KR101148883B1 - Automated teller machine preventing illegal use and method for preventing illegal use in automated teller machine - Google Patents

Abstract

PURPOSE: An automatic teller machine and an illegal use detection method thereof are provided to prevent illegal use by detecting disguised user faces. CONSTITUTION: A camera(10) is arranged on the front side of an automatic teller machine in order to record images of user faces. A storage device(14) permanently stores the images recorded by the camera. A card reader device(20) comprises a card slot for using financial services. The card reader device reads various kinds of information stored in an inserted card when the card is inserted into the card slot. A disguised face detection device(60) determines a disguised user face by analyzing the image of the user face. The disguised face detection device acquires an image which is recorded right before predetermined time from an insertion time point of the card using a real-time buffer list.

Description

Financial automated device with fraud determination function and method for fraud determination of financial automation device {AUTOMATED TELLER MACHINE PREVENTING ILLEGAL USE AND METHOD FOR PREVENTING ILLEGAL USE IN AUTOMATED TELLER MACHINE}

The present invention is a financial automation device and financial automation that can prevent fraudulent use by covering the face by determining whether the user is disguising the face using an image of the front face of the user who uses the financial automation device. The present invention relates to a method for determining illegal use of a device.

Recently, in order to provide convenient financial services to customers, automated teller machines (ATMs) have been widely installed and used, and cash withdrawals, integrated account management, payment of e-commerce bills, electronic authentication, and utility bills are made by such ATMs. Financial services include electronic notices and payments (EBPP) and interpersonal remittances.

On the other hand, ATMs are freely available at any time, and there are no administrators to monitor their use, so fraudulent users disguise their faces and use other people's information (e.g., cash cards, credit cards, bank accounts, etc.) There is a problem in that it is exposed to fraudulent use, such as withdrawal.

Therefore, in order to prevent an unauthorized user from covering their face or covering the camera and using an ATM, a technique for determining a face's identification possibility from a captured image has been a hot topic, and various methods have been proposed according to such needs. .

However, the conventional methods are proposed under the assumption that the camera installed in the ATM can always acquire the front face of the user, and does not explain how to non-forceively acquire the front face image of the user. In addition, there may be a case where a user without criminal intention is identified as a face mask and cannot use the ATM due to the wrong judgment of the system, and in this case, there is no clear solution on how to prepare and handle the ATM.

On the other hand, as an example of a method for detecting the occlusion of the user, Japanese Patent Application No. Hei 4-76690 measures the distance between the camera and the user through a sensor and asks the user to approach the user if the distance is too long. Disclosed is a method of determining whether a user's face is normal by taking an image when the distance is normal. The normality is determined by comparing with a pre-stored criminal list, by determining whether the face is masked using a mask or sunglasses, or by covering the camera with a hand. If there is an abnormality in the identification result of normality, it may be requested to remove the obscured object, or the bank tells the bank to accept the financial request again. Doing.

Another solution is Japanese Patent No. Hei 5-35992, which acquires a user's face image, extracts eye, mouth, ear and nose features and weights them according to the detection level to determine the degree of agreement with the suspected person. do. If it is determined that the person is suspicious, the administrator is notified and the video recording method is disclosed.

Meanwhile, Korean Patent Application No. 1998-82154 describes a method of using skin color information and background image information to extract feature points such as eyes, nose, and mouth from a face image of a user photographed from a surveillance camera attached to an ATM. Doing.

In addition, Korean Patent No. 10-293897 describes a method of determining whether a face is covered by binarizing color information obtained from a face image photographed at an ATM.

Furthermore, in Korean Patent No. 10-473598, when a face is detected in an image taken by an ATM, it is divided into two upper and lower regions, and each divided region is applied to an eigen vector extracted through a principal component analysis technique from a plurality of training images. A method of detecting whether there is sunglasses or a mask on the face by projecting is disclosed.

Lastly, Korean Registration No. 20-393236 discloses a method of determining whether a face is covered by mounting an infrared sensor in an ATM and measuring a user's face temperature.

As described above, the related arts assume that the front face of the user is accurately photographed, and only disclose the process of using the photographed front face, and do not consider a method for actually obtaining the front face.

Therefore, the front face may not be photographed depending on the position of the camera, and the user may not acquire the front face by turning his head by mistake. Therefore, a countermeasure is required.

The present invention is to solve this problem, it is to be able to quickly and accurately determine whether the face of the user who wants to use the financial services through ATM to accurately determine whether the illegal use of the financial automation device .

In addition, the front face image of the user is secured by a non-coercive method, the possibility of face identification is determined by the acquired face front image, and if the acquired face image is identifiable, the use of ATM is allowed; When no identifiable face is found among a plurality of acquired face images, the user is required to gradually cooperate with frontal face photographing to improve the face identification possibility of the user.

A financial automation device having a fraud determination function through a real-time buffering process according to the present invention for achieving the above object, the card reading to read the card information stored in the card when the card for the execution of the financial service is inserted Way; Operation means operated by the user for a business request operation for the user to request processing of the financial service and a password input for the card information; Financial service processing means for processing a financial service according to the business request operation when the password specified for the card information and the password input through the operation means are the same; A camera that is disposed toward the front to continuously photograph the face of the user who wants to use the financial automation device, and continuously outputs an image of the front; A real time buffer list for temporarily storing the video for a predetermined time by storing the video continuously output from the camera and deleting the stored video according to the stored order; and when the card is inserted into the card reading means, the real time An image acquisition unit for acquiring an image photographed immediately before a predetermined time from the insertion point of the card from a buffer list, a face analyzer for analyzing a user's face from the acquired image and determining whether the user's face can be normally identified; Instructs the financial service according to the business request operation to be normally processed when the face of the user is normally identified, and process the financial service according to the business request operation to the financial service processing means when the face of the user is not normally identified. Control command to instruct the user not to Control command output section for outputting comprises a means for determining the stomach comprising a.

In addition, the camera is characterized in that installed in the position close to the inlet for the user to insert the card.

The real-time buffer list may be configured to temporarily store an arbitrary multiple-th frame of consecutive frames of the video output from the camera.

The image acquisition unit may acquire a plurality of images photographed and temporarily stored during a predetermined time period before and after a predetermined time from the time point of insertion of the card from the real-time buffer list, and the face analyzer may include the plurality of images. It is determined whether face identification is possible for each, and the control command output unit instructs the financial service processing unit to normally process the financial service according to the business request operation when there is at least one image capable of identifying the user's face. And outputting a control command.

The face analyzer may determine that the face and the at least one eye in the user's face are normally identifiable faces.

The face analyzer may identify eyes and a mouth from a face of a user, and determine that the face is normally identifiable when it is determined that a positional relationship between the identified eyes and the mouth is an inverted triangle.

In addition, a method of determining fraudulent use of a financial automation device using a real-time buffer list in which a face of a user who wants to use the financial automation device is continuously photographed using a camera installed near a card slot and temporarily stored for a predetermined time. The method may include: (a) sequentially receiving an image of a front photographed and temporarily storing the image in the real-time buffer list; (b) when a card for executing a financial service is inserted, acquiring an image photographed from the real-time buffer list immediately before a predetermined time from when the card is inserted; (c) analyzing the face of the user from the obtained image and determining whether the user's face can be normally identified; (d) If the user's face is normally identifiable, control to process the financial service requested by the user, and if the user's face is not normally identifiable, the user determines that the user requested the financial Controlling to deny processing of the service.

Further, in the step (b), a plurality of images captured and temporarily stored for a predetermined time period before and after a predetermined time from the point of time of the card insertion are obtained from the real-time buffer list, and in the step (c), the The apparatus may determine whether face identification is possible with respect to each of the plurality of images, and determine that the user face identification is possible when at least one image capable of identifying the face of the user exists among the plurality of images.

According to the present invention having the above-described configuration, it is possible to accurately and quickly determine the case where the illegal user uses the financial automation device while disguising his / her face, thereby preventing financial service related accidents.

In addition, according to the present invention, by setting the camera position and the image acquisition time of the financial automation device based on the behavior characteristics of the user, it is possible to acquire the front face of the user non-compulsively, it is inconvenient for the face shooting to the user Can be reduced.

In addition, in executing the face discriminant determination algorithm, the non-coercive one-step verification, the more coercive two-step verification, and the very coercive three-step verification are executed sequentially, so that users who can easily identify the face are allowed to accept transactions early. Even if it is difficult for the user to identify, the transaction can be performed through two or three steps. In this way, even if the face masking judgment fails, the transaction is not immediately blocked, and thus, the erroneous blocking of the financial service due to user's carelessness or a system error can be reduced.

1 is a diagram schematically illustrating a financial automation device having a fraud determination function according to an embodiment of the present invention.
2 is a diagram schematically illustrating an external appearance of a financial automation device according to an embodiment of the present invention.
3 is a view showing images taken by a camera in chronological order.
Fig. 4 is a block diagram showing the configuration of the camouflage determination means in the financial automation device equipped with the illegal use determination function according to the embodiment of the present invention.
5 is a flowchart illustrating a procedure of a method of determining illegal use according to an embodiment of the present invention.
FIG. 6 is a diagram illustrating a face image acquisition situation using the most coercive image acquisition method in order to photograph the front face of the user at an appropriate size.
7 is a flowchart illustrating in detail a valid face detection and identification possibility determination step in the method for determining the use of fraud according to an embodiment of the present invention.

Hereinafter, with reference to the accompanying drawings will be described in detail the present invention.

1 is a diagram schematically illustrating a financial automation device having a fraud determination function through a real-time buffering process according to an embodiment of the present invention. Referring to Fig. 1, a financial automation device having a fraud determination function according to the present invention includes a camera 10, a storage means 14, a card reading means 20, an operation means 30, The financial service processing means 32 and the disguise determination means 60.

The camera 10 is disposed forward to photograph a face of a user who wants to use the financial automation device. The camera 10 continuously photographs the front and continuously captures the captured image regardless of whether the user uses the financial automation device. Will output Of course, the camera 10 may be arranged to photograph the front (ie, the user) only when the user detects a human body (not shown) when the user approaches the financial automation device. On the other hand, the position of the camera can be set near the card slot for the user to insert the card to use the financial services, which will be described later.

The storage means 14 is for permanently storing the front image captured by the camera 10 and may include a hard disk, a flash memory, a temporary storage buffer, and the like. The storage means 14 may include a function of transmitting a captured image stored in connection with a central control center through a network.

The card reading means 20 includes a card inlet through which a card for the execution of a financial service can be inserted, and when the card is inserted into the card inlet, the card reading means 20 reads various kinds of information stored in the inserted card. The card reading means 20 may include not only a method of inserting the card into the card slot, but also a method of bringing the surface of the card into contact with the card reader or slide. In addition, the card reading means 20 may not only read the card, but also read information from the magnetic wire or IC chip attached to the bankbook, and furthermore, it is possible to read fingerprints, irises, mobile phones, IC cards, barcodes, and the like. Various means for reading the information may be included.

The operation means 30 is used to input a request (business request operation) for a financial service that the user wants to be provided, or input a password or the like for the inserted card. The operation means may include a keypad on which numbers or symbols are displayed, a touch screen 31 including a display function (see FIG. 2), and the like.

The financial service processing means 32 transmits card information read by the card reading means 20 and various information input by the user through the operation means 30 to a financial service providing server (not shown) via a network. After being notified of the user authentication result, the user provides or rejects the financial service requested by the user.

Meanwhile, the financial service processing means 32 may receive a control command from the outside, and the control command may control whether the financial service requested by the user using the financial automation device is provided.

The camouflage determination means 60 analyzes the face of the user who has approached to use the financial automation device to determine whether the face is obscured to an unidentifiable degree of who the user is. To this end, the camouflage determination means 60 temporarily stores an image continuously photographed and output by the camera 10, sequentially deletes the stored order from the oldest among the temporarily stored images, and newly outputs the image in the deleted space. By storing the, the real-time buffer list 62 that can temporarily store the image taken by the camera 10 for a predetermined time.

The camouflage determination means 60 uses this real-time buffer list 62, and when a card is inserted into the card insertion port of the card reading means 20, the predetermined time from the point of time of the card insertion from the real-time buffer list 62. Acquire images taken immediately before. In the obtained image, the user's face is extracted, and it is determined whether each part of the extracted user's face is normally identifiable, and if the user's face is normally identifiable, a control command for controlling the financial service requested by the user. If the user's face is so hidden that it is not normally identified, the control unit determines that the financial automation device is used improperly and generates and outputs a control command for controlling the processing of the financial service requested by the user. . The control command is provided to the financial service processing means 32.

The financial automation device having a fraud determination function according to an embodiment of the present invention having such a configuration may determine whether or not to disguise a face by using an image photographed by a user at a time when a card is inserted. It is preferable to determine whether a face is camouflaged using an image taken just before a predetermined time. This is because the user stares at the card slot only immediately before the card is inserted into the card slot, and at the moment the card is inserted, the user looks at the operation means.

On the other hand, the camera in the financial automation device according to the present invention is installed near the card inlet of the financial automation device that the user must stare at. At this time, the position of the camera is set so that the face photographed by the hand inserting the card is not covered. In addition, in order to be easily used in the forced face image acquisition step using the face frame to be described later, the position of the camera as well as the card insertion slot as well as the monitor (for example, the operation means 31) to show the face frame as much as possible. It is preferred to be installed in close proximity.

Here, the meaning that the camera is installed in close proximity to the card slot may be understood that the camera is placed at a position where a front face can be obtained from the user when the user inserts the card into the card slot.

3 is a view showing images taken by a user in a chronological order by a camera disposed near a card slot. The upper row of FIG. 3 shows an image actually photographed by the camera. The camera has to photograph a wide range, and since the distance from the camera is relatively far from the user, all of the upper body of the user is photographed.

In FIG. 3, the images in the lower row are enlarged only in the face part of the user in each image in the upper row, and show face images used to determine whether a face is covered and identify an illegal user. Looking at each face image, it can be seen that a front face image in which the user looks directly at the camera is obtained from the image acquired immediately before the point indicated as the card insertion time point.

Fig. 4 is a block diagram showing the configuration of a camouflage determination means among financial automation devices equipped with a fraud determination function according to an embodiment of the present invention. The disguise determination means 60 for realizing the operation | movement of the financial automation device provided with the illegal use determination function as mentioned above is the real-time buffer list 62, the image acquisition part 64, and the face analysis part 66. ) And a control command output unit 68.

The real-time buffer list 62 has an effect of temporarily storing an image for a predetermined time by storing images continuously output from the camera 10 and sequentially deleting the stored images in the stored order.

That is, the real-time buffer list 62 maintains a video stream which is captured in real time through the camera 10 and continuously flows for a predetermined time, and after that time, first removes the first stored buffer video object in order. In addition, the real-time buffer list 62 provides a function of obtaining an optimal process performance and optimal images while reducing or increasing the number of images stored in the buffer list to 100 → 50 → 30 or the like. .

When the card is inserted into the card reading means 20 by the user, the image acquiring unit 64 acquires the captured image from the real time buffer list 62 just before a predetermined time from the time of inserting the card.

That is, the image acquisition unit 64 acquires an image object of a real time buffer list between 1 second and 5 seconds before the card is inserted. In the case of acquiring an image object for 5 seconds, since a still image of 8 to 15 frames per second can be obtained when the image is acquired, a total of 40 to 75 frames are obtained, and the even or odd image is obtained from the acquired image. Can only take In some cases, an image for every multiple of another integer may be selectively acquired, and the image most likely to obtain a front face is obtained by combining a leading image acquisition method or an ending image acquisition method in a specified time interval. It can also be obtained.

At this time, it is preferable to determine whether or not a card is inserted into the card reading means 20 by detecting a specific signal generated by the card reading means 20 or the financial service processing means 32. That is, when the card is inserted, the card reading means reads the information stored in the card and starts the financial work in the financial automation device, which can be regarded as the time of cart insertion. By such a configuration, it is possible not to install a separate sensing means for determining whether the card has been inserted.

The face analyzer 66 may extract the face region of the user from the image (face image) acquired by the image acquirer 64, and examine the extracted face region to uniquely identify the user. It is judged whether or not photographing is enough.

That is, it is determined whether the mouth and the at least one eye are distinguishable from the user's face, or whether the positional relationship between both eyes and the mouth identified from the user's face is an inverted triangle.

Meanwhile, the face analyzer 66 may preprocess the acquired face image by using various image processing methods to obtain desirable results in subsequent face region inspection.

More specifically, the face analyzer 66 dynamically performs eye detection and mouth detection in order to verify the validity of face identification even when a part of the face is partially hidden for the detection of a normal face. Composite shape detectors ".

The complex shape detector scheme can reduce errors in which face identification is not complete as the face is slightly obscured, thereby determining that the face cannot be identified. For example, if one eye and mouth are visible and the other eye is covered by an eye patch or hair, the face shape can be predicted by using only the visible eye to detect a valid face even though it is difficult to completely detect the user's face. And judge.

The control command output unit 68 performs normal processing of the financial service according to the business request operation input by the user through the operation means 30 when the face of the user analyzed by the face analyzer 66 is normally identified. Generate a control command and transmit it to the financial service processing means 32, and if it is determined that the user's face is not normally identifiable, the financial service according to the operation request operation of the user is processed by the financial service processing means 32; Print a control command that tells you not to.

The control command provided from the disguise determination means 60 may be input to the financial service processing means 32 through a separate input means, or may be transmitted to a financial service providing server (not shown) via a network.

In addition, the above-described configuration may be modified as follows.

That is, the image acquisition unit 64 acquires a plurality of images captured and temporarily stored for a predetermined time period immediately before a predetermined time from the card insertion time from the real-time buffer list 62, and the face analyzer 66 It is determined whether the face can be identified for each of the plurality of images, and if the control command output unit 68 includes any one image that can identify the user's face among the plurality of images, the financial service is normally processed. If the control command is output and there is no image capable of identifying the user's face, the control command may be determined to block the provision of the financial service by determining that the user's face is illegally hidden.

By determining whether a face can be identified using a plurality of images as described above, temporary flickering of illumination, removal of an object covering the face (for example, removal of hair or eye patch), and change of the direction in which the face is directed Etc., it is possible to temporarily solve the problem that the face identification is impossible and the financial service is undesired.

According to the above-described configuration, by analyzing the face using the front face, it is possible to accurately determine the case where the illegal user uses the financial automation device while disguising the face. Accidents can be prevented quickly. It also minimizes the error of unwanted financial denial of service as the face is partially obscured by the user.

Next, a method of determining a situation in which a user illegally uses the financial automation device while covering a face using the financial automation device according to an embodiment of the present invention, which can be configured as described above, will be described.

5 is a flowchart for explaining a method of determining illegal use according to an embodiment of the present invention. First, in the financial automation device as described above, the captured image is continuously stored and removed in the real-time buffer list 62 (S10). Of course, the captured image may continue to be stored in the storage means 14. In this case, when the user approaches and inputs the card to the financial automation device (S20), the real-time buffer list 62 captures the image captured immediately before the predetermined time from the card input time or after the predetermined time from the card input time or the card input time. Acquire one or more (S30), check the position (region) of the user in each of the acquired images, and detects the shape and position of at least eyes and mouth of the components of the face to determine whether the user can identify the face image S40 (first attempt). If the shape and position of the eyes and mouth coincide with the shape and position of the normal face, it is determined that the user's face of the captured image is a normal face which is not artificially masked for illegal use and that the user can be uniquely identified. In operation S41, the financial service requested by the user may be normally processed in operation S50. If, in step S41, the eyes and mouth cannot be identified from the user's face or are determined to be an abnormally hidden face, it is determined that the image of the user's face is not accurately captured to execute the requested financial service. Reserve and execute the subsequent second attempt.

In other words, if it is determined that the face is hidden, the financial automation device may output a request for cooperation to remove the attachment attached to the face or to face the camera (S42). With this request, the user's face is secondarily obtained (S43), and whether the face can be identified from the acquired image is determined (S44). If it is determined that the face image analyzed from the photographed second acquired image is identifiable (S45), the process proceeds to step S50 to allow financial transactions and otherwise executes the next third attempt.

The third attempt is a more coercive cooperation request step, in which the user places the face closer to the camera and guides the correct face to be photographed (S46). At this time, the face frame of a predetermined size is displayed on the operation means 31 of the financial automation device, and the face image photographed by the user is also displayed at the same time. The user must approach the camera closely and look straight at the camera so that his face fits in a similar size to the face frame.

When the user's face image that fits the face frame is obtained (S47), the possibility of identifying the face image is determined again (S48), and if the face is acquired to an identifiable level (S49), the user is allowed to accept the transaction. The transaction is blocked by considering the access (S52).

According to the present invention including such a procedure, by attaching a surveillance camera near the card slot that the user has to stare at least once in order to use the financial automation device, the front face of the user is displayed by using the image before or after card insertion. Can be acquired at the moment it does not recognize (first attempt), without sacrificing user convenience.

In addition, when a large number of people consider using a daily financial automation device, a large number of users may be uncomfortable if the financial service is blocked even in the case of a small error such as when a part of the user's face is covered by hair or the like. . Therefore, in the present invention, not the method of evaluating the face identification possibility at once, the first face identification possibility evaluation is performed using a plurality of images acquired before or after the card is inserted, and the existence of the face that can be identified in the first evaluation is confirmed. If not, request the user's cooperation, such as "Please remove any attachments on the face and stare at the camera". If you can't get an identifiable face in this second attempt, show a face frame on the monitor, ask the user to match your face to this frame, and then get an accurate face image. Gradual evaluation should be carried out gradually.

6 is a view for explaining a process of very coercive third-order face identification possibility evaluation using the face frame. A face frame having a rough face shape on the left is displayed on the monitor (operation means 31, etc.) as shown in the figure on the right, and the user moves the face to a position where his face is aligned with the displayed face frame. The face of the user is photographed to obtain a face image. In the camouflage determination means, the third face identification possibility evaluation is performed on the acquired face image, and finally, whether or not to provide financial service is finally determined.

FIG. 7 is a flowchart illustrating in detail a procedure of valid face detection and identification possibility determination in a method of determining illegal use according to an embodiment of the present invention, in which step S40 and step S41 of FIG. To illustrate.

First, when a plurality of images before and after the card insertion time is obtained from the real-time buffer list 62 in step S30 (S401), one of the plurality of input images is selected (S402).

When the image is selected, image processing is performed in various ways so that the user's face can be accurately identified in the selected image (S403). Thereafter, the user's face is found from the image on which the image processing is performed (S404), and the area of the face is determined (S405). Meanwhile, the face position may be determined by simply specifying the face position in the image by using the result learned from the face identification operation performed previously, and extracting the contour of the face from the specified position.

If the face area is determined, the shape of the user's eyes is detected (S406). The shape of the mouth is detected following the eyes of the user (S407). Also, the shape of the eyes and the mouth is determined based on the edge information of the portion detected as the eyes and the previously stored learning information.

If the eye is not detected or the mouth is not detected, it is determined that the current image under analysis is impossible to identify the face, and the process returns to step S402 to select another image among the input images, in step S403. From step S407, the face identification operation is continued.

On the other hand, if the eyes and mouth of the user is normally detected, the positional relationship between both eyes and the mouth is determined geometrically, and if it is determined that the shape of the inverted triangle is finally determined as a normal face (S408).

If it is determined that the face is normal, the process proceeds to step S50 to allow financial transactions by the user.

When the eyes are not detected or the mouth is not detected at step S406 and step S407, and the process returns to step S402, the face identification operation is performed on all the images taken at the image input step S401, If no normal face is found in any of the images, an enhanced face identification may be attempted such as proceeding to step S42 of the second attempt to guide the user to remove occlusion of the face, or immediately performing step S52. It may be determined that the final failure of the user's face identification.

Claims (8)

Card reading means for reading card information stored in the card when a card for executing a financial service is inserted;
Operation means operated by the user for a business request operation for the user to request processing of the financial service and a password input for the card information;
Financial service processing means for processing a financial service according to the business request operation when the password specified for the card information and the password input through the operation means are the same;
A camera that is disposed toward the front to continuously photograph the face of the user who wants to use the financial automation device, and continuously outputs an image of the front;
A real time buffer list for temporarily storing the video for a predetermined time by storing the video continuously output from the camera and deleting the stored video according to the stored order; and when the card is inserted into the card reading means, the real time An image acquisition unit for acquiring an image photographed immediately before a predetermined time from the insertion point of the card from a buffer list, a face analyzer for analyzing a user's face from the acquired image and determining whether the user's face can be normally identified; Instructs the financial service according to the business request operation to be normally processed when the face of the user is normally identified, and process the financial service according to the business request operation to the financial service processing means when the face of the user is not normally identified. Control command to instruct the user not to Output control command output by an automatic teller machine having a real-time buffer fraud detection function by a process comprising the stomach comprising a determining means for that. The method of claim 1,
The camera is a financial automation device having a fraud determination function through a real-time buffering process, characterized in that the user is installed in the position close to the inlet for inserting the card. The method of claim 1,
The real-time buffer list is
And an illegal use determination function through a real-time buffering process, characterized in that temporarily storing an integer multiple-th frame among successive frames of the image output from the camera. The method of claim 1,
The image acquisition unit,
From the real-time buffer list, a plurality of images captured and temporarily stored during an arbitrary time interval before and after a predetermined time from the time of insertion of the card are obtained.
The face analyzer determines whether a face can be identified for each of the plurality of images,
The control command output unit outputs a control command for instructing the financial service processing unit to normally process the financial service according to the business request operation when there is at least one image capable of identifying the face of the user. Financial automation equipment with fraud determination through time buffering process. 5. The method according to any one of claims 1 to 4,
The face analyzer,
A financial automation device having a fraud determination function through a real-time buffering process, wherein if a mouth and at least one eye in the face of the user are discernible, the face is determined as a normal identifiable face. 5. The method according to any one of claims 1 to 4,
The face analyzer,
Identifying eyes and mouth in the user's face, and determining that the face is normally identified when the positional relationship between the identified eyes and mouth is in the form of an inverted triangle, the fraud determination function using the real-time buffering process Financial automation device equipped with. Determining fraudulent use of a face in a financial automation device using a real-time buffer list that continuously captures a face of a user who wants to use the financial automation device using a camera installed near a card slot and temporarily stores it for a predetermined time. As a method:
(a) receiving a continuous image of the front and storing the temporary image in the real-time buffer list;
(b) when a card for executing a financial service is inserted, acquiring an image photographed from the real-time buffer list immediately before a predetermined time from when the card is inserted;
(c) analyzing the face of the user from the obtained image and determining whether the user's face can be normally identified;
(d) If the user's face is normally identifiable, control to process the financial service requested by the user, and if the user's face is not normally identifiable, the user determines that the user requested the financial And controlling to reject the processing of the service. The method of claim 7, wherein
In the step (b),
From the real-time buffer list, a plurality of images captured and temporarily stored during an arbitrary time interval before and after a predetermined time from the time of insertion of the card are obtained.
In the step (c),
It is determined whether face identification is possible for each of the plurality of images, and if there is at least one image identifying the user's face among the plurality of images, it is determined that the user's face identification is possible. How to determine fraudulent use.

Images