JP2006293842A - Retrieval system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a system which has high scalability and can be efficiently operated and reduces network load and is improved in security. <P>SOLUTION: The system includes an input terminal and a server, and feature data of faces of a plurality of persons to be retrieved are previously registered in a database of the server, and the input terminal generates feature data of a face from a face picture, and instead of face picture data, feature data of the face extracted from face picture data are transmitted from the input terminal to the server, and the server performs collation processing based on feature data of the face received from the input terminal. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  In the present invention, for example, data including generation information such as a person's face is acquired, generation information such as a face image is detected from the acquired data, and a person whose generation information matches the generation information such as the detected face image is detected. The present invention relates to a search system for searching from a database.

  In recent years, an apparatus and a system for authenticating or searching for a person using biological information such as a face image have been proposed. For example, in a face search system that searches for a person using a face image, the face included in the input image is checked by comparing the face image included in the input image with a face image registered in the database in advance. Search for people.

  For example, Japanese Unexamined Patent Application Publication No. 2004-258664 describes a system that performs face image exchange between a plurality of devices via a LAN and performs 1: N face matching. In this system, first, a face image is extracted from an image photographed by a photographing / collating machine, and the face image is transmitted to an operator terminal via a LAN. In the operator terminal, the operator designates a face image to be collated, and returns the designated face image to the photographing / collating machine. The face photographing / collation machine extracts a facial feature quantity from the face image returned from the terminal, and performs a 1: N matching process with the extracted facial feature quantity and N face feature quantities registered in advance. Send the result to the display via LAN. That is, the system has a substantially stand-alone configuration in which image acquisition, face detection, feature amount extraction, and face matching are performed in one apparatus.

  However, the above-described conventional techniques have the following problems (1) to (3).

(1) Since the face photographing / collating machine is substantially a stand-alone configuration, the scale of the apparatus increases as the number of registered DB face information increases.

(2) Since the size of the face photographing / collating machine becomes large, the flexibility of installation is not good, and the degree of freedom of installation of the device for photographing a face image is lost.

(3) Even when it is desired to use face information stored in one database by a plurality of devices, the number of devices that can capture a face image is limited.

  In Japanese Patent Laid-Open No. 2004-318409, a face image captured by a terminal device, a face image stored in an ID card, and personal information (name, etc.) are transferred to a server connected via a LAN. Describes what performs face matching processing. In this system, the server identifies a person based on personal information transferred from the terminal device via the LAN, and the registered face image of the person is compared with the face image transferred from the terminal device via the LAN by 1: 1. And returns the collation result to the terminal via the LAN, and the terminal displays the collation result from the server. In the system, image data (for example, a bitmap file) that can be reproduced as a face image, personal information, a collation result, and the like are transferred on the LAN.

  However, the above-described conventional techniques have the following problems (1) to (4).

(1) Since the amount of face image data for one person is several tens of kilobytes or more, the DB capacity increases as the number of registrants who register face information in the database increases.

(2) When the number of terminals increases or when face search requests from one terminal increase, the amount of information transferred on the network increases.

(3) In the server-side face search algorithm, it is necessary to convert data from a face image to a face feature amount, the number of registered DB face information increases, the number of terminals increases, or face search from one terminal The more requests, the longer the search processing time.

(4) When image data that can be reproduced as an image flows on the network, there is a risk of information leakage due to eavesdropping.
JP 2004-258774 A JP 2004-318409 A

  An object of the present invention is to solve the above-described problems, and to provide a search system that can perform efficient operation and has high expandability or security.

  A search system according to an aspect of the present invention is a system having a terminal device and a server device, wherein the terminal device inputs data including feature values, and data acquired by the data input unit. Feature data generating means for generating feature data from the data, and data transmitting means for transmitting the feature data generated by the feature data generating means to the server device. The server device stores specific feature data in advance. Storing means, data receiving means for receiving feature data transmitted from the terminal device, and feature data similar to the feature data received by the data receiving means by a predetermined reference value or more are stored in the storage means. Search means for searching from the feature data.

  A search system according to an aspect of the present invention is a system including a terminal device and a server device, and the terminal device acquires data including human biometric information, and the data input unit acquires the data. Generated information extraction means for extracting biological information from the generated data, feature data generation means for generating feature data indicating features of the biological information from the biological information extracted by the generated information extraction means, and generated by the feature data generation means Data transmitting means for transmitting the feature data to the server device, the server device receiving the feature data transmitted from the terminal device and storage means storing the feature data of the biometric information of the registrant Data receiving means, and feature data similar to the feature data received by the data receiving means by a predetermined criterion or more. And a retrieval means for retrieving from the feature data of the registrant biometric information stored in the unit.

  An object of the present invention is to solve the above-described problems, and to provide a search system that can perform efficient operation and has high expandability or security.

The best mode for carrying out the present invention will be described with reference to the drawings.
First, a first embodiment of the present invention will be described.
FIG. 1 schematically shows a configuration example of a face search system according to the first embodiment. As shown in FIG. 1, the face search system includes an image input device 1, an input terminal 2, a server 3, an output terminal 4, and the like. The image input device 1 is connected to the input terminal 2. The input terminal 2, the server 3, and the output terminal 4 are connected by a network such as a LAN. Note that the input terminal 2 and the output terminal 4 may be configured by the same terminal device.

First, the image input apparatus 1 will be described.
The image input apparatus 1 is an apparatus for inputting a moving image or a still image, and includes, for example, a video camera or an image scanner. The image input apparatus 1 functions as an image input unit that inputs a human face image (an image including at least a face). Further, when the image input device 1 is installed as a video camera for photographing a human face image, the image input device 1 includes a display unit that displays an image being taken, a guide, and the like, and brightness is unpredictable. It may be configured to be incorporated in a user interface device having auxiliary lighting for illuminating a person to be photographed or the like in the environment where the user is shooting.

Next, the input terminal 2 will be described.
The input terminal 2 can be realized by a configuration of a computer or the like, for example. That is, the input terminal 2 implements various processes as described later by executing various control programs recorded in advance on a storage medium such as a hard disk device or an optical disk by a control unit including a CPU (not shown). .

Hereinafter, a specific configuration example of the input terminal 2 will be described in detail with reference to FIG.
As shown in FIG. 1, the input terminal 2 includes an image acquisition unit 21 as an image acquisition unit, a face detection unit 22 as a face image detection unit, a feature extraction unit 23 as a feature amount extraction unit, and feature data transmission. Part 24.

  The image acquisition unit 21 includes an interface for receiving image data from the image input apparatus 1. That is, the image acquisition unit 21 includes an interface corresponding to the configuration of the image input device 1. For example, when the image input device 1 is a camera that captures an analog image, the image acquisition unit 21 performs A / D conversion on analog image data as a captured image acquired from the camera as the image input device 1. Digitized and supplied to the downstream face detection unit 22.

  The face detection unit 22 detects a human face region from the image supplied from the image acquisition unit 21 and extracts it as a human face image. The face detection unit 22 detects, for example, a location having the highest correlation value as a face region by obtaining a correlation value while moving the template prepared in advance and the image, and detects the face region image as a face region. Extract as an image. In addition, the face detection unit 22 may use a face area detection method such as a face area extraction method using an eigenspace method or a subspace method as a method of extracting a face image.

  The feature extraction unit 23 detects a feature amount of the face from the face image extracted by the face detection unit 22 and generates feature data of the face image. The feature extraction unit 23 detects facial feature amounts such as grayscale information or partial space information in the face image extracted by the face detection unit 22 to generate facial feature data. For example, the feature extraction unit 23 detects the positions of eyes and noses as facial feature amounts from the face image detected by the face detection unit 22. Further, the feature extraction unit 23 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 feature amounts. To detect.

  Further, the feature extraction unit 23 extracts m × n-dimensional information as a feature vector using a gray value in an area of m pixels × n pixels, and generates the feature vector as face feature data as a feature amount. To do. The feature data indicating the feature amount of the face has a smaller data amount than the image data acquired by the image acquisition unit 21 from the image input device 1 or the face image data detected by the face detection unit 22. For example, the face image data detected by the face detection unit 22 is several tens of kilobytes, whereas the facial feature data extracted by the feature extraction unit 23 is several kilobytes.

  The feature data transmission unit 24 transmits the facial feature data generated by the feature extraction unit 23 to the server 3. The feature data transmission unit 24 is configured by a network interface, for example. Further, the feature data transmitting unit 24 may be provided with an encryption function for encrypting data such as facial feature data transmitted to the server 3. That is, it is possible to improve the security of data transmitted by the feature data transmission unit 24 to the server 3 by encrypting and transmitting transmission data such as facial feature data so that the server 3 can decrypt the data. it can.

Next, the server 3 will be described.
The server 3 can be realized by a configuration of a computer or the like, for example. That is, the server 3 realizes various processes as described later by executing various control programs recorded in advance on a storage medium such as a hard disk device or an optical disk by a control unit including a CPU (not shown).

Hereinafter, a specific configuration example of the server 3 will be described in detail with reference to FIG.
As shown in FIG. 1, the server 3 includes a feature data receiving unit 31 as a feature data receiving unit, a face feature database 32 as a storage unit, a face searching unit 33 as a face searching unit, and a search result transmission as an output unit. Part 34 is provided.

  The feature data receiving unit 31 receives facial feature data transmitted from the input terminal 2. The feature data receiving unit 31 is configured by a network interface, for example. Further, the feature data receiving unit 31 may be provided with a decryption function corresponding to the encryption function of the feature data transmitting unit 24. That is, when the feature data transmission unit 24 of the input terminal 2 encrypts data to be transmitted to the server 3, the feature data reception unit 31 has a function of decrypting the received data.

  The face feature database 32 is configured by a storage device such as a hard disk device. In the face feature database 32, face feature data in the registrant's face image is registered in advance in association with the registrant's personal information, etc. Also, in the face feature database 32, for example, the registrant's face is registered. You may make it register an image.

  The face search unit 33 performs a collation process between the face feature data received by the feature data receiving unit 31 and the face feature data registered in the face feature database 32. In the face search unit 33, the facial feature data most similar to the facial feature data received by the feature data receiving unit 31 among the facial feature data registered in the facial feature database 32 is sent to the registrant. Output as search results. That is, the registrant of the feature data having the highest similarity to the facial feature data received by the feature data receiving unit 31 is output as a person search result. Further, when there are a plurality of feature data having a predetermined similarity or higher, all of those registrants may be output as search results.

  In the first embodiment, the face image detected by the image input device 1 is registered in the face feature database 32 as face image matching processing in the face search unit 33. Recognition processing is performed in a processing mode called 1-to-N collation (one-to-N collation mode) for searching (recognizing) who the person is. In the recognition process in the one-to-N matching mode, the face search unit 33 includes the feature data supplied from the input terminal 2 and the face feature data of each registrant registered in the face feature database 32. When the maximum matching degree is extracted from these matching degrees and the maximum matching degree is equal to or greater than a predetermined threshold, the feature data supplied from the input terminal 2 is the maximum similarity degree. It is determined that the data belongs to the registrant corresponding to the feature data.

  The search result transmission unit 34 transmits data indicating the search result of the face by the face search unit 33 to the output terminal 4. The search result transmission unit 34 is configured by a network interface, for example. The hardware that implements the feature data receiving unit 31 and the search result transmitting unit 34 may be configured by a single network interface. Further, the search result transmission unit 34 may be provided with an encryption function for encrypting data such as search results transmitted to the output terminal 4. That is, by encrypting the transmission data so that the output terminal 4 can decrypt it, the security of the data transmitted by the search result transmission unit 34 to the output terminal 4 can be improved.

Next, the output terminal 4 will be described.
The output terminal 4 can be realized by a configuration of a computer or the like, for example. That is, the output terminal 4 implements various processes as described later by executing various control programs recorded in advance on a storage medium such as a hard disk device or an optical disk by a control unit including a CPU (not shown). . The hardware for realizing the input terminal 2 and the output terminal 4 may be configured by a single device such as a computer.

Hereinafter, a specific configuration example of the output terminal 4 will be described in detail with reference to FIG.
As shown in FIG. 1, the output terminal 4 has a search result receiving unit 41 as a search result receiving unit and a display unit 42 as a display unit.
The search result receiving unit 41 receives data indicating a face search result based on face feature data transmitted from the server 3. The search result receiving unit 41 is configured by a network interface, for example. Further, the search result receiving unit 41 may be provided with a decryption function corresponding to the encryption function of the search result transmitting unit 34. That is, when the search result transmission unit 34 of the server 3 encrypts the data to be transmitted to the output terminal 4, the search result reception unit 41 has a function of decrypting the received data. The display unit 42 displays the face search result received by the search result receiving unit 41.

  Note that various mechanisms and functions may be added to the face search system described above depending on the operation mode. For example, when the face search system is applied as an entry / exit management system, a traffic control unit such as a door opening / closing mechanism is added to the server 3 or the output terminal 4, and the face search result by the face search unit 33 of the server 3 is added. Accordingly, the passage control unit can control the passage of a person who has input a face image to the image input device 1. Thereby, the entrance / exit management system by a face image is realizable.

  When the face search system is applied as a person monitoring system, face feature data of a person to be monitored is registered in the face feature database 32, and the face search result by the face search unit 33 of the server 3 is registered. If the person who inputs the face image to the image input device 1 is a person to be monitored, the display unit 42 of the output terminal 4 can display that the person to be monitored has been found. Thereby, the person monitoring system by a face image is realizable.

Next, the flow of processing in the face search system according to the first embodiment configured as described above will be described.
FIG. 2 is a flowchart for explaining the flow of processing in the face search system according to the first embodiment.
First, in the image input apparatus 1, an image including a human face is input. For example, when the image input device 1 is configured as a video camera, the image input device 1 captures an image including a human face. When an image including a human face is input to the image input device 1, the input terminal 2 acquires the image input to the image input device 1 by the image acquisition unit 21 (step S11).

  The image acquired by the image acquisition unit 21 of the input terminal 2 is supplied to the face detection unit 22. The face detection unit 22 detects a face region from the image acquired by the image acquisition unit 21, and extracts the detected face region as a face image (step S12). The face image extracted by the face detection unit 22 is supplied to the feature extraction unit 23. The feature extraction unit 23 detects feature quantities such as facial feature points from the face image extracted by the face detection unit 22 and generates facial feature data based on the detected feature quantities such as facial feature points. (Step S13).

  The facial feature data generated by the feature extraction unit 23 is supplied to the feature data transmission unit 24. The feature data transmission unit 24 transmits the facial feature data generated by the feature extraction unit 23 to the server 3 (step S14). At this time, when data to be transmitted to the server 3 is encrypted, the feature data transmission unit 24 encrypts the facial feature data generated by the feature extraction unit 23 and transmits the encrypted data to the server 3 via the network. .

  When facial feature data is transmitted from the input terminal 2, the server 3 receives the feature data from the input terminal 2 by the feature data receiving unit 31 (step S21). The feature data received by the feature data receiving unit 31 is supplied to the face search unit 33. When the feature data from the input terminal 2 is encrypted, the feature data receiving unit 31 decrypts the received feature data and supplies it to the face search unit 33.

  The face search unit 33 performs a search process (matching process) between the feature data received by the feature data receiving unit 31 and all the feature data registered in the face feature database 32 (step S22). The search process (collation process) is a process of searching the face feature database 32 for facial feature data that matches the facial feature data received by the feature data receiving unit 31.

  For example, the face search unit 33 calculates the matching degree between the feature data received by the feature data receiving unit 31 and all feature data registered in the face feature database 32. If the maximum matching degree among the matching degrees calculated in this way is equal to or greater than a predetermined threshold, the face search unit 33 matches the feature data received by the feature data receiving unit 31 with the feature data giving the maximum matching degree. Judge. In this case, the face search unit 33 outputs information indicating the registrant of the feature data having the maximum matching degree as a search result to the search result transmission unit 34.

  If the maximum matching degree among the calculated matching degrees is less than a predetermined threshold, the face search unit 33 causes the feature data matching the feature data received by the feature data receiving unit 31 to be stored in the face feature database 32. Judge that it is not registered. In this case, the face search unit 33 outputs information indicating that there is no corresponding registrant as a search result to the search result transmission unit 34.

  A search result obtained by the search processing of the face search unit 33 is supplied from the face search unit 33 to the search result transmission unit 34. The search result transmission unit 34 transmits the search result obtained by the face search unit 33 to the output terminal 4 (step S23). At this time, when the data to be transmitted to the output terminal 4 is encrypted, the search result transmission unit 34 encrypts the search result obtained by the face search unit 33 and transmits it to the output terminal 4 via the network. .

  When the search result is transmitted from the server 3, the output terminal 4 receives the search result from the server 3 by the search result receiving unit 41 (step S31). The search result received by the search result receiving unit 41 is supplied to the display unit 42. When the data indicating the search result from the server 3 is encrypted, the search result receiving unit 41 decrypts the received data and supplies it to the display unit 42. The display unit 42 supplied with the search result received by the search result receiving unit 41 displays the search result supplied from the search result receiving unit 41 (step S32).

  As described above, in the face search system according to the first embodiment, the input terminal generates face feature data from the face image, and the server performs matching processing using the face feature data transmitted from the input terminal. I am doing so. Thereby, even if the amount of information in the face feature database as a search target increases, the installation conditions of the input terminal can be relaxed without affecting the scale of the input terminal. Further, since it is easy to connect a plurality of input terminals to one server, the expandability of the system configuration is enhanced and the degree of freedom in system design is increased.

  Furthermore, in the face search system according to the first embodiment, the face feature data extracted from the face image data is transmitted instead of the face image data as information to be sent from the input terminal to the server. In the database, feature data of faces of a plurality of persons to be searched are registered.

  As a result, the processing time required for the matching process for matching the face feature data of the person transmitted from the input terminal with the face feature data stored in the database can be reduced to the matching process when the face image is used. It is possible to shorten the time required. This is because facial feature data (for example, several kilobytes) is smaller and easier to handle than facial image data (for example, several tens of kilobytes), and processing time for extracting feature values from each facial image stored in the database is eliminated. This is because it can. For this reason, the amount of information flowing through the network can be reduced, and the load on the network can be reduced.

  Furthermore, in the face search system according to the first embodiment, the face feature data generated at the input terminal is data that cannot be restored to a face image. That is, in the input terminal, an algorithm for generating facial feature data that cannot be restored to a facial image is used as an algorithm for extracting facial feature data from the facial image. Thereby, even if information transmitted from the input terminal to the server is illegally acquired on the network by eavesdropping or the like, it is possible to reduce the risk of misuse of the information.

Next, a second embodiment will be described.
FIG. 3 schematically shows a configuration example of a face search system according to the second embodiment. As shown in FIG. 3, the face search system according to the second embodiment includes an image input device 1, an input terminal 2, a server 3, an output terminal 4, a personal information input device 5, and the like. In addition to the configuration of the face search system according to the first embodiment shown in FIG. 1, a personal information input device 5 is provided. In the face search system according to the second embodiment shown in FIG. 3, the same components as those of the face search system according to the first embodiment shown in FIG. The explanation will be omitted.

  That is, in the face search system shown in FIG. 3, the image input apparatus 1 is an apparatus for inputting an image including a human face, and has the same configuration as the image input apparatus 1 shown in FIG. In the face search system shown in FIG. 3, the output terminal 4 is a terminal for displaying the face search result by the server 3, and has the same configuration as the output terminal 4 shown in FIG.

  The personal information input device 5 is a device for inputting personal information of a person who inputs an image including a face by the image input device 1. The personal information input device 5 may be any device that inputs personal information or a part of personal information, and can be realized by various devices. For example, the personal information input device 5 includes a keyboard through which personal information is input by a key, a touch panel through which personal information is input by touching the key displayed on the screen, and an ID card storing personal information. A card reader or the like that reads information is applicable.

  In the face search system shown in FIG. 3, the input terminal 2 includes an image acquisition unit 21 as an image acquisition unit, a face detection unit 22 as a face image detection unit, a feature extraction unit 23 as a feature amount extraction unit, and a feature. It has a data and personal information transmission unit 24 and a personal information acquisition unit 25 as personal information acquisition means.

  The image acquisition unit 21 is an interface for acquiring image data from the image input device 1, and has the same function as the image acquisition unit 21 shown in FIG. The face detection unit 22 extracts a human face image from the image acquired by the image acquisition unit 21, and has the same function as the face detection unit 22 shown in FIG. The feature extraction unit 23 generates feature data as a facial feature quantity from the face image extracted by the face detection unit 22, and has the same function as the feature extraction unit 23 shown in FIG. Yes.

  The feature data and personal information transmission unit 24 transmits the facial feature data generated by the feature extraction unit 23 and the personal information acquired by the personal information acquisition unit 25 to the server 3. The feature data / personal information transmission unit 24 is configured by a network interface in the same manner as the feature data transmission unit 24 shown in FIG. Further, the feature data / personal information transmission unit 24 may be provided with an encryption function for encrypting data such as face feature data and personal information transmitted to the server 3.

  In the face search system shown in FIG. 3, the server 3 includes a feature data / personal information receiving unit 31 as feature data receiving means or personal information receiving means, a face feature database 32 as storage means, and face search means. A face search unit 33 and a search result transmission unit 34 as output means are provided.

  The feature data / personal information receiving unit 31 receives facial feature data and personal information transmitted from the input terminal 2, and is similar to the feature data receiving unit 31 shown in FIG. Composed. Further, the feature data / personal information receiving unit 31 may be provided with a decryption function corresponding to the encryption function of the feature data / personal information transmission unit 24.

  The face feature database 32 is configured by a storage device such as a hard disk device, similar to the face feature database 32 shown in FIG. In the configuration shown in FIG. 3, the facial feature database 32 includes personal information for identifying a registrant in the facial image of the registrant associated with the personal information input by the personal information input device 5. Facial feature data and the like are registered.

  The face search unit 33 performs a collation process between the face feature data received by the feature data / personal information receiving unit 31 and the face feature data registered in the face feature database 32. FIG. The face search unit 33 shown in FIG. Further, the face search unit 33 shown in FIG. 3 includes a condition extraction unit 33a for narrowing down the face feature data registered in the face feature database 32 based on the personal information received by the feature data / personal information receiving unit 31. Have. The condition extracting unit 33a narrows down the feature data to be searched from the facial feature data registered in the facial feature database 32 based on the personal information received by the feature data / personal information receiving unit 31.

  That is, the face search unit 33 narrows down the feature data to be searched based on the personal information received from the input terminal 2 by the condition extraction unit 33a, and the search target feature data narrowed down by the condition extraction unit 33a. And the feature data received from the input terminal 2 are collated. As a result, the face search unit 33 stores the feature data of the face most similar to the feature data of the face received by the feature data receiving unit 31 among the feature data of the search target narrowed down by the condition extraction unit 33a. Is output as a face search result.

  The condition extraction unit 33a may narrow down a plurality of feature data from the face feature database 32 as a search target based on personal information, or may search for one feature data from the face feature database 32 based on personal information. You may make it narrow down as. In the former case, the face search unit 33 performs a 1: n (number of feature data narrowed down by personal information) matching process, and in the latter case, a 1: 1 matching process. In the second embodiment, it is assumed that 1: n (the number of feature data narrowed down by personal information) is collated.

  Further, the search result transmission unit 34 transmits data indicating the face search result by the face search unit 33 to the output terminal 4 and, like the search result transmission unit 34 shown in FIG. Consists of. Further, the search result transmission unit 34 may be provided with an encryption function for encrypting data such as search results transmitted to the output terminal 4. The feature data / personal information receiving unit 31 and the search result transmitting unit 34 may be configured by a single network interface.

Next, a processing flow in the face search system according to the second embodiment configured as described above will be described.
FIG. 4 is a flowchart for explaining the flow of processing in the face search system according to the second embodiment. The process shown in FIG. 4 is obtained by adding the process of narrowing down the feature data to be searched based on the personal information input to the input terminal 2 to the process shown in FIG. For this reason, in FIG. 4, the same steps as those in FIG. 2 are denoted by the same reference numerals, and detailed description thereof is omitted.

  First, in the face search system according to the second embodiment, an image including a person's face is input by the image input device 1, and personal information of the person is input by the personal information input device 5. For example, when the image input device 1 is configured as a video camera and the personal information input device 5 is configured with a keyboard, the video camera as the image input device 1 captures an image including a human face, and Personal information is input by the person using the keyboard as the personal information input device 5.

  When an image including a person's face is input to the image input device 1, the image acquisition unit 21 acquires an image input to the image input device 1 in the input terminal 2 (step S11). When the image acquisition unit 21 acquires an image from the image input device 1, the face detection unit 22 extracts a face image from the acquired image (step S12). When the face image is extracted by the face detection unit 22, the feature extraction unit 23 generates face feature data as a feature quantity such as a feature point of the face from the face image (step S13). The facial feature data generated by the feature extraction unit 23 is supplied to the feature data / personal information transmission unit 24.

  When personal information is input to the personal information input device 5, the input terminal 2 acquires the personal information input to the personal information input device 5 by the personal information acquisition unit 25 (step S16). The personal information acquired by the personal information acquisition unit 25 is supplied to the feature data / personal information transmission unit 24.

  The feature data / personal information transmission unit 24 associates the facial feature data generated by the feature extraction unit 23 with the personal information acquired by the personal information acquisition unit 25 and transmits it to the server 3 ( Step S17). At this time, when the data to be transmitted to the server 3 is encrypted, the feature data / personal information transmission unit 24 encrypts the facial feature data and the personal information generated by the feature extraction unit 23 to create a network. To the server 3.

  When the feature data and personal information are transmitted from the input terminal 2, the server 3 receives the feature data and personal information from the input terminal 2 by the feature data / personal information receiving unit 31 (step S25). ). The feature data and personal information received by the feature data / personal information receiving unit 31 are supplied to the face search unit 33. When the feature data and personal information from the input terminal 2 are encrypted, the feature data / personal information receiving unit 31 decrypts the received feature data and personal information and supplies them to the face search unit 33. .

  In the face search unit 33, the condition extraction unit 33 a extracts conditions for narrowing down the feature data registered in the face feature database 32 according to the personal information received by the feature data / personal information receiving unit 31. Based on this condition, the feature data as the search target is narrowed down from the feature data registered in the face feature database 32 (step S26).

  When the search target feature data is narrowed down by the condition extraction unit 33a, the face search unit 33 causes the feature data received by the feature data / personal information reception unit 31 and the search target feature data narrowed down by the condition extraction unit 33a. The search process (collation process) is performed (step S27). The search process (collation process) is a process for searching for facial feature data matching the facial feature data received by the feature data receiving unit 31 from search targets narrowed down by the condition extracting unit 33a.

  For example, the face search unit 33 calculates the matching degree between the feature data received by the feature data receiving unit 31 and all the feature data narrowed down by the condition extracting unit 33a. If the maximum matching degree among the matching degrees calculated in this way is equal to or greater than a predetermined threshold, the face search unit 33 matches the feature data received by the feature data receiving unit 31 with the feature data giving the maximum matching degree. Judge. In this case, the face search unit 33 outputs information indicating the registrant of the feature data having the maximum matching degree as a search result.

  If the maximum matching degree among the calculated matching degrees is less than a predetermined threshold, the face search unit 33 causes the condition extraction unit 33a to send feature data that matches the feature data received by the feature data receiving unit 31. It is determined that the refined feature data (feature data corresponding to the input personal information) does not exist. In this case, the face search unit 33 outputs information indicating that there is no corresponding registrant as a search result.

  The search result by the above search processing is supplied from the face search unit 33 to the search result transmission unit 34, and is transmitted to the output terminal 4 by the search result transmission unit 34 (step S28). In the output terminal 4 to which the search result is transmitted from the server 3, the search result receiving unit 41 receives the search result from the server 3 (step S31), and the received search result is displayed on the display unit 42 ( Step S32).

  In the face search system according to the second embodiment, in addition to the face search system described in the first embodiment, the input terminal acquires personal information input to the personal information input device, and the acquired personal information The information is transferred to the server together with the face feature data, and the face search process is performed on the feature data narrowed down by the personal information by the server.

  Thereby, according to the face search system according to the second embodiment, in addition to the effects described in the first embodiment, it is possible to reduce the time required for the face search process in the server. Even if the amount of information stored in the face feature database increases, an increase in the time required for the face search process can be suppressed, and the accuracy of the search result can be improved.

  The face search system described as the first and second embodiments is configured to send the face search result at the server for the face image acquired at the input terminal to the output terminal. In addition, an input terminal for inputting an image including a face of a person to be collated, a server for performing face search processing, and an output terminal for displaying a search result are separate devices installed at remote locations. Can be realized. Therefore, the input terminal, the server, and the output terminal can be installed at remote locations, and it is possible to remotely monitor the face search status of a person. As a result, it is possible to check the search result with an output terminal installed at a location other than the installation location other than the input terminal and the server.

  In the first and second embodiments, a person search system (face search system) using a face image as biometric information has been described. However, the search system for a person using other biometric information can also be used. The same applies. Furthermore, the first and second embodiments can be applied to a system that performs a search process by comparing feature data obtained from input data with feature data stored in advance in a database. It is.

The figure which shows schematic structure of the face search system which concerns on the 1st Embodiment of this invention. The flowchart for demonstrating the flow of the process in the face search system as 1st Embodiment. The figure which shows schematic structure of the face search system which concerns on the 2nd Embodiment of this invention. The flowchart for demonstrating the flow of a process in the face search system as 2nd Embodiment.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Image input device, 2 ... Input terminal, 3 ... Server, 4 ... Output terminal, 5 ... Personal information input device, 21 ... Image acquisition part, 22 ... Face detection part, 23 ... Feature extraction part, 24 ... Feature data transmission Part (feature data / personal information transmission part), 25 ... personal information acquisition part, 31 ... feature data reception part (feature data / personal information reception part), 32 ... face feature database, 33 ... face search part, 33a ... condition extraction 34: Search result transmission unit 41: Search result reception unit 42: Display unit

Claims (5)

In a search system having a terminal device and a server device,
The terminal device is
Data input means for inputting data including feature values;
Feature data generating means for generating feature data from the data acquired by the data input means;
Data transmission means for transmitting the characteristic data generated by the characteristic data generation means to the server device,
The server device
Storage means for storing specific feature data in advance;
Data receiving means for receiving feature data transmitted from the terminal device;
Search means for searching for feature data similar to the feature data received by the data receiving means by a predetermined reference value or more from the feature data stored in the storage means,
A search system characterized by that. In a search system having a terminal device and a server device,
The terminal device is
Data input means for acquiring data including biological information of a person;
Generation information extraction means for extracting biological information from data acquired by the data input means;
Feature data generating means for generating feature data indicating the characteristics of the biological information from the biological information extracted by the generated information extracting means;
Data transmission means for transmitting the characteristic data generated by the characteristic data generation means to the server device,
The server device
Storage means for storing biometric information feature data of the registrant;
Data receiving means for receiving feature data transmitted from the terminal device;
Search means for searching for feature data similar to the feature data received by the data receiving means over a predetermined reference from the feature data of the biometric information of the registrant stored in the storage means;
A search system characterized by that. In a search system having a terminal device and a server device,
The terminal device is
Data input means for acquiring data including biological information of a person;
Generation information extraction means for extracting biological information from data acquired by the data input means;
Feature data generating means for generating feature data indicating the characteristics of the biological information from the biological information extracted by the generated information extracting means;
Personal information input means for acquiring personal information of the person of the biometric information included in the data input by the data input means;
Data transmitting means for transmitting the characteristic data generated by the characteristic data generating means and the personal information acquired by the personal information input means to a server device;
The server device
Storage means for storing the biometric information feature data of the registrant and personal information in association with each other;
Data receiving means for receiving feature data and personal information transmitted from the terminal device;
Narrowing means for narrowing down feature data to be searched from feature data of biometric information of a registrant stored in the storage means based on personal information received by the data receiving means;
Search means for searching from the feature data narrowed down by the narrowing means for feature data similar to the feature data received from the terminal device by the data receiving means over a predetermined reference,
A search system characterized by that. The biometric information is a face image of a person, and the feature data of the biometric information is feature data indicating a feature of each individual face.
The search system according to claim 2 or 3, wherein the search system is characterized. Furthermore, it has the 2nd terminal device which receives the search result by the search means of the said server apparatus, and displays the received search result,
The search system according to claim 2 or 3, wherein the search system is characterized.

Images