Operator fraud detection system

Abstract

To monitor an entire room of an operator who works remotely so as to detect fraud, and report the fraud to an administrator.SOLUTION: An operator fraud detection system comprises: a fish-eye camera 23 connected to an operator terminal 20 for capturing an operator 2 and a work environment; a camera monitoring unit 221 that detects a fraudulent or abnormal event in the operator 2 or the work environment from an image captured by the fish-eye camera 23; and a warning processing unit 222 that outputs a warning to the operator 2 when the event is detected by the camera monitoring unit 221, and locks a portion related to a contact center business in the operator terminal 20, wherein: the camera monitoring unit 221 identifies a substantially rectangular object held by the operator 2 by image recognition processing from a moving image, and further determines whether or not the substantially rectangular object is a smartphone, and it is assumed that the camera monitoring unit has detected the event when this substantially rectangular object is a smartphone.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a security technology for contact centers, and more particularly to a technology that is effective when applied to an operator fraud detection system that detects fraud committed by an operator who performs work remotely. [Background technology]

[0002] Contact centers (a general term that includes various names such as call centers and help desks; hereafter referred to as "CC") that respond to customer inquiries by telephone and other means may handle personal information including customer contact details. Therefore, to prevent operators from illegally taking out customer information, security measures such as prohibiting the bringing in of mobile information terminals such as smartphones and recording media when entering the work area are taken.

[0003] As a means for detecting attempts by people to obtain images or audio by illegal means, for example, Japanese Patent Application Laid-Open No. 2002-122678 (Patent Document 1) describes a mechanism that can prevent users from having their privacy or confidential information violated by a camera or a microphone, even in cases where the data intercepted or secretly captured is simply recorded in a bugging or secret camera device and is not transmitted wirelessly at all times but is transmitted wirelessly in batches at specified times, or where the data intercepted or secretly captured is transmitted via a wired line rather than wirelessly. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Laid-Open No. 2002-122678 Summary of the Invention [Problem to be solved by the invention]

[0005] According to the conventional technology, for example, it is possible to prevent or deter an operator from illegally obtaining personal information displayed or output on an operator terminal by eavesdropping or secretly taking a photo within a business area where CC business is performed.

[0006] On the other hand, due to the impact of the recent COVID-19 pandemic, companies are now required to consider building infrastructure that takes into account the possibility of operators performing CC work from home as a permanent form of operation, so that CC services can continue even in situations where working from home is strongly requested. In this case, in order to balance the shift to working from home with business security, a system is required that monitors the entire room where the operator works from home and notifies an administrator if any fraud or abnormality by the operator or a third party is detected.

[0007] Therefore, an object of the present invention is to provide an operator fraud detection system that monitors the entire room where an operator works remotely, detects fraud, and notifies a manager.

[0008] The above and other objects and novel features of the present invention will become apparent from the description of this specification and the accompanying drawings. [Means for solving the problem]

[0009] Among the inventions disclosed in this application, the outline of representative inventions will be briefly explained as follows.

[0010] An operator fraud detection system that is a representative embodiment of the present invention is an operator fraud detection system that monitors an operator who performs contact center operations using an operator terminal installed in a remote work environment and the work environment, and includes a camera connected to the operator terminal that captures images of the operator and the work environment, a camera monitoring unit that detects fraudulent or abnormal events in the operator or the work environment from images captured by the camera, and a warning processing unit that, when the camera monitoring unit detects the event, outputs a warning to the operator, locks the parts of the operator terminal that are related to the contact center operations, and notifies a server related to the contact center operations.

[0011] The camera monitoring unit then uses image recognition processing to identify the approximately rectangular object being held by the operator from the video image, and if successful, further uses image recognition processing to determine whether the approximately rectangular object is a smartphone, and if it is a smartphone, it is deemed to have detected the event. [Effects of the Invention]

[0012] The effects obtained by the representative inventions disclosed in this application can be briefly explained as follows.

[0013] That is, according to the representative embodiment of the present invention, it is possible to monitor the entire room where an operator is working remotely, detect any fraud, and notify the administrator. [Brief explanation of the drawings]

[0014] [Figure 1] 1 is a diagram illustrating an overview of an example of the configuration of an operator fraud detection system according to an embodiment of the present invention. [Figure 2] 1(a) to 1(c) are diagrams showing an outline of a method for detecting fraud and abnormalities in one embodiment of the present invention. [Figure 3]FIG. 10 is a diagram illustrating an example in which a substantially rectangular object held by an operator is identified according to an embodiment of the present invention. [Figure 4] FIG. 10 is a diagram outlining an example of a warning screen displayed when fraud or an abnormality is detected by an operator terminal according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0015] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In all drawings used to explain the embodiments, the same parts are generally designated by the same reference numerals, and repeated explanations will be omitted. However, parts that have been designated and explained in one drawing may be referred to by the same reference numerals in the explanation of other drawings, although they will not be shown again.

[0016] <System configuration> 1 is a diagram showing an overview of an example of the configuration of an operator fraud detection system according to one embodiment of the present invention. The operator fraud detection system 1 has a configuration in which, for example, a CC server 10 consisting of a server device that provides various functions related to CC work, and operator terminals 20 consisting of information processing terminals used by each operator 2 who performs CC work remotely while working from home, are mutually connected by a VPN (Virtual Private Network) or the like via a network 30 such as the Internet or a dedicated line.

[0017] The CC server 10 is composed of, for example, server equipment or a virtual server built on a cloud computing service, and realizes various functions related to CC operations by using a CPU (Central Processing Unit) not shown to execute middleware such as an OS (Operating System), DBMS (DataBase Management System), and Web server program that are expanded onto memory from a storage device such as an HDD (Hard Disk Drive), as well as software that runs on top of them.

[0018] This CC server 10 has various units, such as a CC processing unit 11 and an operator management unit 12, implemented as software. The CC processing unit 11 has functions to execute and support CC operations, such as accepting inquiries from customers by telephone or the like, assigning the inquiries to available operators, recording the details of the responses as a history, retaining and appropriately providing information required for the operators to respond to customers, and managing the availability of operators. Since this embodiment is not directly related to the content of CC operations themselves, the content of CC operations will not be described in further detail.

[0019] The operator management unit 12 has a function of responding to the operator 2 when fraud or anomaly is detected in the operator terminal 20 (described later) and a notification to that effect is received. The content of the function here is not particularly limited, and appropriate functions can be implemented depending on the operational design. For example, various functions can be implemented, such as acquiring and saving image data of a screenshot of the screen of the operator terminal 20 at the time fraud or anomaly is detected, or image data captured by the fisheye camera 23 (described later), communicating with the target operator 2 and the operator terminal 20, notifying an administrator or supervisor (hereinafter sometimes referred to as "SV"), unlocking the operator terminal 20, etc.

[0020] The operator terminal 20 is configured, for example, as an information processing terminal such as a PC (Personal Computer) or a tablet terminal, and realizes various functions for the operator 2 to perform CC work by executing middleware such as an OS expanded onto memory from a recording device such as an HDD or SSD (Solid State Drive) using a CPU (not shown), and software running on that.

[0021] This operator terminal 20 has various units, such as a CC client 21, an operator monitoring unit 22, and an operation monitoring unit 24, which are implemented as software. The operator terminal 20 also has a fisheye camera 23 connected via a USB (Universal Serial Bus) or the like. The fisheye camera 23 can be, for example, a commercially available ultra-wide-angle camera with a viewing angle of 170 degrees or 180 degrees, as appropriate, and is installed, for example, above the display of the operator terminal 20, in a position that can capture the front, including the face of the operator 2, and the entire or most of the work environment in the background where the operator 2 performs his or her work (a room where the operator 2 performs CC work from home).

[0022] The fisheye camera 23 is not limited to one equipped with a fisheye lens or an ultra-wide-angle lens, and a camera equipped with a wide-angle lens with a wider viewing angle than a standard lens can be used as appropriate. The expression "wide viewing angle" can also be rephrased as "wide angle of view" or "short focal length." Furthermore, the fisheye camera 23 is not limited to one with an interchangeable lens, and may be a camera with a fixed lens.

[0023] The CC client 21 has a function of providing the operator 2 with various functions and screens required to perform CC operations. Internal processing is performed by the CC processing unit 11 of the CC server 10, and the CC client 21 is implemented as a thin client that communicates with the CC processing unit 11 and only displays the screen related to the processing results and accepts input, so that data such as customer information related to CC operations is not stored on the operator terminal 20 side.

[0024] The operator monitoring unit 22 has the function of monitoring the operator 2 and the operator 2's work environment based on the video captured by the fisheye camera 23, and detecting any fraudulent or abnormal events. By using the fisheye camera 23, it is possible to monitor the operator 2 and his / her movements within the work environment without any blind spots in the work environment. Furthermore, the presence of the fisheye camera 23 acts as a deterrent, which is expected to prevent the operator 2 from committing fraudulent acts. The operator monitoring unit 22 further includes various units such as a camera monitoring unit 221 and a warning processing unit 222. The operator monitoring unit 22 can also be implemented as a thin client.

[0025] The camera monitoring unit 221 has a function of acquiring video data captured by the fisheye camera 23, analyzing it using AI (Artificial Intelligence), and detecting fraud and abnormalities. There are no particular limitations on the AI ​​engine, and any generally available one can be used as appropriate.

[0026] In this embodiment, the AI-based method for detecting fraud and abnormalities includes the following functions: (1) detecting a smartphone, (2) detecting the intrusion of another person into the work environment, and (3) identifying the face of the operator. The functions (1) and (2) prevent information leakage due to surreptitious photography or peeping of the screen displayed on the display of the operator terminal 20 by the CC client 21. Furthermore, the function (3) prevents operation of the operator terminal 20 by someone other than the operator 2.

[0027] In other words, these methods do not primarily detect actual illegal acts such as voyeurism, but rather detect illegal and abnormal events that should be prohibited in order to prevent such illegal acts from occurring. The methods (1) to (3) above are mutually independent, and it is not necessary to have all of them; it is sufficient to have at least one of them.

[0028] FIG. 2 is a diagram outlining a fraud and anomaly detection method according to an embodiment of the present invention. FIG. 2(a) is a diagram outlining an example of smartphone detection described above in (1). The circular diagram schematically shows an image of the operator 2 and the work environment captured by the fisheye camera 23. Due to the characteristics of the fisheye camera 23, the peripheral portion of the circular image (the shaded portion in the example of FIG. 2 (e.g., the walls, ceiling, desk, etc. of the work environment)) is distorted along the circumference. This diagram shows the situation in which the operator 2, shown in the center, has taken out his smartphone and is about to capture an image of the operator terminal 20.

[0029] The camera monitoring unit 221 acquires image data captured by the fisheye camera 23 and constantly (more precisely, at regular time intervals) determines whether a smartphone is in the image using AI image recognition. In the example of FIG. 2(a), the smartphone is enclosed in a rectangle, indicating a state in which the smartphone has been recognized. Note that the rectangle is used for ease of explanation; in reality, it is sufficient if the smartphone can be identified internally. Also, for ease of explanation, the smartphone is shown with its front facing the fisheye camera 23, but considering the general behavior of taking a photo with a smartphone, it is considered more common for the smartphone to be facing its back.

[0030] Regarding the image recognition of smartphones by AI, a learning model may be constructed to directly identify smartphones from circular images with distorted edges captured by fisheye camera 23. However, in this embodiment, a two-stage image recognition is performed in which an approximately rectangular object held by operator 2 in his / her hand is first identified from the image captured by fisheye camera 23, and then it is determined whether the identified approximately rectangular object is a smartphone, and a learning model is constructed for each stage. This makes it possible to more accurately detect fraudulent attempts by operator 2 to take pictures of the screen of operator terminal 20 while holding a smartphone.

[0031] Here, the configuration may further include a function for image recognition of privacy objects from the viewpoint of privacy as an additional function of image recognition by AI. For example, the operator's face, body, the interior of the operator's home, etc. correspond to privacy objects. For privacy objects identified by this additional function, an additional configuration may be implemented in which image processing such as mosaic or blurring is performed to make the privacy objects invisible, and images and videos are saved.

[0032] Instead of the configuration described above that performs image processing to recognize privacy objects and make them invisible, it may be possible to use image recognition by AI to identify unauthorized objects such as smartphones and third parties other than Operator 2, and perform image processing to make all objects other than these unauthorized objects invisible.

[0033] In this embodiment, when recording images or videos that have undergone image processing to make them partially inaccessible as described above, the original images or videos that have not undergone image processing to make them partially inaccessible may also be recorded. For example, the former images or videos may be made viewable by an administrator while protecting privacy, while the latter images or videos may not be normally viewable. Even if the images or videos are normally inaccessible, it is expected that the inaccessible images or videos may be extracted and viewed in the event of an incident or when an external investigative agency conducts an official investigation.

[0034] Fig. 3 is a diagram outlining an example of identifying a substantially rectangular object held by operator 2 in one embodiment of the present invention. In the example of Fig. 3, a take-out coffee cup is identified as the substantially rectangular object held by operator 2 in his / her hand through first-stage image recognition, and a dashed rectangle indicates that the object is determined not to be a smartphone through second-stage image recognition. The substantially rectangular object held (held) by operator 2 is not necessarily a smartphone, and is often other items as shown in the example of Fig. 3. Therefore, by performing two-stage image recognition, the accuracy of identifying smartphones can be improved.

[0035] Returning to Fig. 2, Fig. 2(b) is a diagram outlining an example of the detection of the intrusion of another person into the work environment (2) above. This shows a situation in which another person has intruded behind the operator 2, and there is a risk that the screen of the operator terminal 20 may be peeked at.

[0036] The camera monitoring unit 221 uses AI image recognition to determine whether or not a person other than the operator 2 is present in the image data captured by the fisheye camera 23. In the example of FIG. 2(b), a person who has entered behind the operator 2 is surrounded by a rectangle, indicating that the person has been identified as a different person. As shown in the figure, the person behind the operator 2 is often captured on the periphery of the circular image captured by the fisheye camera 23, and is therefore likely to be a distorted image. Therefore, for example, data of a person captured in a distorted state by a fisheye camera or the like may be used as training data for the AI ​​to learn in advance, thereby improving the accuracy of image recognition. Alternatively, the camera monitoring unit 221 may correct the distortion of the image on the periphery by image processing prior to image recognition by the AI, thereby removing as much distortion as possible.

[0037] 2(c) is a diagram outlining an example of face authentication of the operator himself / herself (3) above. This shows a situation in which the person operating the operator terminal 20 is a different person from the original operator 2 due to impersonation or the like.

[0038] The camera monitoring unit 221 identifies the face of the person operating the operator terminal 20 by AI image recognition of the image data captured by the fisheye camera 23. In the example of FIG. 2(c), the part identified as the face of the person operating the operator terminal 20 is surrounded by a rectangle, but this person is different from the operator 2 in the examples of FIGS. 2(a) and 2(b). The camera monitoring unit 221 then uses AI to determine whether the identified face matches the face of the operator 2 registered in advance, that is, whether a person different from the operator 2 registered in advance is operating the operator terminal 20. The face identification of the operator 2 may be implemented as a face authentication function, and various functions including the CC client 21 may be accessed if face authentication is successful.

[0039] Returning to FIG. 1, the warning processing unit 222 has a function of outputting a warning to the operator 2 when fraud or an abnormality is detected by the camera monitoring unit 221, and locking at least the screens and functions related to the CC client 21 so that they cannot be operated or used. The locking method is not particularly limited as long as it makes the screen related to the CC client 21 invisible. It is also possible to lock not only the screens and functions related to the CC client 21, but also the entire operator terminal 20. In this embodiment, for example, a warning screen such as that shown in the example of FIG. 4 is always displayed in the foreground on the display of the operator terminal 20, hiding or erasing the screen displayed by the CC client 21, and disabling operation of the operator terminal 20.

[0040] The warning processing unit 222 locks the operator terminal 20 and notifies the operator management unit 12 of the CC server 10 that fraud or an abnormality has been detected. Upon receiving the notification, the operator management unit 12 notifies the SV and instructs the warning processing unit 222 to release the lock based on an instruction from the SV. In other words, the operator terminal 20 cannot be operated unless instructed by the SV. To enable the SV to determine whether or not to release the lock, the operator management unit 12 of the CC server 10 may be able to communicate with the operator 2 by voice call, chat, etc. via the warning processing unit 222 of the target operator terminal 20.

[0041] Furthermore, the image or video that served as the basis for detecting fraud or anomaly (the image or video that AI has determined to be fraud or anomaly), or image data of a screenshot of the screen of the operator terminal 20 at the time fraud or anomaly was detected, may be configured to be output to the screen of the terminal used by the SV as a real-time image or video of the current operator 2, as needed, and the SV may refer to that image or video to determine whether or not to unlock. Here, detection has been described as being performed using AI, but for each of the AI ​​models for (1) smartphone detection, (2) detection of intrusion of another person into the work environment, and (3) facial recognition of the operator himself, an AI model may be set and operated for each detection sensitivity level.

[0042] For example, (1) high-sensitivity, medium-sensitivity, and low-sensitivity AI models may be constructed as AI models for smartphone detection, and if the SV determines that the high sensitivity causes many false positives when the high sensitivity AI model is used by default, the configuration may be such that the setting can be changed from the high-sensitivity model to the medium-sensitivity model. This change in the AI ​​model setting may be targeted at all operators 2, or may be possible for each operator 2 being monitored. While a specific example of (1) above has been explained here, the same applies to (2) and (3) above.

[0043] Furthermore, when the SV unlocks the smartphone, the SV may specify a time for unlocking the smartphone, and the anomaly detection / fraud detection function may be turned off for the specified time, allowing the target operator 2 to use the smartphone within the scope necessary for work during that time. However, while the smartphone is unlocked, it is desirable to photograph the operator 2 using, for example, the fisheye camera 23, and store the video image on the CC server 10 side, or to continue to output the video image to the screen of the terminal used by the SV.

[0044] The operation monitoring unit 24 has a function of monitoring the operation status of the process of the operator monitoring unit 22, and restarting the process when it is detected that the process has crashed or is about to crash. This makes it possible to increase availability by preventing, as much as possible, a situation in which the operator monitoring unit 22, which implements a central and important function in this embodiment, is not running and therefore fraud or abnormality cannot be detected.

[0045] As described above, according to the operator fraud detection system 1, which is one embodiment of the present invention, the work environment of the operator 2 is monitored without any blind spots based on the images captured by the fisheye camera 23, and by using AI image recognition to (1) detect smartphones, (2) detect the intrusion of another person into the work environment, and (3) recognize the face of the operator himself, fraud and abnormal events that could lead to fraudulent acts such as voyeurism, peeping, and impersonation by the operator 2 or a third party can be detected in advance, and these fraudulent acts can be prevented.

[0046] In the above explanation, the contact center operator 2 and SV are used as examples, but the present invention can also be applied to a subordinate (user) and a superior (usage monitor) working from home. In that case, instead of the CC client 21 described in this embodiment, the operator terminal 20 is equipped with a general-purpose thin client function, and the server (CC server 10) is also configured to have a processing unit that provides the thin client function instead of the CC processing unit 11. In other words, in a typical thin client system, the server side is equipped with the operator management unit 12, and the client side equipped with the fisheye camera 23 is equipped with the operator monitoring unit 22. Here, as mentioned above, the operator monitoring unit 22 itself may also be configured as a thin client, with its functions transferred to the server side.

[0047] Furthermore, in the above explanation, the detection target is a smartphone, but there are no particular limitations on the detection target, and it is also possible to detect multiple electronic devices with one AI model or multiple AI models, as long as the electronic devices are the detection target of a company or the like that uses the operator fraud detection system 1 of this embodiment, such as a mobile phone, tablet terminal, PDA (Personal Digital Assistant), camera, etc. Among electronic devices, it is also possible to detect imaging devices with a function to take pictures or images, such as a smartphone or camera.

[0048] In the above description, the detection target in the target image is identified as a third party other than the smartphone or operator 2 through image recognition, and fraud is detected. However, fraud may be detected when the detection target is identified through image recognition in multiple consecutive target images, rather than when fraud is detected based on only one target image. For example, if one target image is captured per second, five target images are captured in five seconds, and fraud is detected when the detection target is identified in all five consecutive images. On the other hand, if the detection target is identified in only the first image and cannot be identified in the remaining four images, fraud is not detected. This avoids false detection of the detection target. Note that, while the imaging interval (also referred to as the shooting interval or shooting interval) is exemplified as one image per second, this is not limiting. For example, one frame out of a predetermined FPS (frames per second) when capturing video may be used as the target image. Furthermore, if the detection target is identified or fraud is detected, the imaging interval may be shortened from normal.

[0049] Furthermore, in the above explanation, it has been explained that images and videos in which fraud is detected can be recorded by the operator fraud detection system 1 of this embodiment and confirmed by the SV, etc., but it is also assumed that if communication becomes unavailable between the operator terminal 20 or the CC server 10, it will not be possible to send images or videos in which fraud is detected from the operator terminal 20 to the CC server 10. In this case, for example, the configuration may be such that images or videos in which fraud is detected are continuously recorded by the operator terminal 20 for a predetermined time or until communication is restored, and the recorded images or videos in which fraud is detected are sent to the CC server 10 after communication is restored.

[0050] Furthermore, in the above explanation, it has been assumed that the fraud detection method, imaging interval, image processing to make security objects invisible, etc. operate with the same settings for all operators 2, but it is also possible to change the settings for each operator 2 or each department to which the operator 2 belongs. For example, when the operator monitoring unit 22 of the operator terminal 20 is started, the settings for the target operator 2 or the target department may be read from the CC server 10 and executed.

[0051] Furthermore, in the above description, it has been assumed that fraud detection according to this embodiment is performed regardless of the location of the operator 2. However, the fraud detection operation may be performed after determining whether to perform the fraud detection operation based on the location of the operator 2. For example, the fraud detection operation may not be performed if the operator 2 is at a predetermined location, and may be performed otherwise. Conversely, the fraud detection operation may be performed if the operator 2 is not at a predetermined location, and may not be performed otherwise. Examples of locations where fraud detection operation is not performed include when the operator 2 is at a location with a high security level within the company or when the operator is stationed at a customer's site. There are various well-known and commonly used technologies for identifying the location of the operator terminal 20, and these technologies can be used to achieve this. For example, there are a method using a GPS sensor in the operator terminal 20, a method estimating the location from the IP address assigned to the operator terminal 20, and a method of considering the location of an access point connected to the operator terminal 20 as the location.

[0052] Furthermore, in the above description, the operator terminal 20 has been described as a thin client terminal, but it may also be a non-thin client terminal (FAT terminal). Furthermore, depending on the type of thin client terminal, the local environment of the thin client terminal can still operate even if communication is interrupted, and it is possible to install and run programs in the local environment as appropriate. In addition, it is also possible for the programs in the local environment and the programs in the thin client to run in cooperation with each other. In other words, it is also possible to implement all or part of the means related to the operator fraud detection system 1 of this embodiment in these programs.

[0053] The invention made by the inventor has been specifically described above based on the embodiments, but it goes without saying that the present invention is not limited to the above embodiments and can be modified in various ways without departing from the spirit of the invention. Furthermore, the above embodiments have been described in detail to clearly explain the present invention, and the present invention is not necessarily limited to those having all of the described configurations. Furthermore, it is possible to add, delete, or replace part of the configuration of the above embodiments with other configurations.

[0054] Furthermore, the above-described configurations, functions, processing units, processing means, etc. may be partially or entirely implemented in hardware, for example, by designing them as integrated circuits. The above-described configurations, functions, etc. may also be implemented in software, with a processor interpreting and executing a program that implements each function. Information such as the programs, tables, and files that implement each function can be stored in a storage device such as a memory, hard disk, or SSD, or in a storage medium such as an IC card, SD card, or DVD.

[0055] In addition, in the above figures, the control lines and information lines shown are those that are considered necessary for explanation, and do not necessarily show all the control lines and information lines that are actually implemented. In reality, it can be assumed that almost all components are interconnected.

[0056] Furthermore, the concept of "image" includes both still images and video, and video can also be a video format consisting of multiple still images. The AI ​​model may be configured to input a still image of the target and detect fraud or anomalies, or it may be configured to input a video image and detect fraud or anomalies. [Industrial Applicability]

[0057] The present invention can be used in an operator fraud detection system that detects fraud committed by an operator who performs work remotely. [Explanation of symbols]

[0058] 1...Operator fraud detection system, 2...Operator, 10...CC server, 11...CC processing unit, 12...operator management unit, 20...Operator terminal, 21...CC client, 22...Operator monitoring unit, 23...Fisheye camera, 24...Operation monitoring unit, 30...Network, 221...camera monitoring unit, 222...warning processing unit

Claims

1. An operator fraud detection system that monitors an operator who performs contact center operations using an operator terminal installed in a remote work environment and the work environment, a camera connected to the operator terminal for capturing images of the operator and the work environment; a camera monitoring unit that detects irregularities or abnormalities in the operator or the work environment from images captured by the camera; a warning processing unit that, when the event is detected by the camera monitoring unit, outputs a warning to the operator, locks a portion of the operator terminal related to the contact center operation, and notifies a server related to the contact center operation, The camera monitoring unit An operator fraud detection system that uses image recognition processing to identify the approximately rectangular object held by the operator from the image, and if the object is identified, further uses image recognition processing to determine whether the approximately rectangular object is a smartphone, and if it is a smartphone, it considers the event to have been detected.

2. The operator fraud detection system according to claim 1, The camera monitoring unit An operator fraud detection system that identifies people other than the operator present in the work environment from the image through image recognition processing, and determines that the event has been detected if the people are identified.

3. The operator fraud detection system according to claim 1, The camera monitoring unit The operator fraud detection system identifies the operator's face from the image using image recognition processing, and if the face is identified, determines whether it is the same person as the image of the operator previously registered by the operator, and if it is not the same person, it determines that the event has been detected.

4. In the operator fraud detection system according to any one of claims 1 to 3, The camera monitoring unit An operator fraud detection system that performs image recognition processing on the image after correcting distortion in the peripheral areas of the image.

5. In the operator fraud detection system according to any one of claims 1 to 3, The warning processing unit When the event is detected by the camera monitoring unit, a screenshot of the screen on the operator terminal including the portion related to the contact center operations is acquired and transmitted to the server.

6. In the operator fraud detection system according to any one of claims 1 to 3, The warning processing unit When the event is detected by the camera monitoring unit, the operator fraud detection system acquires a screen showing the fraudulent or abnormal event detected on the operator terminal and transmits the screen to the server.

7. In the operator fraud detection system according to any one of claims 1 to 3, The camera monitoring unit An operator fraud detection system that identifies privacy objects from the image by image recognition processing and performs image processing to make the identified privacy objects invisible.

8. In the operator fraud detection system according to any one of claims 1 to 3, The camera monitoring unit An operator fraud detection system that performs image processing to make it impossible to see anyone other than the smartphone or operator identified by image recognition processing in the image.

9. In the operator fraud detection system according to any one of claims 1 to 3, The camera monitoring unit An operator fraud detection system that changes the interval at which images are taken by the camera when a person other than the smartphone or the operator is identified in the image through image recognition processing.

10. In the operator fraud detection system according to any one of claims 1 to 3, The camera monitoring unit An operator fraud detection system that determines that the event has been detected if a person other than a smartphone or an operator is identified through image recognition processing in each of multiple consecutive images taken for a predetermined period of time or more.

11. A fraud detection system that monitors a user who performs work using a terminal installed in a remote work environment and the work environment, a camera connected to the terminal for capturing images of the user and the work environment; a camera monitoring unit that detects fraudulent or abnormal events occurring in the user or the work environment from images captured by the camera; a warning processing unit that, when the event is detected by the camera monitoring unit, outputs a warning to the user, locks the terminal, and notifies a server related to the business; The camera monitoring unit A fraud detection system that uses image recognition processing to identify the approximately rectangular object held by the user from the image, and if the object is identified, further uses image recognition processing to determine whether the approximately rectangular object is a specified electronic device, and if it is the specified electronic device, it considers the event to have been detected.

Images