JP6905653B2 - Face recognition system, face recognition server and face recognition method - Google Patents

Description

The present invention relates to a face recognition system, a face recognition server, and a face recognition method for recognizing a person's face using an image captured by a camera device.

Conventionally, when determining the customer base of a person who appears in the imaging area set in the store, it is determined from the behavioral form of the person whether or not the customer is the target of analysis, and the person who does not correspond is excluded from the analysis target. Therefore, there is known a customer base analysis system that accurately analyzes the customer base of customers who visit a store (see, for example, Patent Document 1).

In this customer demographic analysis system, for example, the imaging area is set so as to image the customer heading for the audience seat from the guidance waiting area near the entrance / exit of the store from the front. By detecting a person who behaves differently from the behavior of moving toward the audience and excluding that person from the analysis target, a person other than the customer such as a clerk is included in the analysis target, and the same customer is duplicated and analyzed. Avoid being included in.

Japanese Unexamined Patent Publication No. 2014-232495

However, in the configuration of Patent Document 1, when a person does not take a predetermined behavioral form, the person can be excluded from the analysis target, but in order to determine whether or not the person takes a predetermined behavioral form. , A complicated judgment process was required.

In addition, even if it can be determined that the person is taking a predetermined action form, if the same person makes a round trip to the entrance / exit of the store twice or more for the purpose, it is considered that the totals are duplicated. Further, in the configuration of Patent Document 1, it is considered that if a person does not take a predetermined behavioral form, even if the person is a customer to be analyzed, the person is excluded from the analysis target.

The present invention has been devised in view of the above-mentioned conventional situations, and is a face recognition system, a face recognition server, and a face recognition server that can efficiently perform statistical processing so that the same person is not included in duplicate without making a complicated judgment. The purpose is to provide a face recognition method.

The present invention is a face recognition system in which a camera device and a face recognition server are connected, and the face recognition server extracts a feature amount of a person's face appearing in video data captured by the camera device. A face recognition system that compares the extracted facial features with the facial features stored for a predetermined period of time, and discards the extracted facial features if the similarity is less than the predetermined value. be.

Further, the present invention is a face recognition system in which a camera device and a face recognition server are connected, and the face recognition server extracts a feature amount of a person's face appearing in video data captured by the camera device. , The extracted facial feature amount and the facial feature amount stored for the first predetermined period are compared, and the degree of similarity between them is less than the predetermined value, and the face stored for the first predetermined period. This is a face recognition system that discards the extracted facial features when the imaging time of the extracted facial features is stored before the second predetermined period from the shooting time of the extracted facial features. ..

Further, the present invention is a face recognition system in which a camera device and a face recognition server are connected, and the face recognition server extracts a feature amount of a person's face appearing in video data captured by the camera device. , The extracted facial feature amount is compared with the facial feature amount stored for a predetermined period, and when the similarity between them is equal to or more than a predetermined value, the extracted facial feature amount is stored. It is a system.

Further, the present invention is a face recognition server to which a camera device is connected, extracts a facial feature amount of a person appearing in video data captured by the camera device, and determines the extracted facial feature amount. It is a face recognition server that compares the facial features stored for a period of time and discards the extracted facial features if the similarity is less than a predetermined value.

Further, the present invention is a face recognition server to which a camera device is connected, extracts a feature amount of a person's face appearing in video data captured by the camera device, and obtains the extracted face feature amount and the first. The degree of similarity with the facial feature amount stored for a predetermined period of 1 is less than a predetermined value, and the imaging time of the facial feature amount stored for the first predetermined period is the extracted facial feature amount. This is a face recognition server that discards the extracted facial features when the data is stored from the shooting time of the face to before the second predetermined period.

The present disclosure provides a face recognition server camera device is connected, the extracts Kaotoku symptoms of a person appearing in the video data taken by the camera device, the extracted Kaotoku symptoms amount and pre-fixed compares the similarity between the time stored Kaotoku symptom amount, if the similarity is less than a predetermined value, and stores the facial feature quantity the extracted and treated as a statistical object, the similarity When it is equal to or more than a predetermined value, the imaging time when the extracted face of the person is imaged is compared with the imaging time when the face whose similarity is determined to be equal to or greater than the predetermined value is imaged. If the imaging time at which the face determined to be equal to or greater than a predetermined value is captured is within a predetermined time range from the imaging time of the video data from which the facial feature amount has been extracted, the face is excluded from the target of the statistics and described as described above. When the imaging time at which the face whose similarity is determined to be equal to or higher than the predetermined value is not within the predetermined time range from the imaging time of the video data from which the facial feature amount is extracted, the statistics A face recognition server including a processor for statistical processing as a target is provided.

Further, the present invention is a face recognition method in a face recognition system in which a camera device and a face recognition server are connected, and is a process for extracting a feature amount of a person's face appearing in video data captured by the camera device. , The process of comparing the extracted facial feature amount and the facial feature amount stored for a predetermined period, and discarding the extracted facial feature amount when the similarity between them is less than a predetermined value. Is a face recognition method.

Further, the present invention is a face recognition method in a face recognition system in which a camera device and a face recognition server are connected, and is a process of extracting a feature amount of a person's face appearing in video data captured by the camera device. , The extracted facial feature amount and the facial feature amount stored for the first predetermined period are compared, and the degree of similarity between them is less than the predetermined value, and the face stored for the first predetermined period. When the imaging time of the feature amount of the extracted face is stored before the second predetermined period from the shooting time of the feature amount of the extracted face, the process of discarding the feature amount of the extracted face is executed. , Face recognition method.

The present disclosure provides a face recognition method in the face recognition system and the camera device and the face recognition server is connected, and extracts the Kaotoku symptoms of a person appearing in the video data taken by the camera device, wherein compared to pre and extracted Kaotoku symptoms amount the similarity between a certain period stored Kaotoku symptom amount, if the similarity is less than a predetermined value, and stores the facial feature quantity the extracted, When the similarity is processed as a target of statistics and the similarity is equal to or higher than a predetermined value, the imaging time at which the extracted face of the person is imaged and the imaging time at which the face judged to have the similarity is equal to or higher than the predetermined value are imaged. The imaging time when the face whose similarity is determined to be equal to or higher than the predetermined value is within a predetermined time range from the imaging time of the video data from which the facial feature amount is extracted. In this case, the imaging time at which the face whose similarity is determined to be equal to or higher than the predetermined value is excluded from the subject of the statistics is the predetermined from the imaging time of the video data from which the facial feature amount is extracted. Provided is a face recognition method that statistically processes the subject of the statistics when it is not within the time range.

According to the present invention, statistical processing can be performed so that the same person is not included in duplicate without making a complicated judgment.

A block diagram showing in detail an example of the internal configuration of the face recognition system of the present embodiment. A flowchart illustrating an example of an operation procedure of face detection processing in the face recognition server of the present embodiment. Explanatory drawing showing an example of the UI screen displayed on the display of the user terminal A block diagram showing in detail an example of the internal configuration of the face recognition system of the modified example of this embodiment.

Hereinafter, an embodiment (hereinafter, referred to as “the present embodiment”) in which the face recognition system, the face recognition server, and the face recognition method are specifically disclosed in the present invention will be described in detail with reference to the drawings as appropriate. However, more detailed explanation than necessary may be omitted. For example, detailed explanations of already well-known matters and duplicate explanations for substantially the same configuration may be omitted. This is to avoid unnecessary redundancy of the following description and to facilitate the understanding of those skilled in the art. It should be noted that the accompanying drawings and the following description are provided for those skilled in the art to fully understand the present disclosure, and are not intended to limit the subject matter described in the claims.

FIG. 1 is a block diagram showing in detail an example of the internal configuration of the face recognition system 5 of the present embodiment. The face recognition system 5 shown in FIG. 1 has a configuration in which a plurality of camera devices 10, a face recognition server 30, and a user terminal 60 are connected.

Each camera device 10 takes an image of a predetermined place such as in a preset store as an imaging area. Each camera device 10 acquires face image data including the face of a person passing through this imaging region, which appears in the captured image. Each camera device 10 includes an imaging unit 11, a face detection unit 12, a face cutting unit 13, and a communication unit 14. The camera device 10 of the present embodiment may be one or a plurality of cameras.

The image pickup unit has an image pickup device such as a CCD (Charge Couples Device) image sensor or a CMOS (Complementary Metal-Oxide Semiconductor) image sensor, and is incident from a preset imaging region. The light is imaged on the light receiving surface, and the optical image is converted into an electric signal. As a result, a frame of video data representing the video in the imaging region can be obtained.

The face detection unit 12 detects a face included in the image captured by the image pickup unit 11. This face detection process uses known techniques such as a method of detecting facial parts such as eyes, nose, and mouth, a method of detecting skin color, a method of detecting hair, and a method of detecting parts such as the neck and shoulders. It is a process of detecting a face by using it. Further, as the face detection processing method, a pattern recognition technique based on statistical learning may be used.

The face cutting unit 13 cuts out the face image data including the face detected by the face detecting unit 12 from the frame of the image captured by the imaging unit 11. The face image data to be cut out is data including a rectangular image having a size including the captured face. The face detection unit 12 and the face cutting unit 13 are functions executed by a processor 16 such as a CPU (Central Processing Unit), an MPU (Micro Processing Unit), or a DSP (Digital Signal Processor), for example. The processor 16 realizes the functions of the face detection unit 12 and the face cutting unit 13 by executing, for example, an application program stored in the internal memory.

The communication unit 14 is connected to the face recognition server 30 by wire or wirelessly, and transmits the face image data cut out by the face cutting unit 13 to the face recognition server 30. For example, when the camera device 10 is a network camera, the communication unit 14 can transmit face image data via an IP (Internet Protocol) network.

The face recognition server 30 collates the face included in the face image data received from each camera device 10 with the face registered in advance and recognizes the face. The face recognition server 30 includes a communication unit 31, a face feature amount extraction unit 32, a face feature amount comparison unit 33, an imaging time comparison unit 34, a storage unit 35, a pre-passing comparison source face feature amount data memory 41, and a face image data memory 42. Has.

The communication unit 31 receives the face image data from the camera device 10, and in response to a request from the user terminal 60 (for example, a search request described later), the user obtains a search result (that is, statistical data) that matches the request. Send to terminal 60.

The face feature amount extraction unit 32 as an example of the feature amount extraction unit extracts the face feature amount (hereinafter, simply referred to as “face feature amount”) from the face image data received by the communication unit 31. The facial feature amount extraction process is a process of extracting features such as the position of eyes, the positional relationship between eyes and nose and mouth, and how wrinkles are formed by using a known technique.

The face feature amount comparison unit 33 compares the face feature amount extracted by the face feature amount extraction unit 32 with the face feature amount registered in the pre-passing comparison source face feature amount data memory 41, and has a degree of similarity between them. Determines whether or not is higher than a predetermined value. Further, when these similarities are not higher than a predetermined value (that is, when the similarity is less than a predetermined value), the face feature amount comparison unit 33 determines that the similarity is low. The predetermined value (second threshold value) when determining that the similarity is low may be the same value as the predetermined value (first threshold value) when determining that the similarity is high, or is lower than the first threshold value. It may be a value.

As an example of the face feature amount memory, the pre-passing comparison source face feature amount data memory 41 has a storage condition in which the face feature amount extracted by the face feature amount extraction unit 32 from the face image data sent from the camera device 10 is a predetermined storage condition. When the condition is satisfied, this facial feature amount is stored for a certain period of time as a feature amount of facial image data including the face of a person who has passed through the imaging region. At this time, in addition to the feature amount of the face image data, information on the imaging time and the imaging area (position) is also saved as ancillary information.

The start of a certain period is the imaging time when the image data is captured by the camera device 10, and this imaging time is included in, for example, the header of the image data. Further, for a certain period of time, it can be arbitrarily set / changed according to the operation of the user terminal 60 according to the environment of the imaging region. For example, when the imaging area is in a store such as a restaurant, the fixed period is set to 2 hours as an example. When the imaging area is in an office building, the fixed period is set to 6 to 7 hours as an example. This fixed period may be set in advance in the face recognition server 30, or is set by a user such as an administrator operating the user terminal 60 to send a value for the fixed period to the face recognition server 30. May be good.

The face image data memory 42 stores the face image data of the face feature amount registered in the pre-passing comparison source face feature amount data memory 41 in association with the face feature amount for a certain period of time in the same manner as the face feature amount. Here, when a certain period of time elapses and the face feature amount stored in the pre-passing comparison source face feature amount data memory 41 is erased, the face image corresponding to the face feature amount stored in the face image data memory 42 is deleted. The data is also erased at the same time, but the image data may be stored for a longer period of time (for example, one year) without being erased. By accumulating a considerable amount of image data without erasing it, it can be used as big data.

In the imaging time comparison unit 34, the degree of similarity between the face feature amount extracted by the face feature amount extraction unit 32 and the face feature amount registered in the pre-passing comparison source face feature amount data memory 41 is a predetermined value (first). If it is higher than the threshold value), the imaging time of the face image data received from the camera device 10 is acquired. As described above, the imaging time is described in the header information of the face image data.

In addition, the imaging time comparison unit 34 determines that the registered face feature amount is a pre-passing comparison source face within a predetermined time from the imaging time of the face image data from which the face feature amount is extracted by the face feature amount extraction unit 32. It is determined whether or not the data is stored in the feature data memory 41. The predetermined time is a time set to prevent duplication of counting (double counting) of the same person when performing statistical processing described later, and is set to an appropriate value according to the environment of the imaging area. NS. For example, when the fixed period in which the facial feature amount can be stored in the pre-passing comparison source facial feature amount data memory 41 is 24 hours, the predetermined time is set to a shorter time such as 2 hours. The fixed period and the predetermined time may be the same period. In this case, the facial feature amount before the predetermined time does not exist in the pre-passing comparison source facial feature amount data memory 41, so that the imaging time comparison unit It is also possible to omit 34.

The storage unit 35 is connected to the pre-passing comparison source face feature data memory 41 and the face image data memory 42, stores the feature amount of the face image data in the pre-passing comparison source face feature data memory 41, and stores the feature amount in the pre-passing comparison source face feature data memory 41. The corresponding face image data is stored in the face image data memory 42. As will be described later, even if the face feature amount is not statistically processed, the face feature amount is saved if the face image data corresponding to the face feature amount satisfies the storage condition.

The face recognition server 30 further includes an age / gender determination unit 36, a statistics unit 37, a storage unit 38, a statistical data search unit 39, and an age / gender statistical data memory 43. The age / gender determination unit 36 estimates the age and gender using a known technique based on the facial feature amount extracted by the facial feature amount extraction unit 32. The estimated age may be expressed by a certain age range or by a representative value.

As an example of the statistical processing unit, the statistical unit 37 performs statistical processing that targets age and gender. In this statistical processing, for example, the number of faces included in the image captured by the camera device 10 is aggregated for each age and gender and registered as age-gender statistical data. Statistical processing may include not only counting up the number of people (number of faces) for each age and gender, but also calculating the ratio of each age group to the total number of people, the ratio of men and women, and the like. .. In addition, statistical processing may be aggregated so that it can be searched according to the AND condition or OR condition of the search conditions (date, period, place, age, gender, camera ID, etc.) described later, and may be used as age / gender statistical data. .. The age-gender statistical data is statistical data classified into at least age and gender, and may be subdivided by other elements. The storage unit 38 is connected to the age-gender statistical data memory 43, and stores the age-gender statistical data statistically processed by the statistical unit 37 in the age-gender statistical data memory 43. The age-gender statistical data memory 43 is used when the age-gender statistical data is stored for a long period of time and the statistical data is used.

The statistical data search unit 39 searches the age / gender statistical data memory 43 under predetermined search conditions in accordance with the request from the user terminal 60, and returns the search result to the user terminal 60.

The user terminal 60 is a general-purpose computer device that makes various settings and requests to the face recognition server 30, and includes a communication unit 61, a control unit 62, a display 63, and an operation unit 64. The communication unit 61 is connected to the face recognition server 30 by wire or wirelessly and can communicate with the face recognition server 30. For example, the communication unit 61 is connected to the face recognition server 30 via an IP network.

The control unit 62 comprehensively controls the operation of the user terminal 60, executes the application, and requests the face recognition server 30 to perform a search process according to the user's input operation. The display 63 displays various types of information, and displays search conditions, search results, and the like, which will be described later, on the UI screen. The operation unit 64 is a keyboard, a mouse, or the like, and accepts input operations from the user such as search conditions.

When the user terminal 60 is composed of a portable tablet terminal, the operation unit 64 is integrated with the display 63 as a touch panel, and input operations can be performed directly on the screen of the display 63.

The operation of the face recognition system 5 having the above configuration is shown.

FIG. 2 is a flowchart illustrating an example of an operation procedure of the face detection process in the face recognition server 30 of the present embodiment. The communication unit 31 in the face recognition server 30 receives the face image data sent from the camera device 10 (S1).

The face feature amount extraction unit 32 extracts the face feature amount from the received face image data (S2). As the facial features, for example, features such as the position of the eyes, the positional relationship between the eyes, the nose, and the mouth, and how to wrinkle are extracted. The face feature amount comparison unit 33 compares the face feature amount extracted by the face feature amount extraction unit 32 with the face feature amount registered in the pre-passing comparison source face feature amount data memory 41 (S3), and these It is determined whether or not the similarity of is higher than a predetermined value (first threshold value) (S4).

When the degree of similarity is not higher than a predetermined value (that is, when it is determined that the degree of similarity is low), the storage unit 35 passes the facial feature amount extracted by the facial feature amount extraction unit 32 in advance. It is saved in the data memory 41 (S7). The age / gender determination unit 36 estimates the age and gender using a known technique based on the facial feature amount extracted by the facial feature amount extraction unit 32, and the statistical unit 37 estimates the estimated age and gender. Perform statistical processing to be the target of statistics (S8). In this statistical processing, the number of age-gender statistical data classified by age and gender of a person, which is estimated from the extracted facial features, is counted so as to increase (count up) by a value of 1.

The storage unit 38 stores the age-gender statistical data statistically processed by the statistics unit 37 in the age-gender statistical data memory 43 (S9).

On the other hand, when it is determined in step S4 that the similarity is higher than a predetermined value (first threshold value), the imaging time comparison unit 34 captures the facial image data whose facial feature amount has been extracted by the facial feature amount extracting unit 32. The time is compared with the imaging time of the face feature amount registered in the pre-pass comparison source face feature data memory 41 (S5), and the features with high similarity stored in the pre-pass comparison source face feature data memory 41 are compared. It is determined whether or not the amount is stored within a predetermined time (within a predetermined time range) from the imaging time of the face image data from which the facial feature amount has been extracted by the facial feature amount extraction unit 32 (S6).

When it is out of the range of the predetermined time (S6, NO), the statistics unit 37 performs the above-mentioned statistical processing in step S8, assuming that it is the face feature amount of the new face image data. On the other hand, when it is determined in step S6 that the time is within the predetermined time range (S6, YES), the statistics unit 37 determines that the extracted facial feature amount has already been saved. Then, the statistics unit 37 determines whether or not the face image data of the face feature amount satisfies the storage condition (S10). When the storage condition is not satisfied (S10, NO), the face recognition server 30 discards the face feature amount without storing it in the pre-passing comparison source face feature amount data memory 41 (S11). On the other hand, when the face image data satisfies the storage condition (S10, YES), the storage unit 35 stores the face feature amount in the pre-passing comparison source face feature amount data memory 41 (S12).

Here, if, for example, it is set to save only the first image as a saving condition, the face image data determined in step S10 is not the first image, so the corresponding face feature amount is not saved. , Will be discarded. Further, when the storage condition is that the orientation of the face included in the face image data is front and the face is reflected in a large size, if the face image data does not correspond to this storage condition, the corresponding facial feature amount. Is discarded. Further, when it is a storage condition that all are saved, all the facial features are saved regardless of the image data. In addition to being set in advance in the storage unit 35, this storage condition may be changed by setting information from the user terminal 60 during the operation.

When the storage of the age-gender statistical data and the facial feature amount is completed in the processes of steps S1 to S12, the statistical data search unit 39 receives the request from the user terminal 60, and according to this request, the age-gender statistical data memory 43. The age-gender statistical data registered in is searched, and the corresponding age-gender statistical data is returned to the user terminal 60 (S13). After this, the face recognition server 30 ends this operation.

The user terminal 60 requests the face recognition server 30 to search the age / gender statistical data, and receives the search result from the face recognition server 30. FIG. 3 is a diagram showing a UI (user interface) screen displayed on the display 63 of the user terminal 60. On the left side of the screen of the display 63, search conditions for the user to search the age / gender statistical data registered in the age / gender statistical data memory 43 are displayed.

Here, as search conditions, keywords such as the date of imaging, the time zone (period), the place, the age, the gender, and the camera ID can be selected. The user can search for these search items under preset AND conditions or OR conditions by inputting a check mark in the input box of each search item.

When the user terminal 60 transmits the search condition set by the user to the face recognition server 30, the statistical data search unit 39 searches the age-gender statistical data registered in the age-gender statistical data memory 43 according to the search condition. The face recognition server 30 extracts the face feature amount of the person who matches the search condition from the age / gender statistical data memory 43, and is associated with the face feature amount registered in the face image data memory 42. The face image data is extracted from the face image data memory 42. The face recognition server 30 transmits the extracted face feature amount and face image data to the user terminal 60.

As a result, on the right side of the screen of the display 63 of the user terminal 60, the facial features and facial image data of the person who matches the search conditions are displayed. Here, as the search results, the facial feature amounts TK1, TK2, TK3 of the three persons and the facial image data G1, G2, G3 are displayed.

As described above, in the face recognition system 5 of the present embodiment, the face feature amount extraction unit 32 in the face recognition server 30 uses the face of the face image data including the face appearing in the image captured by the camera device 10 with respect to the imaging region. Extract the feature amount. The storage unit 35 stores the face feature amount of the face image data including the face that has passed through the imaging region in the pre-passing comparison source face feature amount data memory 41. The statistics unit 37 captures the facial features that are imaged by the camera device 10 and have a high degree of similarity to the facial features of the facial image data extracted by the facial feature extraction unit 32 by the camera device 10. When the pre-passing comparison original facial feature amount data memory 41 is not saved by a predetermined time before the time, the facial feature amount of the face image data extracted by the face feature amount extraction unit 32 is statistically processed.

As a result, the face recognition server 30 determines that the face features with high similarity stored at a time close to the imaging time belong to the same person, and the face features obtained at this imaging time are not statistically processed. , Efficient statistical processing can be performed so that the same person is not included twice without making complicated judgments. Further, in the face recognition system 5, since the face image data cut out from the video frame by the camera device 10 is transmitted to the face recognition server 30, the amount of data transmission can be reduced as compared with the case of transmitting the video data. ..

Further, when the face feature amount having a high degree of similarity is stored in the pre-passing comparison source face feature amount data memory 41 by a predetermined time, the face image data captured by the camera device 10 satisfies the predetermined storage condition. If not, the storage unit 35 discards the face feature amount of the face image data extracted by the face feature amount extraction unit 32. As a result, unnecessary facial features can be eliminated and the amount of stored data can be reduced.

Further, when the face image data captured by the camera device 10 satisfies the predetermined storage conditions, the face feature amount of the face image data extracted by the face feature amount extraction unit 32 is passed in advance to the comparison source face feature amount data memory. Save to 41. As a result, the data of the effective facial features increases, and the accuracy of determining the similarity of the facial features improves.

Further, the statistics unit 37 has a face feature having a low degree of similarity that does not reach a predetermined value (second threshold value) with respect to the face feature amount of the face image data imaged by the camera device 10 and extracted by the face feature amount extraction unit 32. When the amount is stored in the pre-passing comparison source face feature amount data memory 41, the face feature amount of the face image data extracted by the face feature amount extraction unit 32 is statistically processed. As a result, statistical processing can be performed without missing the face of the first person.

In addition, the predetermined time can be set according to the environment of the imaging area such as in a restaurant or a building. As a result, statistical processing can be performed so that the same person is not included in duplicate regardless of the environment of the imaging region.

Further, the face recognition server 30 searches the age-gender statistical data (feature amount of statistically processed face image data) registered in the age-gender statistical data memory 43 in response to a request from the user terminal 60 (terminal device). Then, the search result is returned to the user terminal 60. This can contribute to the use of users and is expected to be used in various ways.

(Modified example of this embodiment)
FIG. 4 is a block diagram showing in detail an example of the internal configuration of the face recognition system 5A of the modified example of the present embodiment. The same components as those in the present embodiment are designated by the same reference numerals, and the description thereof will be omitted. In the face recognition system 5A of the modified example of the present embodiment, unlike the above-described embodiment, the camera device 10A has only the image pickup unit 11 and the communication unit 14, and the image data captured by the image pickup unit 11 is displayed. It is simply transmitted to the face recognition server 30A by the communication unit 14 as it is.

Unlike the present embodiment, the face recognition server 30A has a face detection unit 52 and a face cutting unit 53 in the processor 40. The face detection unit 52 detects the face included in the image with respect to the image data (video) transmitted from the camera device 10A, as in the face detection unit 12 of the present embodiment. Similar to the face cutting section 13 of the present embodiment, the face cutting section 53 cuts out face image data including the face detected by the face detecting section 52 from the frame of the video.

As described above, in the face recognition system 5A of the modified example of the present embodiment, the load-bearing processing is efficiently concentrated on the face recognition server 30A, so that the load on the camera device 10A can be reduced. In this way, the camera device 10A only transmits the captured image data (video) to the face recognition server 30, so that a simple configuration is sufficient. For example, the camera device 10A can be used even if the camera device 10A is already installed in the imaging area. Is. Further, in the face recognition system 5A of the modified example of the present embodiment, the camera device 10A can also be used with a general-purpose network camera.

Although various embodiments have been described above with reference to the drawings, it goes without saying that the present invention is not limited to such examples. It is clear that a person skilled in the art can come up with various modifications or modifications within the scope of the claims, which naturally belong to the technical scope of the present invention. Understood.

INDUSTRIAL APPLICABILITY The present invention is useful because when using an image captured by a camera device, statistical processing can be efficiently performed so that the same person is not included in duplicate without making a complicated judgment.

5, 5A Face recognition system 10, 10A Camera device 11 Imaging unit 12, 52 Face detection unit 13, 53 Face cutting unit 14 Communication unit 16, 40, 66 Processor 30, 30A Face recognition server 31 Communication unit 32 Face feature amount extraction unit 33 Face feature comparison unit 34 Imaging time comparison unit 35, 38 Storage unit 36 Age gender determination unit 37 Statistics department 39 Statistical data search unit 41 Pre-pass comparison source Face feature data memory 42 Face image data memory 43 Age gender statistical data memory 60 User terminal 61 Communication unit 62 Control unit 63 Display 64 Operation unit TK1, TK2, TK3 Face feature amount G1, G2, G3 Face image data

Claims (3)

A face recognition server to which a camera device is connected
Extracting Kaotoku symptoms of a person appearing in the video data taken by the camera device,
Compares the similarity between the Kaotoku symptom amount that has been stored for a period of time in advance and Kaotoku symptoms amount the extracted,
When the similarity is less than a predetermined value, the extracted facial feature amount is saved and processed as a statistical object.
When the similarity is equal to or greater than a predetermined value, the imaging time at which the extracted face of the person is imaged is compared with the imaging time at which the face determined to have the similarity is equal to or greater than the predetermined value.
When the imaging time when the face whose similarity is determined to be equal to or higher than the predetermined value is within a predetermined time range from the imaging time of the video data from which the facial feature amount is extracted, the statistics Exclude from the target
When the imaging time of the face whose similarity is determined to be equal to or higher than the predetermined value is not within the predetermined time range from the imaging time of the video data from which the facial feature amount is extracted, the statistics. Equipped with a processor that performs statistical processing as an object of
Face recognition server. This is a face recognition method in a face recognition system in which a camera device and a face recognition server are connected.
Extracting Kaotoku symptoms of a person appearing in the video data taken by the camera device,
Compares the similarity between the Kaotoku symptom amount that has been stored for a period of time in advance and Kaotoku symptoms amount the extracted,
When the similarity is less than a predetermined value, the extracted facial feature amount is saved and processed as a statistical object.
When the similarity is equal to or greater than a predetermined value, the imaging time at which the extracted face of the person is imaged is compared with the imaging time at which the face determined to have the similarity is equal to or greater than the predetermined value.
When the imaging time when the face whose similarity is determined to be equal to or higher than the predetermined value is within a predetermined time range from the imaging time of the video data from which the facial feature amount is extracted, the statistics Exclude from the target
When the imaging time of the face whose similarity is determined to be equal to or higher than the predetermined value is not within the predetermined time range from the imaging time of the video data from which the facial feature amount is extracted, the statistics. Statistical processing as the target of
Face recognition method. To the face recognition server, which is a computer connected to the camera device so that it can communicate with the camera device.
A step of extracting the facial features of a person appearing in the video data captured by the camera device, and
A step of comparing the degree of similarity between the extracted facial features and the facial features stored in advance for a certain period of time, and
When the similarity is less than a predetermined value, the step of storing the extracted facial feature amount and processing it as an object of statistics, and
When the similarity is equal to or greater than a predetermined value, a step of comparing the imaging time when the extracted face of the person is imaged and the imaging time when the face determined to have the similarity equal to or greater than the predetermined value is imaged. ,
When the imaging time when the face whose similarity is determined to be equal to or higher than the predetermined value is within a predetermined time range from the imaging time of the video data from which the facial feature amount is extracted, the statistics Steps to exclude from the target and
When the imaging time of the face whose similarity is determined to be equal to or higher than the predetermined value is not within the predetermined time range from the imaging time of the video data from which the facial feature amount is extracted, the statistics. To realize the step of statistical processing as the target of
Face recognition program.

Images