates to a face recognition method and apparatus, and more particularly, to a face recognition method and apparatus for searching for and identifying a face in an image input through a camera module.

Biometric technology is a technology that identifies a person using biometric information. Until now, the most widely used method is to acquire an image of a part of the body and extract and compare features. Examples of biometric technologies include fingerprint recognition, face recognition, retina recognition, iris recognition, vein recognition, palm print recognition, voice recognition, and the like.

Among them, face recognition technology is a technology that uses a face to identify a person, and is widely used to verify an individual's characteristics or to verify an identity. Currently, a face recognition method is well known for detecting a face region from an image input through a camera module, extracting feature information from the detected face region, and comparing it with feature information of a face registered in a database. .

On the other hand, in order to increase the face recognition rate, an image must be acquired in an illumination environment having the same brightness and direction as when registering an image. However, in the face recognition environment, the video coming through the camera module is not. For example, if you take a picture in very bright outdoors, you can see people with white faces appear dark in the picture. This is because the brightness of the image is determined by the camera. If the entire image containing the face is bright or dark, the feature information of the face to be recognized may be contaminated, and thus the recognition rate may be significantly reduced. Therefore, there is a need for a method for preventing contamination of facial feature information in various environments including light and background.

An object of the present invention is to provide a face recognition method and apparatus capable of preventing contamination of feature information of a face by a lighting environment of the entire subject and thus lowering a recognition rate.

In order to achieve the above object, the present invention provides a face recognition method of an image input from a camera module, the process of searching for a face region by receiving an image from the camera module, calculating the position and size in the image of the detected face region. A process of measuring a brightness of a subject portion corresponding to the calculated position and size of the face region, a process of receiving a next image by setting an exposure condition corresponding to the brightness of the subject region, and a face region from the next input image. It provides a face recognition method comprising the step of extracting the face identification.

In addition, in order to achieve the above object in the present invention, in the face recognition method of the image input from the camera module, the process of extracting the face region by receiving the image from the camera module, by extracting the facial feature information from the face region of the input image A face recognition process for determining whether or not to register the image, a process of measuring a part of a subject within the position and size range of the extracted face region when the image is detected as an unregistered face by using a metering window, and metering brightness. A face recognition method comprising setting a corresponding exposure condition and receiving a next image and performing a face recognition process again is provided.

In order to achieve the above object, the present invention provides a face recognition device having a camera module, the exposure control unit for controlling the camera module to be set to alternately set the exposure conditions respectively corresponding to the brightness of the image and the brightness of the face in the image, If the image input through the camera module is an image that is input as the exposure condition corresponding to the overall brightness of the image, the face area is extracted, and the brightness of the subject part corresponding to the position and size is measured so that the exposure condition corresponding to the brightness is set. According to an aspect of the present disclosure, there is provided a face recognition apparatus including a face recognition unit for controlling a exposure control unit and recognizing a face when the image is received as an exposure condition corresponding to the brightness of the face in the image.

Hereinafter, exemplary embodiments of a face recognition method and apparatus will be described in detail with reference to the accompanying drawings. However, in the drawings and the description of the embodiments, illustration and description of parts not related to the central technical idea of the present invention will be omitted so as not to obscure the gist of the present invention.

In addition, prior to the description of the embodiment, the term "feature information" in the present specification is to be understood that it means the unique feature information that each individual has to be distinguished from other people in the face of a person. For example, the facial feature may include eye, nose, mouth, and a specific area of the skin. In addition, the term "standby mode" is understood to mean the operation mode of the state before performing the face recognition.

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

Referring to FIG. 1, the face recognition apparatus 100 of the present exemplary embodiment includes a camera module 10, a controller 20, a storage unit 30, a key input unit 40, and a display unit 50.

The camera module 10 acquires image data and transmits the image data to the controller 20. The camera module 10 has an image sensor 11, an image signal processor 12, and a lens unit 13. The image sensor 11 converts the optical signal of the subject input through the lens unit 13 into an electrical signal. The image sensor 11 may be a CCD sensor, a CMOS sensor, or the like. The lens unit 13 performs a function of projecting an image and may include a plurality of optical lenses. The image signal processor 12 converts an analog signal input from the image sensor 11 into digital data. Also, the image signal processor 12 calculates an auto exposure target under the control of the controller 20, measures the brightness of the subject, and sets the exposure accordingly. The image signal processor 12 may be implemented as an ISP (Image Signal Processor), and may be installed in the image sensor 11 or installed outside the camera module 10.

The storage unit 30 performs a function of storing the face database 31. The face database 31 may have the facial feature information data of the registrant, and may have a facial image or related information data of the registrant. The face database 31 may register or update data under the control of the controller 20. The storage unit 30 detects the face area in the face recognition mode and stores programs and data for automatically correcting the exposure according to the brightness of the face area. The storage unit 30 also provides a storage location for processing an image input from the camera module 10. In addition, the storage unit 30 stores various programs and data for controlling the overall operation of the face recognition apparatus 100 and performs a temporary storage function of data generated during the operation. The storage unit 30 may be various memory devices such as a flash memory and a hard disk.

The key input unit 40 generates a command signal related to the overall operation of the face recognition apparatus 100 according to the key input. The key input unit includes keys for inputting numeric or character information and function keys for setting various functions. The key input unit 40 may be implemented as a keypad, a touch screen, a touch pad, or the like. Here, the key input unit 40 may include a key for entering the face recognition mode and a shutter key for photographing in the photographing mode.

The display unit 50 controls the control unit 20 to output image data to the screen. The display unit 50 displays the captured image of the subject on the screen. Here, the image of the subject may be a preview image, a captured image, and a stored image. In addition, the display unit 50 may display information about an operation, an operation, and a result of the face recognition apparatus 100. The display unit 50 may be implemented as a display device such as a liquid crystal display (LCD), organic light emitting diodes (OLED), a plasma display panel (PDP), or the like.

The controller 20 has a signal processor (not shown) for converting an image signal output from the camera module 10 into image data. The signal processor may be implemented as a digital signal processor (DSP). The controller 20 may have an image processor (not shown) for converting image data output from the signal processor into screen data for display on the display unit 50. The image processing unit performs size adjustment, color conversion, image compression, and the like. In addition, the controller 20 sets a face recognition mode according to a key input and controls the display unit 50 to output an image input through the camera module 10. In addition, the controller 20 controls the overall operation of the face recognition apparatus 100.

In particular, the controller 20 searches for and extracts a face region from an input image, extracts facial feature information, pre-registers a face database, calculates brightness for a face region, calculates an exposure compensation value, and controls exposure setting. To this end, the controller 20 may have an exposure controller 21 and a face recognition unit 22. The face recognizer 22 may include a face region extractor 23, a face region calculator 24, a feature information extractor 25, a registration determiner 26, and the like.

The face region extractor 23 searches for and extracts a face region from the input image. In the search and extraction of the face region, a face extraction technique using contour, skin color, texture, template, or the like may be used. For example, the face region extractor may find a face region using the skin color in the input image. The feature information extractor includes a transducer for emphasizing the color of the face, a transducer for emphasizing black and white, a preprocessor to enhance the image of a specific part to extract the feature information of the face, and a background labeling for separating the face region and the background region ( labeling) group and the like.

The face area calculator 24 calculates the distribution position and size of the face area searched by the face area extractor 23 in the input image.

The feature information extractor 25 extracts feature information of the face from the face region extracted by the face region extractor. Facial feature information extraction may use well-known Principal Component Analysis (PCA) techniques or template matching techniques. For example, facial feature information extraction is performed by extracting a multiple PCA vector by performing principal component analysis on a set face region, and analyzing an edge distribution of a face image to obtain an edge distribution vector. Can be extracted.

The registration determining unit 26 compares the feature information of the face extracted from the input image with the feature information of the face stored in the face database 31 to determine whether the registration is already performed and output the result.

The exposure controller 21 controls the image signal processor 12 of the camera module 10 to control the exposure setting of the camera module 10. The exposure control unit 21 transmits the metering window information to be set as the automatic exposure target in the camera module 10 to the camera module 10 and controls the exposure setting according to the metering result. The metering window information may be position and size information of the face area acquired by the face area calculator 24. In addition, the exposure controller 21 may control the exposure compensation of the camera module 10 by calculating the brightness of the face area in the input image and calculating the exposure compensation value based on the value. Here, the brightness of the face area is the average brightness of pixels located in the face area.

The exposure controller 21 controls the exposure setting of the camera module 10 by calculating an exposure compensation value of the next input image based on the brightness of the face region. Here, the exposure compensation value may be an exposure compensation value, or a value corresponding to a difference between the general average exposure value or the face database average exposure value and the exposure value corresponding to the brightness of the face area, or a predetermined level thereof. Alternatively, the exposure compensation value may be an exposure value of a preset increase or decrease ratio.

The embodiment of the face recognition device according to the present invention has been described so far. Hereinafter, an embodiment of a face recognition method according to the present invention will be described.

2 is a flowchart illustrating a face recognition method according to a first embodiment of the present invention.

Referring to FIG. 2, in the process of performing the standby mode (S201), the controller 20 checks a face recognition request according to a key input (S203). Here, the face recognition request may be a signal for performing interworking with another application or a key input generated while performing another application. In this process, the controller 20 may output a menu screen to the display unit 50 to guide execution of face recognition.

When the face recognition request is confirmed, the controller 20 activates the function of the camera module 10 to receive an image (S205). In this process, the exposure controller 21 of the controller 20 sets an exposure condition based on the brightness of the entire subject and receives an image by an auto exposure (AE) function. The exposure conditions can be set by measuring and metering an autoexposure target for the entire subject. For example, by performing multi photometry on the subject, an exposure condition corresponding to the entire brightness of the subject may be set. The image input process may be a preview process of a camera photographing mode or an image input process for security authentication in a security camera. The controller 20 may output the input image to the display unit 50 so that the user can check it.

When the image is input, the controller 20 searches for a face region in the input image (S209). As described above, the controller 20 may search for a face region in the input image by applying a face extraction technique using an outline, skin color, texture, template, and the like. When the preview image is input, the controller 20 may convert the input preview image into preview screen data and extract a face region from the preview screen data.

Subsequently, the controller 20 checks whether the searched face area exists (S211). If it is determined that the searched face area does not exist, the controller 20 ends face recognition. If it is determined that the face region exists, the controller 20 calculates distribution information of the face region in the image (S213). For example, the controller 20 calculates the position and size of the face area in the image.

Subsequently, the controller 20 sets the photometric window corresponding to the distribution range of the face region (S215). The distribution information (position, size) is calculated by analyzing the distribution information of the face area, and the photometric window belonging to the distribution range is set. The metering window may have a rectangular shape or other shape. In addition, the light metering window may be smaller than the face area in consideration of human movement. When the metering window is formed in a rectangle, it is smaller than the rectangle inscribed in the face area.

When the metering window is set, the controller 20 transmits the metering window information (position, size, etc.) to the camera module 10 and performs metering for measuring the brightness of the corresponding subject part (S217). In this process, the brightness of the face existing in the subject is measured. This photometric process may be replaced by a process of calculating an average brightness of an area corresponding to the photometric window in the input image.

When the metering is completed, the controller 20 proceeds to set exposure of the camera module 10 under an exposure condition corresponding to the measured brightness (S219). Under the control of the controller 20, the camera module 10 is set to an exposure condition corresponding to the brightness of the face area.

When the exposure setting is completed, the control unit 20 receives an image generated from the camera module 10 (S221). In this case, the controller 20 may receive an image in which the exposure is adjusted to the face area.

When the image is input, the controller 20 extracts a face region from the input image and performs face recognition (S223). Since the input image is an image in which the exposure is matched to the face area, the controller 20 may extract face feature information accurately and perform face recognition. The facial recognition process is described in detail below.

When the face recognition is completed, the control unit 20 confirms the end of the face recognition request of the above-described work (S239). If the end request is confirmed, the face recognition is terminated. Otherwise, the process proceeds again from the image input process S205.

3 is a flowchart illustrating a face recognition process in the face recognition method according to the first embodiment of the present invention.

Referring to FIG. 3, in the face recognition process (S223), first, the controller 20 extracts a face region from the input image (S224). The face region extractor 23 may extract a face region by using an outline, skin color, texture, template, and the like.

Subsequently, the controller 20 checks whether the extracted face region exists (S226). If there is no extracted face area, the controller 20 ends face recognition. If the extracted face area exists, the controller 20 extracts face feature information from the extracted face area (S228). The feature information extractor 25 may extract facial feature information by principal component analysis or template matching.

Subsequently, the controller 20 compares the extracted facial feature information with the facial feature information stored in the face database 31 to determine whether the face is a registered face (S230). If it matches the registered facial feature information, and outputs the registered face to the display unit 50 (S234). In this case, the controller 20 may display an image or related information of the corresponding face on the display unit 50. When the face recognition mode is performed in the preview process, the controller 20 may display the face recognition result along with the preview screen data on the preview screen.

On the other hand, if it is determined that the face is not registered, the control unit 20 outputs the unregistered face to the display unit 50 (S232). In this case, the controller 20 may guide the screen through whether to register a face and determine a request for face registration (S236). If there is a face registration request, the controller 20 may register the face database 31 (S238). If there is no registration request, the following process proceeds. Meanwhile, a process of confirming a face registration request (S236) and a process of registering a face database (S238) may be omitted.

In the face recognition method according to the first embodiment of the present invention as described above, the face region is extracted from the input image, the exposure condition corresponding to the face region is set, and the face recognition is performed with the next input image. Face recognition is received by receiving an image with the exposure set based on the face, which increases the accuracy of face recognition and improves the face recognition rate. Meanwhile, the aforementioned face recognition method may be repeatedly performed with one cycle from the image input process S205 to the face recognition process S223 until the end of the face recognition mode request.

4 is a conceptual diagram illustrating a face recognition method according to a first embodiment of the present invention.

For example, suppose the real face recognition environment is very bright. The controller 20 darkens the exposure setting of the input image by controlling the camera module 10 in order to obtain an overall clear image because the entire subject is bright. As a result, the brightness of the face is darker than in the real environment.

When the object 91 as shown in FIG. 4 is input through the camera module, the face image is dark in the primary input image 92 as shown in FIG. 4. In this image 92, the control unit 20 searches for a face region. After the search, the controller 20 acquires distribution information of the face region in the image and sets the photometric window Wpm as shown in FIG. 4. When the photometric window Wpm is metered and the secondary image is input, as shown in FIG. 4, the secondary input image 93 whose exposure is matched to the face region may be obtained. When face recognition is performed, accurate facial feature information can be extracted and recognition rate is increased.

Meanwhile, in the face recognition method of the present invention, a process of registering a face in a face database must be preceded. This process is described below.

5 is a flowchart illustrating a face registration process in the face recognition method according to the first embodiment of the present invention.

Referring to FIG. 5, while performing the standby mode S301, the controller 20 confirms a face registration request according to a user selection (S303). When the face registration request is confirmed, the controller 20 performs a preview of activating the camera module to receive an image and output the image to the display unit 50 (S305).

During the preview, the controller 20 checks the photographing key input generated by the key input unit 40 (S307). When the photographing key input is confirmed, the controller 20 receives image data from the camera module 10 at the key input time point and outputs the photographed image to the display unit 50 as a photographed image (S309). The output of the captured image continues until user selection is made.

Subsequently, the controller 20 confirms a face registration request for the image output to the display unit 50 (S311). If it is determined that there is a registration request, the controller 20 extracts a face region from the input image (S313). The controller 20 extracts feature information of the face from the face region (S315), and determines whether the feature information is registered in the face database 31 (S317). If there is no registered data, the controller 20 registers facial feature information to the face database 31 (S319). In this case, a face image, identification information, and related information may be registered in the face database 31. If there is already registered data, it informs that it is a registered face (S321).

On the other hand, unlike the first embodiment described above, the face recognition method of the present invention receives an image by automatic exposure and performs face recognition first, and when it is determined that the face is not registered in the face database, the metering of the face and its corresponding After setting the exposure conditions, it is also possible to face-detect the next input image. Such an embodiment is described below. In the following description, the description overlapping with the description of the first embodiment will be omitted.

6 is a flowchart illustrating a face recognition method according to a second embodiment of the present invention.

Referring to FIG. 6, in the standby mode S401, the controller 20 confirms a request for the face recognition mode according to a key input (S403). When the request for the face recognition mode is confirmed, the controller 20 activates the camera module function and receives an image through the camera module (S405). In this process, the control unit 20 sets the entire subject as an autoexposure target by the autoexposure function and performs light metering to receive an image.

When the image is input, the controller 20 extracts a face region from the input image (S407). The controller 20 checks whether the extracted face region exists (S409). If it is determined that the extracted face region does not exist, the process of receiving an image (S405) is performed again. If it is determined that the face region exists in the input image, the controller 20 extracts face feature information from the face region of the input image (S411).

Next, the extracted facial feature information is compared with the data of the face database 31 to determine whether it is a registered face (S413). If the determination result is the registered face, the face is output to the display unit 50 (S415). In this case, the controller 20 may output an image or related information of the corresponding face.

If the determination result is that the face is not registered, the control unit 20 determines whether an unregistered determination order is 2 or less (S417). Here, the unregistered determination order may be increased or decreased as necessary. If it is determined to be 2 or less in this process, the control unit 20 calculates distribution information (position, size, etc.) of the face area in the image (S419). The control unit 20 sets a photometric window corresponding to the calculated distribution information of the face area (S421). Subsequently, the photometric window information is transmitted to the camera module 10 to perform photometry for measuring the brightness of the subject part (S423).

When the metering is completed, the control unit 20 controls the camera module to set the exposure to the exposure condition corresponding to the metering result (S425). The controller 20 sets the exposure conditions corresponding to the brightness of the face area. If there is no termination request in the process of checking the end of the face recognition request (S429), the controller 20 determines whether the face is registered in the face database in the process of receiving an image from the camera module (S405) Repeat until (S413). By performing the above process again, face recognition is performed on the face in the image. On the other hand, if it is determined to be 2 in the process of determining whether the unregistered determination order is 2 or less (S417), the control unit 20 displays that it is an unregistered face (S427). If there is an end request in the process of confirming the end request of the face recognition (S429), the face recognition ends.

The second embodiment of the face recognition method according to the present invention has been described above. Unlike the first embodiment, the face recognition method of the present invention performs face recognition again by performing a certain level of exposure correction on the basis of the face when it is determined as an unregistered face after face recognition on the input image. Accordingly, the accuracy of face recognition can be improved and the face recognition rate can be improved. In the meantime, in the face recognition method according to the present invention, face recognition may be performed by performing exposure compensation according to the preset exposure compensation value as in the third embodiment.

7 is a flowchart illustrating a face recognition method according to a third embodiment of the present invention.

Referring to FIG. 7, in the standby mode S501, the controller 20 confirms a face recognition request according to a key input (S503). When the face recognition request is confirmed, the controller 20 activates the camera module function to receive an image (S505).

Subsequently, the controller 20 extracts a face region from the input image (S509). In operation S511, the controller 20 determines whether the extracted face region exists, and if the face region exists in the input image, the controller 20 extracts feature information of the face from the face region (S513). If the face area does not exist, the controller 20 proceeds again in step S505 of receiving an image.

When face feature information is extracted, the controller 20 compares the extracted feature information with data stored in the face database 31 to determine whether a face in the input image is a face registered in the face database 31 (S515). . If it is confirmed that the registered face, the control unit 20 outputs the registered face to the display unit 50 (S517).

If it is determined that the face is not registered, the controller 20 checks whether the correction order is less than or equal to the set order (S519). Here, the setting order may be provided as a setting item for the face recognition mode and may be selected by the user. Alternatively, it may be provided as a default at the manufacturing stage of the face recognition device.

If the correction order is less than or equal to the set order, the controller 20 calculates the brightness of the extracted face region (S521) and calculates an exposure correction value based on the calculated brightness (S523). When the exposure compensation value is calculated, the control unit 20 performs exposure setting control for correcting the exposure of adding or subtracting the exposure compensation value to the exposure value by automatic exposure (S525). The exposure compensation value calculation process will be described later in detail.

Meanwhile, in step S519, the controller 20 notifies that the face is unregistered if the correction order is the set order (S527). The exposure compensation value is set to 0 (S529).

Subsequently, the controller 20 checks whether there is a request for terminating face recognition (S531). If there is no termination request, the control unit 20 performs the operation again from the process of receiving an image (S505). In the case of an unregistered face, exposure compensation is performed up to a set order to recognize the face of the input image in which contamination of the facial feature information is prevented.

In the above-described face recognition method according to the third embodiment, the exposure condition value is set by reflecting the exposure compensation value of a preset condition when discriminating as an unregistered face, and the face recognition is performed again by receiving a new image. Such face re-recognition may be repeatedly performed a predetermined number of times in a manner of checking whether the degree of exposure compensation has been set. Repetitive exposure correction and face recognition may improve the accuracy of face recognition. Hereinafter, a process of receiving an image in the face recognition method of the above-described third embodiment will be described in more detail below.

8 is a detailed flowchart of an image input process in the face recognition method according to the third embodiment of the present invention.

Referring to FIG. 8, the controller 20 sets an entire subject as an autoexposure target in the image input process and performs light metering to calculate an exposure value (S551). An exposure value corresponding to the average brightness of the entire input image may be calculated. At this time, the exposure value is called an automatic exposure value.

Next, the controller 20 checks whether there is exposure compensation (S53). If there is exposure compensation, the controller 20 adds or subtracts a preset exposure compensation value to the automatic exposure value to set the exposure (S555). The exposure compensation may be performed by omitting the process of calculating the automatic exposure value and adding or subtracting the exposure compensation value to the previously obtained automatic exposure value. If there is no exposure compensation in step S553, the control unit 20 sets the exposure to the automatic exposure value (S557). When the exposure setting is completed, the controller 20 generates an image by controlling the camera module (S559) and receives the image (S561).

The exposure compensation value setting process in the face recognition method of the third embodiment will be described in more detail below. In the following description, the average average exposure value refers to an average exposure value determined by statistics on brightness values of a general picture. The average exposure value of the face database means an exposure value corresponding to an average value of brightnesses of faces registered in the face database.

9 is a detailed flowchart of an exposure compensation value setting process in the face recognition method according to the third embodiment of the present invention.

Referring to FIG. 9, in operation S521 of setting an exposure correction value, the controller 20 calculates an exposure value corresponding to brightness of a face area (S571). The controller 20 determines whether the preset exposure compensation form is a quantitative exposure compensation (S573). If it is set to quantitative exposure compensation, the control unit 20 sets the exposure compensation value to the preset exposure compensation level (S575).

On the other hand, if it is determined that the set exposure correction form is not quantitative correction, the controller 20 checks whether the setting of the exposure correction form is a general average correction (S577). If the average correction is normal, the controller 20 calculates a difference value by subtracting the exposure value calculated in step S571 from the general average exposure value (S579). If the correction form is not the average average exposure, the difference value is calculated by subtracting the correction value calculated by the S571 process from the average exposure value of the face database 31 (S581).

When the calculation of the difference value is completed, the controller 20 checks whether the preset exposure compensation level is set ratio correction (S567). If it is confirmed by the set ratio correction, the exposure value corresponding to the preset ratio from the difference value is set as the exposure compensation value (S585). Each time the face re-recognition is repeatedly performed, it increases or decreases step by step in the difference between the total image exposure value and the face region exposure value. On the other hand, if it is not the set ratio correction, the control unit 20 sets the difference value to the exposure correction value (S569).

For example, assume that the exposure compensation form is set to quantitative correction. Assume that the average exposure value is 135, the exposure value based on the face area is 145, and the set exposure compensation level is 5. The controller 20 subtracts the exposure value 145 based on the face area from the general average exposure value 135. The difference is -10. Since the negative difference value, the control unit 20 sets -5 as the exposure compensation value and receives the next image as the exposure value darkened by 5 from the automatic exposure value.

As can be seen from the above-described embodiments, the face recognition method and apparatus of the present invention recognizes the face with an image set according to the brightness of the face using a continuous input image. Alternatively, face recognition is performed again on an unregistered face with an image obtained by setting exposure based on a face. Contamination of characteristic information of the face region by the brightness of the entire subject can be minimized, so that accurate face recognition results can be obtained. Typically, an image is output at a speed of 10-15 frames / sec from the camera module. Images are displayed one frame at 0.067-0.1 seconds. Therefore, it is possible to recognize the face of the same person as a continuous input image. The speed at which the image is input may be appropriately set in consideration of the moving speed of the person, the number of times of face recognition, the number of exposure compensation, and the like. When a plurality of faces exist in the input image, face recognition may be sequentially performed for each face.

Meanwhile, the embodiments of the face recognition method and apparatus according to the present invention have been described above. However, the face recognition method and apparatus of the present invention are not limited to the above-described embodiment. Various modifications can be made without departing from the spirit of the invention.

For example, the face recognition apparatus of the present invention can be applied to a portable terminal having a digital camera, a security camera or a camera module. Here, the mobile terminal having a camera module includes a mobile communication terminal, a mobile phone, a personal digital assistant (PDA), a smart phone, an IMT-2000 (International Mobile Telecommunication 2000) terminal, and a UMTS (Universal Mobile). Telecommunication service terminals, digital broadcasting terminals, information communication devices such as notebooks and personal computers, and multimedia devices.

Such facts will be readily apparent to those skilled in the art. Therefore, the scope of the present invention should not be limited to the embodiments described above, but should be defined not only by the claims below but also by the equivalents of the claims.

As described above, according to the face recognition method and apparatus according to the present invention, the brightness of an image coming from a camera module in a real environment is set as a face reference to minimize contamination of feature information required for face recognition. Therefore, the accuracy and recognition rate of face recognition can be improved, so that related application operations can be performed smoothly.

Claims (18)

In the face recognition method of the image input from the camera module, A search process of searching for a face region by receiving an image from the camera module; Calculating a distribution range in the image of the searched face region; A photometric process of measuring brightness of a subject part corresponding to the calculated distribution range of the face region; and And a face recognition process of setting an exposure condition corresponding to the measured brightness and identifying a face by receiving an image under the set exposure condition. The method of claim 1, wherein the searching process is performed. An image input process for receiving an image by setting exposure conditions corresponding to the overall brightness of the subject; And a face region searching process of searching for a face region in the input image. The method of claim 2, wherein the image input process is performed. Face recognition method characterized in that the process of receiving a preview image (preview image) from the camera module. The method of claim 1, wherein the photometric process is Calculating a distribution range in the image of the face region; Setting a photometric window corresponding to a distribution range in the image of the face region, and And measuring brightness of a subject part corresponding to the photometric window. The method of claim 4, wherein the setting of the photometric window is performed. And a light metering window smaller than the size of the face area within the distribution range of the face area. In the face recognition method of the image input from the camera module, An input process of receiving an image from the camera module; A face recognition process of extracting a face region from the input image to determine whether or not a face is registered; A metering process of measuring brightness of a subject part corresponding to a distribution range of the face area when discriminating as unregistered, and And a face recognizing process of setting an exposure condition corresponding to the measured brightness and receiving an image under the set exposure condition and performing the face recognition process again. The method of claim 6, wherein the input process An automatic exposure setting process of setting the camera module to an exposure condition corresponding to the entire brightness of the subject; And a face image input process of receiving an image generated under the exposure condition. The method of claim 7, wherein the image input process Face recognition method characterized in that the process of receiving the preview image from the camera module. The method of claim 6, wherein the photometric process is Calculating a distribution range in the image of the face region; Setting a photometric window corresponding to a distribution range in the image of the face region, and Face recognition method characterized in that the process of measuring the brightness of the subject portion corresponding to the metering window. The method of claim 9, wherein the setting of the photometric window is performed. And a light metering window smaller than the size of the face area within the distribution range of the face area. The method of claim 6, wherein the facial recognition process is performed. And performing a face recognition process again, if it is determined that the face is not registered, informing that the face is an unregistered face. In the face recognition method of the image input from the camera module, An input process of receiving an image from the camera module; A face recognition process of extracting a face region from the input image to determine whether or not a face is registered; When it is determined as not registered, the exposure compensation value corresponding to the preset condition corresponding to the result of subtracting the exposure value of the face area corresponding to the brightness of the extracted face area from the exposure value corresponding to the face average brightness of the face database and the brightness of the entire subject And a face recognition process of setting an exposure condition by adding corresponding exposure values, receiving an image under the set exposure condition, and performing the face recognition process again. delete The method of claim 12, wherein the facial recognition process Face recognition method characterized in that it is repeatedly performed a predetermined number of times. The method of claim 14, wherein the facial recognition process Face recognition method characterized in that the same exposure compensation value is added or subtracted step by step every iteration. The method of claim 14, wherein the exposure compensation process is performed. Face recognition method characterized in that for each iteration it increases or decreases step by step in the difference between the total image exposure value and the face area exposure value. In the facial recognition device having a camera module, An exposure controller for controlling the camera module so that exposure conditions corresponding to the overall brightness of the image and the brightness of the face area in the image are set; If the image inputted through the camera module is an image received as an exposure condition corresponding to the overall brightness of the image, the face region is extracted, and the exposure condition corresponding to the brightness is set by measuring the brightness of the subject part corresponding to the position and size. And a face recognition unit for controlling the exposure control unit and recognizing a face if the image is received under an exposure condition corresponding to the brightness of the face area in the image. In the facial recognition device having a camera module, An exposure controller for controlling the camera module so that exposure conditions corresponding to the overall brightness of the image and the brightness of the face area in the image are set; If the face recognition result for the image received through the camera module is an unregistered face, a face recognition unit for receiving the next image under the exposure conditions corresponding to the brightness of the face area in the image and proceeding face recognition again, the face characterized in that Recognition device.

Images