JP2005292994A - Person recognition device and traffic control device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To easily input an ID sign or a password sign without directly touching a device by a person to be recognized.
A unique ID sign and a password sign indicated by a registrant's action are registered in advance together with a face image, and an ID sign or password for a recognition target person registered in the registration means in an authentication process. The sign is pointed at the camera, the sign indicated by the person to be recognized is detected from the image taken by the camera, and the detected sign is collated with each registered registrant's ID sign or password sign. By doing this, the registrant is specified or the recognition target person is authenticated as a registrant.
[Selection] Figure 1

Description

  The present invention, for example, performs a person authentication device for recognizing a person by collating biometric information such as a face image, and entering / exiting important facilities (rooms, floors, buildings, etc.) and device access management using biometric collation. The present invention relates to a traffic control device.

  2. Description of the Related Art Conventionally, a person recognition apparatus has been proposed that recognizes a person using biological information (biometrics) such as a face image, fingerprint, handprint, voiceprint, handwriting, retina, and iris. Such biological information is classified into a contact type and a non-contact type. The contact-type biological information is biological information obtained when the person touches the device. Non-contact type biometric information is biometric information obtained without the person touching the device. For example, fingerprints, handprints, handwriting, and the like belong to contact-type biological information. The person recognition apparatus using such contact-type biometric information has a drawback that the person to be authenticated must touch the apparatus in order to input the biometric information.

  On the other hand, face images, voiceprints, retinas, irises, and the like belong to non-contact type biological information. Such a person recognition device using non-contact type biological information has an advantage that the person to be recognized can input the biological information without touching the device. In particular, the face image is superior to other biometric information in that the face image as biometric information can be input without the recognition target person being conscious.

  However, even if there is a person recognition device using non-contact type biometric information as described above, it is used when collation using identification information (ID information) or biometric information for narrowing the target person for collation to one person fails. The current password is entered by the person to be recognized directly touching an operation unit such as a numeric keypad. In such a case, there is a problem that the effect of inputting biometric information without contact cannot be effectively used.

As a technique for solving such a problem, Japanese Patent Application Laid-Open No. 2002-251614 (Patent Document 1) prescribes a specific operation in which eyes are closed and opened for a predetermined time, and the detection of this specific operation is a condition. As a method, a method for collating face images has been proposed. However, since the method described in Patent Document 1 is based on the condition that a specified operation is detected, registration and verification of a user's own signature is impossible, and the person to be recognized must always be specified at the time of verification. There is a problem that it is not convenient because the operation must be performed.
JP 2002-251614 A

  The present invention solves the above-mentioned problems, and a person recognition device and traffic that can easily input and operate information other than non-contact biological information without impairing convenience An object is to provide a control device.

  The person recognition device according to the present invention includes a registration unit in which a registrant's face image and a sign indicated by the registrant's own operation are registered, an imaging unit that captures the person to be recognized, and an image captured by the imaging unit. The image input means for inputting the sign, the sign detection means for detecting the sign indicated by the person to be recognized from the image input by the image input means, the sign detected by the sign detection means, and the signature registered in the registration means Sign collating means for collating, face image detecting means for detecting a face image from the image input by the image input means, face image detected by the face image detecting means, and the sign registered in the registration means Face collation means for collating the face image of the registrant whose signature is matched by the collation means, and the face image of the person to be recognized and the sign collation by the face collation means Depending on whether or not the registered person's face image signature matches the stage match, the recognized person having a recognizing means for recognizing whether the registrant signs match with the signature verification unit.

  The person recognition device according to the present invention includes a registration unit in which a registrant's face image and a sign indicated by the registrant's own operation are registered, an imaging unit that captures the person to be recognized, and an image captured by the imaging unit. An image input means for inputting a face image, a face image detection means for detecting a face image from an image input by the image input means, a face image detected by the face image detection means, and a face image registered in the registration means If the face image of the person to be recognized does not match the face image registered in the registration means by the face matching means, the recognition from the image input by the image input means A signature detection means for detecting a signature indicated by the subject person, and a signature verification means for verifying the signature detected by the signature detection means and the signature of each registrant registered in the registration means And when the face image of the person to be recognized matches the face image of the registrant registered in the registration unit by the face verification unit, or the signature detected by the signature detection unit by the signature verification unit Recognizing means for recognizing that the person to be recognized is a registrant whose face image or signature matches when the signature of the registrant registered in the registration means is matched.

  The person recognition device according to the present invention has a variety of operation modes, and has a registration means in which a sign for mode change indicated by a person's operation for changing the operation mode to a specific operation mode is registered, and a person is photographed Photographing means, image input means for inputting an image of the person photographed by the photographing means, sign detection means for detecting a sign indicated by the person from the image of the person inputted by the image input means, and the sign detection means A signature collating means for collating the signature detected by the above-mentioned sign for mode change registered in the registering means, and the signature detected by the signature detecting means by the sign collating means and the mode registered in the registering means When the change sign matches, the operation mode is changed to the specific operation mode corresponding to the mode change sign. And a processing means that.

  The person recognition apparatus according to the present invention has a variety of operation modes, and a registration means in which a sign indicated by a person's action is registered as an operation sign corresponding to various operations effective in a specific operation mode A photographing means for photographing a person in the specific operation mode, an image input means for inputting an image of the person photographed by the photographing means, and a sign indicated by the person from the image of the person inputted by the image input means. A signature detection means for detecting, a signature verification means for verifying a signature detected by the signature detection means and an operation signature registered in the registration means, and a signature detected by the signature detection means by the signature verification means And the operation sign registered in the registration means coincide with the operation corresponding to the operation sign. And a processing means for performing the operations Flip was.

  The traffic control device according to the present invention includes a registration unit in which a registrant's face image and a sign indicated by a registrant's unique operation are registered, an imaging unit that captures a person to be recognized, and an image captured by the imaging unit. The image input means for inputting the sign, the sign detection means for detecting the signature indicated by the person to be recognized from the image input by the image input means, the sign detected by the sign detection means, and the signature registered in the registration means Sign collating means for collating, face image detecting means for detecting a face image from the image input by the image input means, face image detected by the face image detecting means, and the sign registered in the registration means Face collation means for collating the face image of the registrant whose signature is matched by the collation means, and the face image of the person to be recognized and the sign collation by the face collation means Recognizing means for recognizing whether the person to be recognized is a registrant whose signature is matched by the sign collating means, based on whether or not the face image of the registrant whose signature is matched by the step is matched. And a passage control means for permitting passage of the recognition target person when the means recognizes that the recognition target person is a registered person registered in the registration means.

  The traffic control device according to the present invention includes a registration unit in which a registrant's face image and a sign indicated by a registrant's unique operation are registered, an imaging unit that captures a person to be recognized, and an image captured by the imaging unit. An image input means for inputting a face image, a face image detection means for detecting a face image from an image input by the image input means, a face image detected by the face image detection means, and a face image registered in the registration means If the face image of the person to be recognized does not match the face image registered in the registration means by the face matching means, the recognition from the image input by the image input means A signature detection means for detecting a signature indicated by the subject person, and a signature verification means for verifying the signature detected by the signature detection means and the signature of each registrant registered in the registration means And when the face image of the person to be recognized matches the face image of the registrant registered in the registration unit by the face verification unit, or the signature detected by the signature detection unit by the signature verification unit A recognizing means for recognizing that the person to be recognized is a registrant whose face image or sign coincides with the sign of the registrant registered in the registration means; A traffic control unit that permits the recognition target person to pass when the user is recognized as a registered person registered in the registration unit;

  According to the present invention, it is possible to provide a person recognition device and a traffic control device capable of easily inputting and operating information other than non-contact biological information without impairing convenience.

  The best mode for carrying out the present invention will be described with reference to the drawings.

  FIG. 1 schematically shows configurations of a person recognition device 1 and a traffic control device 2 according to the present embodiment. The person recognition device 1 includes a user interface unit 10 and a processing unit 20. Further, the traffic control device 2 is realized by connecting the door opening / closing mechanism 200 to the processing unit 20 of the person recognition device 1.

  The user interface unit 10 includes a camera 110, a display unit 111, an auxiliary illumination 112, and the like. The camera 110 is constituted by a television camera using an image sensor such as a CCD sensor, and takes a moving image or a continuous still image. The camera 110 functions as a photographing unit that photographs a face image of the recognition target person P (an image including at least a face), an operation indicated by the recognition target person P, and the like.

  The display unit 111 is a display device that displays operation guidance for the recognition target person P, images taken by the camera 110, and the like. For example, when displaying guidance that optimizes the distance of the recognition target person P with respect to the camera 110 and the position in the left-right direction, the display unit 111 displays an optimal facial image together with the image being captured by the camera 110. The position is indicated by a circle or the like.

  The auxiliary illumination unit 112 is a lighting device such as a fluorescent lamp that emits light as auxiliary illumination to the person to be recognized P photographed by the camera 110 when the brightness of the place photographed by the camera 110 is not sufficient. It is. The auxiliary illumination unit 112 is, for example, auxiliary illumination that irradiates light toward at least the face of the recognition target person P from the upper right or upper left of the camera 110, and toward at least the face of the recognition target person P from the lower side of the camera 110. And auxiliary illumination that emits light. Note that the auxiliary illumination unit 112 may be omitted as long as the illuminance in the area photographed by the camera 110 is sufficient.

Next, the processing unit 20 will be described.
The processing unit 20 can be realized by a configuration of a computer or the like, for example. That is, the processing unit 20 functions as a processing unit when a control unit including a CPU (not shown) executes various control programs recorded in advance on a storage medium such as a hard disk device or an optical disk. The operation is realized. Hereinafter, a specific configuration example of the processing unit 20 will be described in detail with reference to FIG.

  The processing unit 20 includes an image input unit 121 as an image input unit, a face detection unit 122 as a face image detection unit, a feature extraction unit 123 as a feature amount extraction unit, a registration information holding unit 124 as a registration unit, and a recognition unit. A recognition unit 125 as a face verification unit or a signature verification unit, an automatic authentication start determination unit 126 as an authentication start determination unit, a traffic control unit 127 as a traffic control unit, and a sign detection unit 128 as a signature detection unit. doing.

  The image input unit 121 includes an interface for transmitting and receiving image data. For example, the image input unit 121 performs A / D conversion on analog image data as a captured image obtained from the camera 110 and digitizes the data, and then performs subsequent components (a face detection unit 122, an automatic authentication start determination unit 126, or a signature). It is supplied to the detection unit 128).

  The face detection unit 122 detects the face image of the person P to be recognized from the image captured by the camera 110 supplied from the image input unit 121. The face detection unit 122 detects the face area having the highest correlation value as a face area by, for example, obtaining a correlation value while moving the template prepared in advance and the image. The image of the face area is set as a face image. In addition, as the face detection unit 122, a face area detection method such as a face area extraction method using an eigenspace method or a subspace method may be used.

  The feature extraction unit 123 uses the face image detected by the face detection unit 122 to extract facial feature amounts such as grayscale information or partial space information. For example, the feature extraction unit 123 detects the positions of eyes and noses as facial feature amounts from the face image detected by the face detection unit 122. Further, the feature extraction unit 123 cuts out a face region into a certain size and shape based on the positions of eyes, nose, and the like as face feature amounts detected from the face image, and uses the density information as a feature amount. Extract. For example, the feature amount is extracted as feature vector using m × n-dimensional information, using the gray value of an area of m pixels × n pixels as information as it is.

  The registration information holding unit 124 is configured by a storage device such as a hard disk device, and recognition data relating to a registrant is registered in advance. For example, personal information such as the name of the registrant and recognition data for recognizing the registrant are registered in the registration information holding unit 124 in processing in the administrator mode described later. The recognition data includes, for example, a registrant's face image, a registrant's face feature amount extracted by the feature extraction unit 123, an ID sign as identification information for identifying the registrant, and the registration For example, a password sign as the password information of the person. The ID sign and the password sign will be described in detail later.

  The recognizing unit 125 performs processing for recognizing (authenticating) the recognition target person P photographed by the camera 110. As a process for recognizing the person P to be recognized from the face image, the recognizing unit 125 uses the feature amount of the registrant registered as recognition data in the registration information holding unit 124 and the feature amount obtained by the feature extracting unit 123. The collation process by the face image which collates (comparison) is performed. In addition, the recognizing unit 125 is configured to verify a signature that is identified by the person to be recognized P photographed by the camera 110 and an ID sign or a password sign registered as recognition data in the registration information holding unit 124. It also has a function of performing verification processing.

  Further, as a collation process using a face image, the recognition unit 125 recognizes who is a registered person registered in the registration information holding unit 124 as a recognition target person P photographed by the camera 110. A function of performing recognition processing in a processing mode called 1-to-N collation (one-to-N collation mode), and a specific registrant registered in the registration information holding unit 124 by the person to be recognized P photographed by the camera 110 It has a function of performing authentication processing in a processing mode (one-to-one matching mode) called one-to-one matching for recognizing whether or not they match (identify).

  For example, in the recognition process in the 1-to-N matching mode, the recognition unit 125 includes the face image of the person P to be recognized photographed by the camera 110 and the face image of each registrant registered in the registration information holding unit 124. When the maximum similarity is calculated by calculating the similarity as the matching degree with the maximum similarity, the recognition target person P corresponds to the recognition data corresponding to the maximum similarity. It is judged that it is a person.

  In the recognition process in the one-to-one verification mode, the recognition unit 125 first selects a registrant to be verified among the registrants registered in the registration information holding unit 124 with an ID sign. A similarity is calculated as a matching degree between the identified registrant's face image and the recognition target person's P face image captured by the camera 110, and the recognition is performed when the similarity is equal to or greater than a predetermined threshold. It is determined that the target person P is a registered person.

  The automatic authentication start determination unit 126 detects the presence or absence of a person based on the image taken by the camera 110, and determines the start of recognition (authentication) processing based on the detection result. For example, the automatic authentication start determination unit 126 is continuously supplied with images taken by the camera 110 via the image input unit 121, and based on the difference information of the continuous images, It is determined whether or not a person exists in the shooting range (captured image).

  Moreover, the said passage control part 127 controls the passage of the said recognition subject person P by controlling the door opening / closing mechanism 200 based on the recognition result with respect to the recognition subject person P by the said recognition part 125. FIG. The door opening / closing mechanism 200 is a mechanism that is provided at a door for entering and exiting a facility and controls opening / closing of the door or unlocking / locking of the door.

  For example, when the recognition target person P is a registrant registered in the registration information holding part 124 in the recognition part 125, the traffic control part 127 receives the recognition target person P and the registrant from the recognition part 125. And the door opening / closing mechanism 200 opens the door. In addition, when only a part of registrants (persons having authority to pass through the door) can pass through the door, the traffic control unit 127 uses the recognition unit 125 to recognize the person P to be recognized as a door. When it is recognized that the person has the authority to pass, the door is opened by the door opening / closing mechanism.

  The sign detection unit 128 detects the sign (ID sign or password sign) indicated by the person P to be recognized from the continuous images (video) taken by the camera 110. The sign detection unit 128 acquires continuous images captured by the camera 110 via the image input unit 121, and detects the movement of the recognition target person P from the difference image of the continuous images. It is.

  Next, the signature detected by the signature detection unit 128 will be described.

  The signature used as the ID sign and the password sign is indicated by a person's action or the like. For example, the sign used as the ID sign and the password sign may be a blink pattern, a finger (shape or movement) pattern, a hand (shape or movement) pattern, a facial expression pattern, or a combination thereof. And so on. The person recognition apparatus 1 is characterized in that a specific pattern is not given to the registrant as an ID sign or a password sign, but a pattern as a sign determined by the registrant is registered as authentication data in advance.

  For example, in the pattern by blinking, the registrant himself / herself determines the order of blinking the left and right eyes, the interval between those blinks, and the like to form a sign. In addition, as for the pattern by the finger, the registrant himself / herself may form a sign by determining the shape indicated by the finger, the change pattern of the movement, the change interval, and the like. In addition, as for the pattern based on facial expressions, the registrant himself / herself determines the change pattern of facial expressions, the interval at which the facial expressions are changed, and the like to form a sign.

  Thus, the signature used as the ID sign and the password sign in the person recognition apparatus 1 is specified by the pattern of physical movement, and is determined by the registrant himself. Therefore, the physical movement pattern used as a signature can be set independently by each registrant, and the degree of freedom of the physical movement pattern used as a signature can be increased.

  Further, in the person recognition apparatus 1, an operation mode change operation performed by the administrator, an operation by the administrator in the registration mode, and the like can be performed using a signature registered in the registration information holding unit 124 in advance. Suppose that In this case, for example, the registration information holding unit 124 replaces an operation input for changing the person recognition apparatus 1 from a normal operation mode (for example, a collation mode) to a specific operation mode (for example, a registration mode). A sign for mode change or an operation sign corresponding to various operations in a specific operation mode (for example, various operations of the administrator in the registration mode) is registered in advance. Note that the sign for mode change and the sign for operation are also indicated by the action of a person in the same manner as the sign for ID and the sign for password.

Next, basic face image registration processing and authentication processing in the person recognition apparatus 1 configured as described above will be described.
2 and 3 are flowcharts for explaining the operation when registering an ID sign and a password sign in the registration information holding unit 124. FIG. FIG. 2 is a flowchart for explaining an operation example when shifting from the verification mode as the normal operation mode to the operation mode for registration (registration mode) as the specific operation mode. FIG. 3 is a flowchart for explaining an example of operation using a signature for operation in the registration mode and a signature registration process.

  First, an operation example when shifting from the verification mode to the registration mode will be described with reference to FIG.

  It is assumed that the person recognition device 1 normally operates in an operation mode (collation mode) for performing face collation. During the operation in the collation mode, in the person recognition device 1, whether or not a person has been detected because the automatic authentication start determination unit 126 determines the start of authentication processing based on the image taken by the camera 110. Is determined (step S1). When the automatic authentication start determination unit 126 determines that a person has been detected, the image input unit 121 captures an image captured by the camera 110 and supplies the image to the sign detection unit 128 (step S2). ).

  Then, the sign detection unit 128 detects a sign indicated by the person photographed from the image supplied from the image input unit 121. When the signature is detected by this, the recognition unit 125 determines whether or not the signature detected by the signature detection unit 128 is a predetermined signature for shifting to the registration mode (step S3). For example, when an administrator signature is set as a signature for shifting to the registration mode, it is determined whether or not the signature detected by the signature detection unit 128 is an administrator signature.

  If it is determined that the sign detected by the sign detection unit 128 is not a predetermined sign for shifting to the registration mode (NO in step S3), the recognition unit 125 determines the person's character based on normal face matching. Recognition processing is executed (step S4). The person recognition process based on the face matching will be described later in detail.

  Further, when it is determined by the above determination that the sign is a predetermined sign for shifting to the registration mode (step S3, YES), the recognizing unit 125 performs processing for shifting the operation mode of the person recognition device 1 from the verification mode to the registration mode. (Step S5). As a result, the operation mode of the person recognition apparatus 1 shifts to the registration mode, and the ID sign, the password sign, or other information can be registered in the registration information holding unit 124.

  As described above, the operation mode has been changed from the normal mode to the registration mode for registering various types of information using pre-registered operation signs. It is possible to perform the operation of shifting. Further, as a result, it is possible to perform operation even for a handicapped person who is hygienic without touching the apparatus and who cannot perform manual operation.

Next, an operation example using an operation sign in the registration mode and an operation example of an ID sign or password sign registration process will be described with reference to FIG.
When the registration mode is entered, a menu screen for the registration mode is displayed on the display unit 111 (step S11). In this state, the person recognition device 1 is in a state where the menu can be operated by an operation sign indicated by the administrator. The operation sign is registered in advance in the registration information holding unit 124 corresponding to various operations, and is indicated by the operation of the administrator. Therefore, the manager operates the menu by showing an operation sign corresponding to a desired operation in front of the camera 110.

  That is, in the state where the registration mode menu screen is displayed, the image input unit 121 captures the video imaged by the camera 110 and supplies the video image to the sign detection unit 128. On the other hand, the sign detection unit 128 detects the sign indicated by the administrator from the video supplied from the image input unit 121 (step S12). Thus, when the operation sign indicated by the administrator is detected, the recognizing unit 125 collates the sign with the operation sign registered in the registered information holding unit 124, so that the administrator The operation content corresponding to the indicated sign (operation sign) is determined (steps S13, S15, S21, and S23).

  Here, when the recognition unit 125 determines that the sign indicated by the administrator is an operation sign corresponding to an operation for inputting information on the registrant (step S13, YES), the display unit 111 displays the registration sign. A guide to input information related to the person (information for identifying the person to be registered and personal information of the person to be registered) is displayed.

  In this state, the administrator inputs information for specifying the registered person or personal information such as a name as information on the registered person through an operation unit (not shown). Information regarding the registered person such as information for identifying the registered person input by the administrator and personal information of the registered person is registered in the registration information holding unit 124 (step S14). When the registration of information relating to the registrant is completed, a menu screen for the registration mode is displayed again on the display unit 111 (step S11), and a sign for operation by the administrator can be input.

  In addition, in the state where the registration mode menu screen is displayed on the display unit 111, the signature indicated by the administrator detected by the signature detection unit 128 by the recognition unit 125 is the ID sign or password for the registrant. When it is determined that the signature is an operation signature corresponding to an operation for shifting to signature registration as a signature (step S15, YES), an ID signature or a password signature registered by the registered person is displayed on the display unit 111. Guidance is provided as shown in front of the camera 110.

  In response to this guidance, the person to be registered indicates a sign to be registered as an ID sign or a password sign in front of the camera 110. As a result, the image input unit 121 captures and captures an image of the registered person showing the sign with the camera 110 (step S16).

  The image captured from the camera 110 is supplied to the sign detection unit 128. The sign detection unit 128 detects a sign from the image supplied from the image input unit 121, and the recognition unit 125 determines whether the sign is valid (step S17). For example, the recognition unit 125 determines whether or not the signature is valid by determining whether the signature detected by the signature detection unit 128 does not match the signature already registered in the registration information holding unit 124. Determine whether.

  If it is determined by this determination that the signature detected by the signature detection unit 128 is not valid (step S17, NO), the display unit 111 displays a guidance indicating a different signature, and the step S16. Then, the signature detection process by the signature detection unit 128 is executed again.

  If it is determined by the above determination that the signature detected by the signature detection unit 128 is valid (YES in step S17), the sign detection unit 128 causes the registered person to confirm the detected signature. Then, an image showing the detected sign is displayed on the display unit 111 (step S18). If the signature displayed on the display unit 111 is satisfactory, the registered person or administrator instructs the registration of the signature (step S19, YES). If the signature displayed on the display unit 111 is not satisfactory, the registered person or administrator instructs the registration of the signature (step S19, NO), and the processing from step S16 is executed again. To do.

  When the signature detected by the signature detection unit 128 is confirmed (step S19, YES), the signature detection unit 128 registers the detected signature in the registration information holding unit 124 corresponding to information on the registered person. Is performed (step S20).

  In this signature registration process, the detected signature is recorded in the registration information holding unit 124 as the ID signature or password signature of the registered person. When the signature detected by the signature detection unit 128 is registered, a menu screen for the registration mode is displayed again on the display unit 111 (step S11), and a sign for operation by the administrator can be input. It becomes.

  In the state where the registration mode menu screen is displayed on the display unit 111, the signature indicated by the administrator detected by the signature detection unit 128 by the recognition unit 125 shifts to registration of the registrant's face image. If it is determined that the sign is an operation sign (step S21, YES), the person recognition apparatus 1 registers the face image and the face feature amount as authentication data of the registered person in the registration information holding unit 124. The process which performs is performed (step S22).

  In the registration processing of the face image and the facial feature amount, the facial feature amount is extracted by the face detection unit 122 and the feature extraction unit 123 from the face image of the registered person photographed by the camera 110, and the facial image is extracted. And the feature amount are recorded in the registration information holding unit 124 corresponding to the registered person. When the registration of the face image is finished, the menu screen for the registration mode is displayed again on the display unit 111 (step S11), and the operation sign by the administrator can be input.

  In addition, in the state where the registration mode menu screen is displayed on the display unit 111, the sign indicated by the administrator detected by the sign detection unit 128 by the recognition unit 125 is an operation sign indicating the end of the registration mode. When the person recognition apparatus 1 determines that it is (step S23, YES), the person recognition apparatus 1 performs a process of shifting from the registration mode to the normal mode (step S24), and is capable of performing a collation process using a face image as a normal operation. And

  As described above, in the registration mode for registering information related to the registered person, ID sign, password sign, and face image as registered person authentication data, various operations can be performed according to pre-registered operation signs. Therefore, it is possible to easily register information on a registered person, register an ID sign, a password sign, and a face image without directly touching the apparatus. Further, as a result, it is possible to perform operation even for a handicapped person who is hygienic without touching the apparatus and who cannot perform manual operation.

  Next, a person recognition process using a face image and a signature (ID sign and password sign) in the person recognition apparatus will be described.

  First, an operation example of the person recognition apparatus 1 in the one-to-one matching mode will be described. FIG. 4 is a flowchart for explaining an operation example of the person recognition apparatus 1 in the one-to-one matching mode.

  As shown in FIG. 4, in the one-to-one verification mode, first, the recognition target person P shows an ID sign toward the camera 110. On the other hand, an image captured by the camera 110 is supplied to the automatic authentication start determination unit 126 via the image input unit 121. Therefore, when a person is photographed by the camera 110 (step S31, YES), the automatic authentication start determination unit 126 determines the start of authentication.

  When the automatic authentication start determination unit 126 determines the start of authentication, the image input unit 121 supplies an image captured by the camera 110 to the sign detection unit 128 (step S32). As a result, the video of the ID sign that the recognition target person P shows toward the camera 110 is supplied to the sign detection unit 128.

  The sign detection unit 128 performs a process of detecting the ID sign indicated by the recognition target person P from the video captured from the camera 110 via the image input unit 121 (step S33). In the sign detection process, when the ID sign cannot be detected from the video captured by the sign detection unit 128 from the camera 110 (NO in step S33), an ID sign input guidance is displayed on the display unit 111. Thus, the recognition target person P is again guided to show the ID sign (step S34), and the process returns to step S32.

  In the sign detection process, when the sign detection unit 128 detects a sign from the video captured from the camera 110 (YES in step S33), the recognition unit 125 detects the sign detected by the sign detection unit 128 and the above-described sign. Signature verification processing is performed with the ID signature registered in the registration information holding unit 124 (step S35).

When the signature detected by the signature detection unit 128 does not match the ID signature registered in the registration information holding unit 124 in this signature verification process (step S36,
NO), by displaying the ID sign input guidance on the display unit 111, the recognition target person P is again guided to show the ID sign (step S34), and the process returns to step S32.

  Further, when the signature detected by the signature detection unit 128 in the signature verification process matches one of the ID signatures registered in the registration information holding unit 124 (step S36, YES), the person recognition device 1 Performs a face matching process between the face image of the recognition target person P and the face image of the registrant whose ID sign matches. That is, when the signature detected by the signature detection unit 128 matches the ID sign of the registrant (step S36, YES), the face detection unit 122 supplies the camera supplied via the image input unit 121. The face image of the person to be recognized is detected from the image photographed by 110 (step S37). When the face image of the person to be recognized is detected in this way, the feature extraction unit 123 extracts a facial feature amount from the face image (step S38) and supplies it to the recognition unit 125.

  The recognizing unit 125 includes the facial feature amount of the person to be recognized extracted by the feature extracting unit 123 and the facial feature amount of the registrant whose ID sign registered in the registered information holding unit 124 matches. Is performed (step S39).

  When the feature amount of the face of the person to be recognized matches the feature amount of the face of the registrant specified by the ID sign by this face matching process, that is, when the similarity between the two is equal to or greater than a predetermined similarity ( In step S40, YES), the recognizing unit 125 recognizes that the person to be recognized is a registrant specified by the ID sign, displays the recognition result on the display unit 111, and the traffic control unit. 127 is notified. Accordingly, if the registrant is a person who has the authority to pass, the passage control unit 127 opens the door by the door opening / closing mechanism 200 and permits the recognition target person to pass (step S41).

  Further, when the facial feature amount of the person to be recognized does not match the facial feature amount of the registrant specified by the ID sign by the face matching process, that is, the similarity between the two is less than the predetermined similarity (Step S40, NO), the display unit 111 displays password sign input guidance (Step S41). In response to the input guidance, the person to be recognized shows a password sign toward the camera 110.

  In this state, the image input unit 121 supplies an image captured by the camera 110 to the sign detection unit 128. Accordingly, the password sign image shown by the recognition target person P toward the camera 110 is supplied to the sign detection unit 128 (step S43).

  The sign detection unit 128 performs a process of detecting the password sign indicated by the recognition target person P from the video captured from the camera 110 via the image input unit 121 (step S44). In the sign detection process, when the password sign cannot be detected from the video captured by the sign detection unit 128 from the camera 110 (NO in step S44), the process returns to step S42 and the password sign is displayed on the display unit 111. By displaying the input guidance, the recognition target person P is again guided to show the password sign as an input state of the password sign.

  In the sign detection process, when the sign detection unit 128 detects a sign from the video captured from the camera 110 (step S44, YES), the recognition unit 125 detects the sign detected by the sign detection unit 128 and the above-described sign. Signature verification processing is performed with the registrant password signature specified by the ID signature registered in the registration information holding unit 124 (step S45).

  If the signature detected by the signature detection unit 128 and the password sign of the registrant registered in the registration information holding unit 124 do not match in the signature verification process (NO in step S46), the recognition unit 125 recognizes that the person P to be recognized is not a registrant specified by the ID sign, displays on the display unit 111 that the authentication is impossible and the door cannot be passed, and displays the authentication process. Exit. In this case, the passage control unit 127 does not permit the recognition target person P to pass by keeping the door closed by the door opening / closing mechanism 200 (step S47).

  Also, when the signature detected by the signature detection unit 128 in the signature verification process matches the registrant password signature specified by the ID signature registered in the registration information holding unit 124 (step S46, YES), the recognizing unit 125 recognizes that the person to be recognized P is a registrant specified by the ID sign, and that the display unit 111 is authenticated and allows passage of the door. And notifies the passage control unit 127 that the passage of the recognition subject person's door is permitted. Accordingly, the passage control unit 127 opens the door (or unlocks the door) by the door opening / closing mechanism 200, and permits the recognition target person P to pass (step S41).

  As described above, a plurality of registrants each register their own ID sign together with the face image, and in the authentication process in the one-to-one matching mode, the ID sign registered by the person to be recognized in the registration means The signature of the person to be authenticated is detected from the image shown to the camera and photographed by the camera, and the detected signature and the ID signature of each registrant registered in advance are collated. The registrant to be verified is specified.

  In addition, the registrant registers in advance a unique password sign together with the face image in the registration means, and in the authentication process in the one-to-one verification mode, the face image of the person to be authenticated is registered in the registration means. If it does not match the face image, the person to be recognized shows the password sign to the camera, detects the sign indicated by the person to be authenticated from the image taken by the camera, and the detected sign and the registration means By verifying with the password sign of the registered registrant, it is authenticated whether the recognition target person is a registrant registered in the registration means.

  According to the person recognition device or traffic control device that realizes these operations, the person to be recognized can easily input an ID sign or password sign without touching the device directly. It is hygienic without the subject touching the device, and even a handicapped person who is unable to operate by hand can easily perform the operation. In addition, since it can be operated completely in contact, it is possible to improve the durability of the device, and it is not necessary to install an operation unit such as a numeric keypad in the user interface unit. Can be reduced, and the mounting position can be saved.

  Next, an operation example of the person recognition device 1 in the 1-to-N collation mode will be described. FIG. 5 is a flowchart for explaining an operation example of the person recognition apparatus 1 in the 1-to-N collation mode.

  As shown in FIG. 4, in the one-to-N collation mode, first, the recognition target person P captures a face image with the camera 110. On the other hand, an image captured by the camera 110 is supplied to the automatic authentication start determination unit 126 via the image input unit 121. The automatic authentication start determination unit 126 detects whether or not a person has been photographed by the camera 110 based on an image photographed by the camera 110 supplied via the image input unit 121 (step S51). ). Thus, when it is detected that a person is photographed by the camera 110 (step S31, YES), the automatic authentication start determination unit 126 determines the start of authentication.

  When the automatic authentication start determination unit 126 determines the start of authentication, the image input unit 121 supplies an image captured by the camera 110 to the face detection unit 122 (step S52). The face detection unit 122 detects the face image of the person to be recognized from the image captured by the camera 110 supplied via the image input unit 121 (step S53). Thus, when the face image of the person to be recognized is detected, the feature extraction unit 123 extracts a facial feature amount from the face image (step S54), and supplies it to the recognition unit 125.

  The recognizing unit 125 performs face collation to collate the facial feature amount of the recognition target person extracted by the feature extracting unit 123 with the facial feature amounts of all registrants registered in the registered information holding unit 124. Processing is performed (step S55). In this face matching process, the degree of similarity between the facial feature amount of the person to be recognized and the facial feature amount of each registrant registered in the registration information holding unit 124 is calculated, and the maximum similarity degree is predetermined. If the maximum similarity is equal to or greater than a predetermined threshold, it is determined that the recognition target person matches the registrant who gives the maximum similarity.

  When it is determined by the face matching process that the recognition target person matches the registrant registered in the registration information holding unit 124 (step S56, YES), the recognition unit 125 determines that the recognition target person is the maximum similarity. It recognizes that it is a registrant who gives a degree, displays the recognition result on the display part 111, and notifies the said traffic control part 127. FIG. Accordingly, if the registrant is a person who has the authority to pass, the passage control unit 127 opens the door (or unlocks the door) by the door opening / closing mechanism 200, and passes the recognition target person. Is permitted (step S57).

  If it is determined by the face matching process that the person to be recognized does not match the registrant registered in the registration information holding unit 124 (NO in step S56), a password sign is input to the display unit 111. A guidance is displayed (step S58). In response to the input guidance, the person to be recognized shows a password sign toward the camera 110.

  In this state, the image input unit 121 supplies an image captured by the camera 110 to the sign detection unit 128. As a result, the password sign image shown by the recognition target person P toward the camera 110 is supplied to the sign detection unit 128 (step S59).

  The sign detection unit 128 performs a process of detecting the password sign indicated by the recognition target person P from the video captured from the camera 110 via the image input unit 121 (step S60). In the sign detection process, if the sign detection unit 128 cannot detect the password sign from the video captured from the camera 110 (NO in step S60), the process returns to step S58, and the password sign is displayed on the display unit 111. By displaying the input guidance, the recognition target person P is again guided to show the password sign as an input state of the password sign.

  In the sign detection process, when the sign detection unit 128 detects a sign from the video captured from the camera 110 (YES in step S60), the recognition unit 125 detects the sign detected by the sign detection unit 128 and the above-described sign. Signature verification processing is performed with the password signatures of all registrants registered in the registration information holding unit 124 (step S61).

  When the signature detected by the signature detection unit 128 does not match the password sign of each registrant registered in the registration information holding unit 124 in this signature verification process (NO in step S62), the recognition unit 125 recognizes that the recognition target person P is not a registrant registered in the registration information holding unit 124, and displays on the display unit 111 a message indicating that the authentication is impossible and that the door cannot be passed. The authentication process ends. In this case, the passage control unit 127 does not permit the recognition target person P to pass by keeping the door closed by the door opening / closing mechanism 200 (step S63).

  If the signature detected by the signature detection unit 128 in the signature verification process matches the password sign of any registrant registered in the registration information holding unit 124 (YES in step S62), The recognizing unit 125 recognizes that the person P to be recognized is a registrant whose password sign is the same, and displays on the display unit 111 information indicating that authentication is possible and permitting passage of the door, The passage control unit 127 is notified that the recognition subject person is allowed to pass through the door. Accordingly, the passage control unit 127 opens the door (or unlocks the door) by the door opening / closing mechanism 200, and permits the recognition target person P to pass (step S57).

  As described above, each registrant registers in advance a unique password sign together with the face image in the registration means, and the face image of the person to be authenticated is registered in the registration means in the authentication process in the 1: N matching mode. If the face image of any registrant does not match, the person to be recognized shows the password sign toward the camera, and the sign shown by the person to be authenticated is detected from the image taken by the camera. By verifying whether the person to be recognized is a registrant registered in the registration means by checking the signature against the password sign of each registrant registered in advance in the registration means It is.

  As a result, the person to be recognized can easily input a password sign without directly touching the device. As a result, the person to be recognized is hygienic without touching the device, and further, it can be operated by hand. Even a disabled person can be easily operated. In addition, since the operation can be performed completely in contact with the device, it is possible to improve the durability of the device, and further, it is not necessary to install an operation unit such as a numeric keypad in the user interface unit. Can be reduced, and the mounting position can be saved.

The figure which shows schematic structure of the person recognition apparatus and traffic control apparatus which concern on embodiment of this invention The flowchart for demonstrating the operation example at the time of transfering to the registration mode which registers a signature etc. from a normal operation mode. It is a flowchart for demonstrating the operation example of the registration process of the authentication data including a signature in registration mode. The flowchart for demonstrating the operation | movement procedure of the collation process by one-to-one collation. The flowchart for demonstrating the operation | movement procedure of the collation process by one-to-many collation.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Person recognition apparatus, 2 ... Traffic control apparatus, 10 ... User interface part, 20 ... Processing part, 110 ... Camera, 111 ... Display part, 121 ... Image input part, 122 ... Face detection part, 123 ... Feature extraction part, 124: Registration information holding unit, 125: Recognizing unit, 127: Traffic control unit, 128: Sign detecting unit, 200: Door opening / closing mechanism

Claims (6)

A registration means in which the face image of the registrant and a sign indicated by the registrant's own action are registered;
Photographing means for photographing the person to be recognized;
Image input means for inputting an image photographed by the photographing means;
Sign detection means for detecting a sign indicated by the person to be recognized from the image input by the image input means;
A signature verification means for verifying the signature detected by the signature detection means and the signature registered in the registration means;
Face image detection means for detecting a face image from the image input by the image input means;
A face collation unit for collating the face image detected by the face image detection unit with the face image of the registrant whose signature is matched by the sign collation unit registered in the registration unit;
Depending on whether or not the face image of the person to be recognized matches the face image of the registrant whose signature is matched by the sign collating means by the face collating means, the signature of the person to be recognized is matched by the sign collating means. A recognition means for recognizing whether or not a registrant,
A person recognition apparatus characterized by comprising: A registration means in which the face image of the registrant and a sign indicated by the registrant's own action are registered;
Photographing means for photographing the person to be recognized;
Image input means for inputting an image photographed by the photographing means;
Face image detection means for detecting a face image from the image input by the image input means;
Face collation means for collating the face image detected by the face image detection means with the face image registered in the registration means;
When the face collation means does not match the face image of the person to be recognized and the face image registered in the registration means, a sign indicated by the person to be recognized is detected from the image input by the image input means. Sign detection means;
A signature verification means for verifying the signature detected by the signature detection means and the signature of each registrant registered in the registration means;
When the face image of the person to be recognized matches the face image of the registrant registered in the registration unit by the face verification unit, or the signature detected by the sign detection unit by the signature verification unit and the registration Recognizing means for recognizing that the person to be recognized is a registrant whose face image or signature is matched when the sign of the registrant registered in the means matches.
A person recognition apparatus characterized by comprising: In a person recognition device having various operation modes,
A registration means in which a sign for mode change indicated by the action of a person who changes the operation mode to a specific operation mode is registered;
Photographing means for photographing a person;
Image input means for inputting an image of a person photographed by the photographing means;
Sign detecting means for detecting a sign indicated by the person from the image of the person input by the image input means;
A signature verification means for verifying the signature detected by the signature detection means and the mode change signature registered in the registration means;
When the sign detected by the sign detection means matches the sign for mode change registered in the registration means by the sign collating means, the specific operation mode corresponding to the sign for mode change corresponds to the operation mode. Processing means to change to,
A person recognition apparatus characterized by comprising: In a person recognition device having various operation modes,
Registration means in which a sign indicated by a person's action is registered as an operation sign corresponding to various operations that are valid in a specific operation mode;
Photographing means for photographing a person in the specific operation mode;
Image input means for inputting an image of a person photographed by the photographing means;
Sign detection means for detecting a sign indicated by the person from the image of the person input by the image input means;
A signature verification means for verifying the signature detected by the signature detection means and the operation signature registered in the registration means;
Processing means for executing an operation corresponding to the operation corresponding to the operation signature when the signature detected by the signature detection means matches the signature for operation registered in the registration means When,
A person recognition apparatus characterized by comprising: A registration means in which the face image of the registrant and a sign indicated by the registrant's own action are registered;
Photographing means for photographing the person to be recognized;
Image input means for inputting an image photographed by the photographing means;
Sign detection means for detecting a sign indicated by the person to be recognized from the image input by the image input means;
A signature collating means for collating the signature detected by the signature detecting means with the signature registered in the registration means;
Face image detection means for detecting a face image from the image input by the image input means;
A face collation unit for collating the face image detected by the face image detection unit with the face image of the registrant whose signature is matched by the sign collation unit registered in the registration unit;
Depending on whether the face image of the person to be recognized matches the face image of the registrant whose signature is matched by the sign collating means by the face collating means, the sign of the person to be recognized is matched by the sign collating means. A recognition means for recognizing whether or not a registrant,
When the recognition means recognizes that the person to be recognized is a registrant registered in the registration means, a passage control means for permitting passage of the person to be recognized;
A traffic control device comprising: A registration means in which the face image of the registrant and a sign indicated by the registrant's own action are registered;
Photographing means for photographing the person to be recognized;
Image input means for inputting an image photographed by the photographing means;
Face image detection means for detecting a face image from the image input by the image input means;
Face collation means for collating the face image detected by the face image detection means with the face image registered in the registration means;
When the face collation means does not match the face image of the person to be recognized and the face image registered in the registration means, a sign indicated by the person to be recognized is detected from the image input by the image input means. Sign detection means;
A signature verification means for verifying the signature detected by the signature detection means and the signature of each registrant registered in the registration means;
When the face image of the person to be recognized matches the face image of the registrant registered in the registration unit by the face verification unit, or the signature detected by the sign detection unit by the signature verification unit and the registration Recognizing means for recognizing that the person to be recognized is a registrant whose face image or signature coincides with the signature of the registrant registered in the means;
When the recognition means recognizes that the person to be recognized is a registrant registered in the registration means, a passage control means for permitting passage of the person to be recognized;
A traffic control device comprising:

Images