JP7054075B2 - Information processing system, information processing method, and program - Google Patents

Description

The present invention relates to an information processing system, an information processing method, and a program.

In recent years, an increase in the amount of damage caused by shoplifting in supermarkets and retail stores (hereinafter collectively referred to as stores) has become a problem. Therefore, various measures have been taken to reduce the damage caused by shoplifting.

The detection of general shoplifting criminals is realized by allocating specialized personnel called shoplifting G-men to stores and using shoplifting G-men to control the shoplifting site. However, the number of stores that can be assigned to professional personnel called shoplifting G-men is limited because they require special skills and labor costs. Therefore, the shoplifters detected at present are only the tip of the iceberg, and a method of dealing with shoplifting without human intervention such as shoplifting G-Men is desired.

As a method of dealing with shoplifting without arranging shoplifting G-men, there is a method using a crime prevention management system using face recognition technology. In this method, the store staff monitors the image of the surveillance camera placed in the store, and if a person who seems to be suspicious is found, the face information (mainly the face image) of that person is manually obtained each time. Register with the security management system. After that, face recognition is performed on the image of the surveillance camera by face recognition technology, and when an image matching the face image of the registered person is detected, the store staff is notified to that effect.

However, with this method, the appearance and unrelated behavior of a person who seems to be suspicious can also be seen, so when store staff who do not have special skills such as shoplifting G-men perform registration work, personal emotions may affect it. It is difficult to avoid, and the credibility of the registered data may be unfounded. Therefore, from the viewpoint of shoplifting prevention where misrecognition is not allowed, it is difficult to operate such a method, and a method without room for personal feelings is desired.

A technique for actually capturing the shoplifting site with image recognition technology has been proposed, but it is difficult to reliably detect a skillful shoplifter with the current image recognition technology.

As a technique related to shoplifting prevention, there is a technique disclosed in Patent Documents 1 to 3, which will be briefly described below.

Patent Document 1 discloses a product management system that manages the inventory status of products displayed in a showcase using a computer. This system includes a product acquisition detection means that detects whether or not a product has been acquired from a showcase, and a first imaging means that captures an image of the person who acquired the product when the product acquisition detection means detects that the product has been acquired. It is provided with a sales acquisition means for acquiring sales information of a product at a cash register when the product acquisition detection means detects that the product has been acquired. This system further shoplifts based on the second image pickup means that images the purchaser of the product when the sales information is acquired by the sales acquisition means, the image pickup information by the first image pickup means, and the image pickup information by the second image pickup means. It is provided with a control means for determining an action and an information storage means for storing the image pickup information by the two image pickup means, and determines the shoplifting action based on the image pickup information.

Patent Document 2 discloses a sales management system capable of reducing shoplifting damage. This system is equipped with a store controller. When the settlement image data from the settlement camera recorded in the recording device and the display shelf image pickup data from the display shelf camera match, the store controller deletes the settlement image pickup data and the display shelf image pickup data from the recording device. However, if the display shelf imaging data remains in the predetermined period, the store controller notifies that fact and calls attention. After that, the face recognition system is used to identify a person who is likely to be shoplifted.

Patent Document 3 discloses a security system that can be easily adopted even in a store that sells a large number of products and sells a large number of products. This system uses RFID (Radio-Frequency Identification Device) to grasp the products in the store, and periodically scans the RFID in the store to grasp the products that disappeared from the store even though they did not go through the cash register. do. If the system determines that there is a product that has disappeared from the store even though it has not passed through the cash register, the person who was in the store between the current scan time and the previous scan time is listed as a candidate shoplifter. .. Then, by repeating the above operation, candidates are piled up to increase the detection accuracy.

Japanese Unexamined Patent Publication No. 2013-188401 Japanese Unexamined Patent Publication No. 2012-2429112 Japanese Unexamined Patent Publication No. 2013-134644

According to the disclosure of Patent Document 1, the system of Patent Document 1 contributes not only to the identification of shoplifters but also to the recidivism of shoplifters who have become habitual. However, in the system of Patent Document 1, it is necessary to take an image of a person who has taken out a product every time the product is taken out from the product display unit even in a normal product taking out, and an operation for shoplifting action determination is required. In addition, the system of Patent Document 1 does not make any diagnosis or determination of the person's behavior itself, so that even if a person has suspicious behavior that leads to shoplifting, it is not taken into consideration.

Even in the system of Patent Document 2, since no diagnosis or discrimination is made regarding the behavior of a person itself, even if a person has a suspicious behavior that leads to shoplifting, it is not taken into consideration.

Since the system of Patent Document 3 does not consider reducing the accumulated information of candidates, if it is accidentally listed as a candidate many times, it may make a wrong judgment. Further, like the systems of Patent Documents 1 and 2, the system of Patent Document 3 does not make any diagnosis or discrimination regarding the behavior of a person itself, so that a person may behave suspiciously as if it leads to shoplifting. If so, it is not taken into account.

Further, any of the systems of Patent Documents 1 to 3 is considered to be executed only in one store.

In view of the above problems, the present invention is an information processing system and information processing that can contribute to the prevention of shoplifting by focusing on the detection of actions that are thought to lead to shoplifting rather than the shoplifting action itself. It seeks to provide methods and programs.

According to the first aspect of the present invention, a motion line detecting means for extracting face information and detecting motion line information for each person included in the image based on an image acquired from a camera is not purchased. Despite this, the product detection means that detects the lost product and generates the information processing information including the information about the detected product and the information of the time zone when the product is lost, and the flow line information and the information processing information. Based on this, an information processing system including a control means for updating a suspicious numerical value stored in association with each person included in the image is provided.

The second aspect of the present invention is an information processing method executed by an information processing system, in which face information is extracted for each person included in the image and movement line information is obtained based on the image acquired from the camera. To detect, to detect a lost product even though it has not been purchased, to generate shoplifting information including information on the detected product and information on a time zone in which the product is lost, and the above-mentioned operation. It is an information processing method including updating a suspicious numerical value stored in association with each person included in the image based on the line information and the shoplifting information .

A third aspect of the present invention is a motion line detecting means for extracting face information and detecting motion line information for each person included in the image based on an image acquired from a camera, which has not been purchased. Based on the product detection means, the flow line information, and the shoplifting information, which detects the lost product and generates the shoplifting information including the information about the detected product and the information of the time zone when the product is lost. , Is a program for functioning as a control means for updating a suspicious numerical value stored in association with each person included in the image .

According to the present invention, it is possible to provide an information processing system that can contribute to the prevention of shoplifting by emphasizing the detection of an act that seems to lead to shoplifting rather than the shoplifting act itself.

It is a figure for demonstrating the schematic structure of the shoplifting prevention system which is an information processing system which concerns on this invention. It is a figure which showed an example of the layout of the store to which this invention is applied, and the arrangement example of the camera installed inside and outside the store in order to construct the shoplifting prevention system in this store. It is a block diagram which shows the structural example of the store-side apparatus which concerns on embodiment of this invention. It is a block diagram which shows the structural example of the center side apparatus which concerns on embodiment of this invention. It is a figure which showed an example of the information stored in the memory of the center side apparatus which concerns on embodiment of this invention. It is a flowchart for demonstrating the operation of the store-side apparatus which concerns on embodiment of this invention. It is a flowchart for demonstrating the operation of the center side apparatus which concerns on embodiment of this invention.

The shoplifting prevention system, which is an information processing system according to the present invention, has information on the face of a person who visited the store, information on the date and time and time zone, information on the date and time when the product disappeared from the store, and information on the movement line of the person in the store. By combining these, it is possible to statistically derive a person who may repeat shoplifting and prevent re-offending.

Hereinafter, embodiments of the shoplifting prevention system according to the present invention will be described.

FIG. 1 is a diagram for explaining a schematic configuration of a shoplifting prevention system, and is one or more of a center-side device installed in a data center DC and one or more installed in each store and connected to the center-side device via a communication line. It is composed of equipment on the store side. Here, only two stores α and β are shown, and the configuration of the center-side device and the store-side device will be described later.

FIG. 2 is a diagram showing an example of the layout of a store and an example of the arrangement of cameras installed inside and outside the store to construct a shoplifting prevention system in this store.

In FIG. 2, in addition to the product display shelves 100 being installed along the wall in the store, a plurality of product display shelves 200 are installed at intervals. In addition, one or more checkout counters 300 for performing settlement processing are installed in the store. In a store with such a layout, a plurality of cameras CM1 are installed so that all areas of the product display shelves 100 and 200 can be photographed, a checkout counter 300, a visitor at the entrance / exit 400 of the store, and an exit. A plurality of cameras CM2 and CM3 are installed so that a person can be photographed. The camera CM1 and its shooting area, in other words, the display location of the product, are associated in advance. In addition, when installing the camera, it is devised so that there is no blind spot area.

FIG. 3 is a block diagram showing a configuration example of the store-side device.

In FIG. 3, as described above, the store-side device 10 according to the present embodiment includes two cameras CM3 for entrance / exit imaging, which are installed at the entrance / exit 400 of the store and capture images of a person entering and exiting, respectively. It is provided with a plurality of cameras CM1 and CM2 that are installed at predetermined locations in the store and capture images in a predetermined shooting area.

The store-side device 10 also includes a store-side face recognition unit 20, a flow line detection unit 30, a product detection device 40, a suspicious behavior detection unit 50, and a store-side control device 60. The store-side face recognition unit 20, the flow line detection unit 30, the product detection device 40, the suspicious behavior detection unit 50, and the store-side control device 60 can be realized by one computer or CPU, but in the following, each unit is individually. The case of preparing for the above will be described. The functions and operations of the above parts will be briefly described below.

The store-side face authentication unit 20 receives image information from the two entrance / exit imaging cameras CM3 and image information from the plurality of cameras CM1 and CM2, and based on the image information of the two entrance / exit imaging cameras CM3. Face information is extracted from the face images of the person entering and leaving, face authentication is performed, and the entry date and time, the exit date and time, and the face information are output to the store side control device 60 for each person. The "hour" of the date and time is indicated by the time zone. In addition, although the face image itself is used as the face information here, the feature point information used for face recognition such as the contour of the face extracted from the face image, the distance between the eyes and the length of the cheekbones is used. May be.

The motion line detection unit 30 receives image information from two cameras CM3 for entrance / exit imaging and image information from a plurality of cameras CM1 and CM2, and includes the image information from the plurality of cameras CM1 and CM2 in the image. The face information is extracted from the face image for each person, the movement line of the person is detected, and the movement line information detected for each person and the face information of the corresponding person are output to the store side control device 60.

The product detection device 40 detects whether or not there is a lost product even though the product in the store has not been purchased, and if there is a lost product (lost product), the product is related to the lost product. The information and the lost date / time and time zone information are output to the store side control device 60. The information about the lost product includes the place (display place) information where the lost product was displayed. As a specific example of the product detection device 40, RFID is attached to each product displayed on the product display shelves 100 and 200, and an RFID reader is installed on the product display shelves 100 and 200 to periodically present the presence or absence of products. There is a form of detecting. Alternatively, since this type of store is usually equipped with a POS (Point of Sales) device, the number of products is confirmed by a mobile terminal (with a barcode reading function) for store staff connected to the POS device. But it's okay.

The suspicious behavior detection unit 50 receives the flow line information detected for each person and the face information of the corresponding person detected by the flow line detection unit 30 via the store side control device 60, and uses these information to obtain a product. When the presence or absence of a person who takes suspicious behavior is checked near the display shelves 100 and 200 and it is determined that there is a person who takes suspicious behavior, the face information and the location information of the person corresponding to the determined flow line information are input to the store side control device 60. Output to.

The store-side device 10 further includes a communication unit 70 for wirelessly or wiredly communicating with the center-side device, a notification unit 80 for notifying store staff of warning against shoplifting based on warning information from the center-side device, and a store. A memory (store side memory) 90 for recording various detection information and discrimination information from each of the above parts under the control of the side control device 60 is provided. The memory 90 also includes a memory for storing a program necessary for the store-side control device 60 to execute a control operation.

The store-side control device 60 reads a program for executing the control operation from the memory 90 and executes the following control operation.

The store-side control device 60 controls the imaging operation of the cameras CM1 to CM3. The store-side control device 60 also acquires the entry date / time, exit date / time, and face information for each person from the store-side face authentication unit 20, and outputs (transmits) to the center-side device together with the store information. Store information is identification information given to each store in order to identify the store. The store-side control device 60 further transmits the flow line information detected for each person acquired by the flow line detection unit 30 and the face information of the corresponding person to the center-side device together with the store information, and detects each person. The flow line information and the face information of the corresponding person are output to the suspicious behavior detection unit 50, and the suspicious behavior is detected. When the store-side control device 60 further receives the information on the lost product and the lost date / time and time zone information from the product detection device 40, the store-side control device 60 outputs these information as shoplifting information to the center-side device together with the store information. The store-side control device 60 further outputs the face information, location information, and store information of the corresponding person, which is sent when the suspicious behavior detection unit 50 determines that there is a suspicious person, to the center-side device.

As described above, when the store-side control device 60 receives the warning information from the center-side device via the communication unit 70, the store-side control device 60 urges the store-side staff to warn against shoplifting based on the received warning information to the notification unit 80. Notify the notification.

FIG. 4 is a block diagram showing a configuration example of the center side device 500.

In FIG. 4, the center-side device 500 according to the present embodiment includes a center-side control device 510, a center-side face recognition unit 520, a communication unit 530 that communicates wirelessly or by wire with the store-side device of each store, and a memory that serves as a database. (Center side memory) 540 is provided. The memory 540 also includes a memory that stores a program necessary for the center-side control device 510 to execute a control operation. The center-side control device 510 and the center-side face recognition unit 520 can be realized by one computer or CPU, but the case where each unit is individually provided will be described below.

The center-side control device 510 reads a program for executing the control operation from the memory 540 and executes the following control operation.

The center-side control device 510 receives the entry / exit date / time, face information, and store information of each person to the store sent from the store-side device, and these information are given to each person together with the suspicious numerical value. It is stored in the memory 540. The center-side control device 510 also receives information on lost products, lost date / time and time zone information, and store information sent from the store-side device, and stores these information in the memory 540. The center-side control device 510 further receives face information, location information, and store information of the corresponding person sent from the store-side device when the suspicious behavior detection unit 50 of the store-side device determines that there is a suspicious person. While accumulating this information in the memory 540, the suspicious numerical value described later is updated.

The center-side face recognition unit 520 obtains information about lost products, lost date / time and time zone information, store information, movement line information detected for each person, and face information of the corresponding person, which is obtained via the center-side control device 510. And, using the store information, it is checked whether there is a person who has approached the display place of the lost product at the time and time when the product disappears, and if it is determined that there is a person, is the face information of the corresponding person stored in the memory 540? investigate. That is, when the center side face recognition unit 520 collates the face information of the corresponding person obtained as a result of the above determination with the face information of each person stored in the memory 540 and detects the corresponding face information. It is output to the center side control device 510. In response to this, the center-side control device 510 executes an operation of updating the memory 540 so as to increase the suspicious value corresponding to the person in the detected face information. If the center side face recognition unit 520 determines that there is a person, but the face information of the corresponding person does not exist in the memory 540, the center side control device 510 displays the face information and the store information together with a suspicious numerical value. It is newly stored in the memory 540.

Next, the operations of the store-side device 10 and the center-side device 500 of the store α will be described step by step with reference to FIGS. 6 and 7 which are flowcharts thereof.

1. 1. The store-side control device 60 acquires face information of a person entering and exiting by a camera CM3 installed at the entrance / exit 400 of the store α, and stores the face information, the date and time of entry / exit, and the store information in the memory 90. , Is transmitted to the center side device 500 of the data center via the communication unit 70 (step S101). The center-side control device 510 of the center-side device 500 stores the face information, entry / exit date / time, and store information of all persons transmitted from the store α in the memory 540 together with the suspicious data (step S201).

FIG. 5 shows an example of the storage form in the memory 540. For the storage of information, the storage area of the memory 540 is divided for each person, and the face information with a symbol or number for identifying the person for each section is associated with the store information, the entry / exit date and time, and the suspicious data described later. And remember. Hereinafter, the storage area divided for each person is referred to as a person storage area.

When new face information and entry / exit date / time and store information are transmitted, the center side control device 510 collates the face information with the face information already stored, and if there is matching face information. In, new store information and entry / exit date / time are added to the person storage area (section) of the matching face information. In the following description, "match" includes the case where the degree of similarity is equal to or more than a predetermined value as a result of collation based on the degree of similarity.

2. 2. The store-side control device 60 of the store α causes the product detection device 40 to periodically detect whether or not there is a product (lost product) that has been lost from the store α even though it has not been purchased (step). S102). When the store-side control device 60 receives the output of the lost product from the product detection device 40, the store side control device 60 stores the lost product name and date and time, the lost time zone and location information in the memory 90, and the lost product name and date and time are lost. Store information is added to the time zone and location information and transmitted to the center side device 500 (step S103). In the following, information obtained by adding store information to lost product name, date and time, lost time zone and place information may be collectively referred to as lost information.

3. 3. When the center-side control device 510 detects that the lost information has been received (step S202), the center-side control device 510 stores the received lost information in the memory 540 (step S203). Lost information is stored for each store.

The center-side control device 510 periodically records and updates the suspicious data of the person who entered the store α based on the information acquired in the above 1 and 2. The suspiciousness data is shown here as a numerical value, and will be referred to as a suspiciousness numerical value hereafter.

In the store α, the store side control device 60 causes the flow line detection unit 30 to acquire the flow line information and the face information of the person in the store α, stores these information in the memory 90, and stores the information in the center side device 500. Transmit (step S104).

When the center side control device 510 detects that the flow line information and the face information of the person in the store α transmitted from the store α are received (step S202), the received information is stored in the memory 540 (step S202). S203).

The store-side control device 60 also uses the flow line information detected for each person and the face information of the corresponding person detected by the flow line detection unit 30 to the suspicious behavior detection unit 50 for the product display shelf 100. , Check for the presence or absence of a person who behaves suspiciously near 200 (step S105), and when the determination result of the presence or absence of a suspicious person is obtained, the face information of the corresponding person, the location information where the person was, and the store information are obtained. It is transmitted to the center side device 500 as suspicious information (step S106).

When the center-side control device 510 receives the suspicious information (step S204), the center-side control device 510 uses the face information in the suspicious information to collate the face information of all the stored persons with the face corresponding to the person. Check for information (step S205). When the center side control device 510 can also extract the corresponding face information (Yes in step S205), the suspicious value is added to the person in the extracted face information in the memory 540 by a predetermined value (Yes). Step S206). When the corresponding face information cannot be extracted, the center-side control device 510 stores the face information included in the suspicious information in a new person storage area of the memory 540 together with the store information (S207).

The center side control device 510 further sets the date and time when the product is lost and the time zone based on the information acquired in 1 and 2 above, the store information, the date and time when the product is lost, the flow line information of the time zone, and the face information. , The product display shelves 100 and 200 on which the lost products are placed, that is, the operation of determining whether or not there is a person who has stayed for a certain period of time in front of the display place is executed (step S208). The center side face recognition unit 520 is used for this determination operation. When the center side control device 510 determines that there is a corresponding person, the center side control device 510 uses the face information of the person to collate the face information of all the persons stored in the memory 540, and the face corresponding to the person. Extract information. When the center-side control device 510 can also extract the corresponding face information, the center-side control device 510 also adds a predetermined value to the suspicious value of the person storage area that stores the extracted face information (step S209). As a result of the determination, in the case of a person who has just passed in front of the product display shelf or has not even passed, the suspicious value may not be added but also decreased.

4. By repeating the above steps 1 to 3 daily with the center-side device 500 for a plurality of stores, the center-side device 500 updates the suspicious value of a person who may repeat recidivism in the stored information as shown in FIG. .. When updating the suspiciousness value, the suspiciousness value can be prevented from increasing in the case of a person unrelated to shoplifting by the device described in 3 above.

Based on this information, when there is a person whose suspicious value has risen above a certain threshold value (Yes in step S210), the center-side device 500 corresponds to that fact as suspicious person information together with the face information of the corresponding person. Not only the store (α) but all the stores are notified (step S211). At each store that received the notification (Yes in step S107), the face information of the notified person is displayed on the display dedicated to the store staff and the store staff is notified (step S108).

After that, the store-side control device 60 collates the face information of the person photographed at the time of admission with the face information of the notified person by the store-side face recognition unit 20 (step S109), and the corresponding person is admitted. If it is determined that the item is, the store clerk is notified to that effect to call attention, so that shoplifting can be prevented (step S110).

As described above, in each of the plurality of stores α and β, the face information of the person who has entered each store and the suspicious data (suspicious numerical value) related to the person are shared through the center side device. It is possible to detect a person with a high suspicious value not only in store α but also in other store β.

5. As a method for preventing erroneous detection, the following methods including the items described above can be considered.

5-1. When a person in front of the shelf where the lost product was placed stays for a certain period of time due to the detection process using the flow line information on the center side device, the suspicious value is added to the plus, and conversely, in front of the shelf. If it just passed or did not even pass, it controls not only not to add but also to decrease.

5-2. When the suspicious behavior detection unit 50 of the store-side device 10 detects a suspicious behavior that is suspected that a certain person may be deciding the item for shoplifting by the detection process using the flow line information, the suspicious behavior is suspicious. Information indicating behavior detection, face information, and store information are transmitted to the center side device 500 to add a suspicious value in the memory 540. The detection of suspicious behavior is realized as follows. The motion patterns to be detected as suspicious behaviors are categorized into several suspicious behavior patterns and stored in the memory 90 in advance. When the suspicious behavior detection unit 50 detects an unusual movement in the flow line information sent from the flow line detection unit 30 under the control of the store-side control device 60, the flow line information becomes a suspicious behavior pattern. Determine if it is applicable. When it is determined to be applicable, the suspicious behavior detection unit 50 immediately displays information indicating that the existence of a person who takes suspicious behavior via the store side control device 60 is detected, along with face information and store information, in the center side device 500. Send to. Of course, the store staff may be notified by a display dedicated to the store staff that there is a person who behaves suspiciously. The suspicious behavior detection unit may be installed in the center-side device 500, and the suspicious behavior detection may be executed in the center-side device 500.

5-3. In the case of another person who accidentally visits the store multiple times at the same time as the person who shoplifted, the suspicious value may be mistakenly recorded, so shorten the time interval to detect the loss of the product and shorten the time interval. Improves the accuracy of suspicious numbers.

5-4. As a suspicious numerical value, it is not only recorded as an individual, but by managing information associated with a person who existed in the store at the same time, behavior that deviates from that (for example, it existed in the store when the suspicious numerical value was recorded). If there is nothing when you visit the store α or visit another store β when you are not with a person, the suspiciousness value is subtracted.

5-5. Even if the suspiciousness value exceeds the threshold value, the data can be cleared manually if it is found to be a false positive or if a police officer arrests the shoplifting site.

The handling of suspicious numbers will be explained using an example.

It is assumed that three people, X, Y, and Z, each stay in the store α in FIG. 1 at the time zone shown in FIG. 1 on the same day. It is assumed that the products (beer, ice cream) shown in FIG. 1 are suitable for shoplifting at the store α. Then, the three people X, Y, and Z who stayed at the store α during the time and date (2016/7/5, 11:00 to 14:00) when the product (beer) disappeared are suspicious. Record.

At that time, by the flow line detection unit and suspicious behavior detection unit of the person installed in the store α, in front of the product display shelf where the product disappeared during the time zone (11:00 to 14:00) when the product disappeared. When the presence of the person X who has stayed for a certain threshold time or longer is detected, the suspicious value of the person X corresponding to the suspicious behavior detection is counted as a large value. On the contrary, the suspicious figures are not recorded for the persons Y and Z who have just passed or did not pass in front of the product display shelves where the products have disappeared even though they stayed at the store α at the same time. By repeating this, the credibility of the suspicious value increases.

Then, when the suspicious value of a certain person exceeds a predetermined threshold value, the center-side device notifies all the corresponding stores together with the face information of the corresponding person. After that, each store that received the notification will collate the face information of the person photographed at the time of admission with the face information of the person who received the notification, and if the entry of the corresponding person is detected, the store clerk will be alerted. , Shoplifting can be prevented by strengthening the patrol in the store and broadcasting about the strengthening of crime prevention in the store.

In the memory 90 in the store-side device 10 and the memory 540 in the center-side device 500, the information whose storage time has passed a certain period may be automatically erased. In particular, the information stored in the person storage area of the memory 540, the store information related thereto, the entry / exit date / time information, and the suspicious numerical value information are automatically deleted if the suspicious numerical value does not change for a certain period. Is also good.

The shoplifting prevention system, the store-side device, and the center-side device according to the present invention can be applied to various sales facilities such as supermarkets and retail stores.

DC data center CM1 to CM3 Camera 100, 200 Product display shelf 300 Registration counter 400 Doorway 10 Store side device 60 Store side control device 500 Center side device 510 Center side control device

Claims (6)

Based on the image acquired from the camera, the flow line detecting means for extracting face information and detecting the flow line information for each person included in the image, and
A product detection means that detects a lost product even though it has not been purchased and generates shoplifting information including information on the detected product and information on a time zone in which the product is lost.
A control means for updating the suspicious numerical value stored in association with each person included in the image based on the flow line information and the shoplifting information.
Information processing system equipped with. The control means identifies a person who has stayed at a display place where the product was displayed for a predetermined time in the time zone when the product disappears, and updates the suspicious value associated with the specified person . The information processing system according to claim 1. The information processing system according to claim 1 or 2, wherein the control means reduces the suspicious value of a person who has not stayed at the display place where the product was displayed for a predetermined time during the time when the product disappeared. .. The information processing system according to any one of claims 1 to 3, further comprising a notification means for notifying that the person has visited the store when the visit of a person whose suspiciousness value is equal to or higher than a predetermined value is detected by face authentication. An information processing method executed by an information processing system.
Based on the image acquired from the camera, face information is extracted for each person included in the image and flow line information is detected.
Detecting lost products even though they have not been purchased, and generating shoplifting information including information on the detected products and information on the time zone when the products are lost.
Based on the flow line information and the shoplifting information, updating the suspicious numerical value stored in association with each person included in the image, and
Information processing methods including. Computer,
A flow line detecting means for extracting face information and detecting flow line information for each person included in the image based on an image acquired from a camera.
A product detection means that detects a lost product even though it has not been purchased, and generates shoplifting information including information on the detected product and information on a time zone in which the product is lost.
A control means for updating a suspicious numerical value stored in association with each person included in the image based on the flow line information and the shoplifting information.
A program to function as.

Images