Monitoring device and monitoring device method

Abstract

An ATM 10 having embedded therein a function as a monitoring apparatus comprises: an imaging unit 23 that includes, in an imaging range, an operator of a transaction apparatus for handling money and the surrounding thereof; a posture detection unit 21a for detecting a joint position of a skeleton in an image of a person included in an imaging result of the imaging unit 23; a determination unit 21b that determines the motion of the person on the basis of an operator detection result of the posture detection unit 21a; and a command unit 21c that generates a command for instructing a response to the person, on the basis of a determination result of the determination unit 21b. This monitoring apparatus can realize high-function monitoring of operation of a transaction apparatus for handling money.

Description

Technical Field

[0001] The present invention relates to a monitoring device and a monitoring method for monitoring an operator of a transaction device that handles money and the surroundings thereof.

Background Art

[0002] Conventionally, there is a technique for monitoring operations using images. For example, Patent Document 1 describes "An abnormal behavior detection device used for a gaming machine operated by an operator at an operator position, the abnormal behavior detection device characterized by having the following elements: (1) an imaging unit that acquires image information including a subject at the operator position; (2) a subject motion detection unit that detects subject motion information indicating the motion of the subject based on the image information acquired by the imaging unit; (3) an abnormal behavior determination unit that determines the occurrence of abnormal behavior by the operator based on the subject motion information detected by the subject motion detection unit."

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, the conventional technology has not been able to sufficiently take measures against fraud and assist operations for a transaction device that handles money. Fraud against a transaction device that handles money is diverse, such as peeping from behind the operator, attaching a device for fraud, and acts of destruction. In addition, it is also required to provide support such as operation guidance and fraud prevention in real time when the operator is confused in the operation or trying to perform a transfer operation due to transfer fraud. Furthermore, in a trading device, there may be people waiting in line around the operator. Unlike passersby, these people stay near the operator, which can make it difficult to distinguish them from the operator. The technology described in Patent Document 1, for example, cannot identify and respond to a wide variety of fraudulent activities, nor can it provide support tailored to the operator's condition. For these reasons, monitoring of trading devices has historically been limited to simply capturing and recording images for subsequent investigations.

[0005] Therefore, the present invention aims to provide a highly functional monitoring device and monitoring method that can implement measures against fraud and provide operational support for the operation of currency handling devices. [Means for solving the problem]

[0006] To achieve the above objective, one representative monitoring device of the present invention comprises: an imaging unit whose imaging range includes an operator of a currency handling device and the area around the operator; a posture detection unit that detects joint positions related to the skeleton of a person's image included in the imaging result of the imaging unit; a determination unit that determines the person's movements based on the detection result of the posture detection unit; and a command unit that generates a command indicating the response to the person based on the determination result of the determination unit. The determination unit determines the person's movements by comparing the detection result of the posture detection unit with operator movement pattern data defined in advance for the operator's movements, and the command unit generates the command by referring to command pattern data that associates the movements with the command. Furthermore, one representative monitoring method of the present invention includes an imaging step in which the monitoring device images an operator of a currency handling device and the area around the operator; a posture detection step in which it detects joint positions related to the skeleton of the image of the person included in the imaging result of the imaging step; a determination step in which it determines the movement of the person based on the detection result of the posture detection step; and a command step in which it generates a command indicating a response to the person based on the determination result of the determination step, wherein the determination step determines the movement of the person by comparing the detection result of the posture detection step with operator movement pattern data defined in advance for the operator's movement, and the command step generates the command by referring to command pattern data that associates the movement with the command. [Effects of the Invention]

[0007] According to the present invention, highly functional monitoring of the operation of a currency handling device can be achieved. Other problems, configurations, and effects will be clarified by the following description of the embodiments. [Brief explanation of the drawing]

[0008] [Figure 1] Diagram illustrating the monitoring of automated teller machines (ATMs). [Figure 2] Configuration diagram showing the ATM configuration [Figure 3] Flowchart showing ATM processing procedures [Figure 4] Detailed diagram of command pattern data [Figure 5] Specific examples of images and actions (Part 1) [Figure 6] Specific examples of images and actions (Part 2) [Figure 7] Specific examples of images and actions (Part 3) [Figure 8] Specific examples of images and actions (Part 4) [Figure 9] Specific examples of images and actions (Part 5) [Figure 10] Configuration diagram showing the ATM configuration of Example 2 [Figure 11]Specific Examples of Operation Support (Part 1) [Figure 12] Specific Examples of Operation Support (Part 2) [Figure 13] Specific Examples of Operation Support (Part 3) [Figure 14] Explanatory Diagram of Sharing of Irregular History [Figure 15] Explanatory Diagram of Object Detection [Figure 16] Explanatory Diagram of Advertisement Display

Mode for Carrying Out the Invention

[0009] Hereinafter, embodiments will be described with reference to the drawings.

Embodiment

[0010] FIG. 1 is an explanatory diagram regarding the monitoring of an automated teller machine (ATM). The ATM 10 shown in FIG. 1 is a transaction device that handles money and incorporates a function as a monitoring device for monitoring an operator and the surroundings thereof.

[0011] The ATM 10 includes an imaging unit that includes an operator and the surroundings thereof in an imaging range. The ATM 10 detects the skeleton of the operator and the skeletons of surrounding persons from an image that is an imaging result by the imaging unit. The skeleton of the operator indicates the posture of the operator. Similarly, the skeletons of surrounding persons indicate the postures of surrounding persons.

[0012] Based on the skeleton of the operator, the skeletons of surrounding persons, the state of the transaction, etc., the ATM 10 determines the actions of the operator and surrounding persons, and generates a command indicating a response to the operator and surrounding persons based on the determination result. The commands include those related to reporting when the operator is performing fraud, alerting the operator when a surrounding person is acting suspiciously, guiding the operator in operations, etc.

[0013] In this way, by imaging the operator and the surroundings of a transaction device, identifying the skeletons of people, determining actions, and giving warnings and guidance according to the actions, it is possible to implement countermeasures against fraud and operation support.

[0014] Figure 2 is a diagram showing the configuration of ATM 10. ATM 10 includes a monitoring control unit 21, a storage unit 22, an imaging unit 23, a vibration sensor 24, a microphone 25, a transaction control unit 31, a display and operation unit 32, a speaker 33, a coin storage unit 34, and a deposit / withdrawal unit 35.

[0015] The transaction control unit 31 is a control unit that processes various transactions, including those involving currency. Transactions include deposits, withdrawals, transfers, balance inquiries, and passbook updates. The transaction control unit 31 can be connected to a card reader, passbook processing unit, communication unit, receipt printer, etc. (not shown), and these can be used as needed in transactions.

[0016] The display and operation unit 32 is connected to the transaction control unit 31 and provides display output to the operator and accepts input from the operator. The display and operation unit 32 includes a touch panel display and buttons. The speaker 33 is connected to the transaction control unit 31 and provides audio output to the operator.

[0017] The currency storage unit 34 stores currency, i.e., banknotes and coins, sorted by denomination. The deposit and withdrawal unit 35 handles the deposit and withdrawal of currency. When currency is inserted into the deposit and withdrawal unit 35, the currency storage unit 34 counts the inserted currency by denomination and stores it, and notifies the transaction control unit 31 of the total amount and the balance of each denomination. The currency storage unit 34 also receives control from the transaction control unit 31 and withdraws the currency to the deposit and withdrawal unit 35.

[0018] The monitoring control unit 21 is connected to the transaction control unit 31, the storage unit 22, the imaging unit 23, the vibration sensor 24, and the microphone 25. The imaging unit 23 is a camera whose imaging range includes the operator and the area around the operator, and outputs the resulting image to the monitoring control unit 21. The vibration sensor 24 detects vibrations from the ATM 10 and outputs the detection result to the monitoring and control unit 21. The microphone 25 collects sounds from the surrounding area of ​​the ATM 10 and outputs the collected sound results to the monitoring and control unit 21.

[0019] The storage unit 22 is, for example, a hard disk drive, and stores operator action pattern data 22a, surrounding person action pattern data 22b, transaction operation procedure data 22c, and command pattern data 22d.

[0020] Operator action pattern data 22a is data that predefines the actions an operator takes when operating the ATM 10. The surrounding person behavior pattern data 22b is data that predefines the behavior of people located around the operator. The behavior of surrounding people includes, for example, appropriate behavior such as waiting in line, as well as behavior that suggests cheating, such as peeking.

[0021] Transaction operation procedure data 22c is data that shows the operation procedure for a transaction on the ATM 10. For example, the operation procedure shows the order of operations when performing that transaction. For example, the operation procedure for a withdrawal is "card acceptance," "amount input," and "cash withdrawal." Whether each operation has been performed can be obtained from the transaction control unit 31. In addition, the actions in each operation can be registered in the operator action pattern data 22a.

[0022] Command pattern data 22d is data that associates the actions of the operator and surrounding people with commands. Details of command pattern data 22d will be described later.

[0023] The monitoring control unit 21 is a control unit that controls the monitoring of the ATM 10, and can be implemented, for example, by a CPU (Central Processing Unit). The monitoring control unit 21 implements the functions of the attitude detection unit 21a, the determination unit 21b, and the command unit 21c.

[0024] The posture detection unit 21a detects the joint positions of the skeleton in the image of a person included in the imaging results of the imaging unit 23. Specifically, the posture detection unit 21a extracts the image of a person from the image and detects the joint positions of the skeleton. At this time, the distance to the person may be identified based on the position of the person's image. For example, the operator can be identified as being in the vicinity of the ATM 10 based on the size of the image in the image. For surrounding people, the distance to the ATM 10 can be estimated based on the size of the image in the image and the position of their feet.

[0025] The determination unit 21b determines the movement of the person based on the detection result of the posture detection unit 21a. The determination unit 21b determines the operator's movements by comparing the detection result of the posture detection unit 21a with the operator movement pattern data 22a. Furthermore, the determination unit 21b determines the movement of surrounding people by comparing the detection result of the posture detection unit 21a with the surrounding person movement pattern data 22b. The determination unit 21b can also determine a person's actions using the transaction status obtained from the transaction control unit 31. For example, it can determine that a card insertion action is being performed based on the operator's posture and the state of the card reader. Furthermore, the determination unit 21b can compare the detection result of the posture detection unit 21a with the transaction operation procedure data 22c to identify the transaction status, and can also use the identified transaction status to determine the person's movements. Furthermore, it is possible to determine a person's movements by further using the output of the microphone 25 and the vibration sensor 24. For example, if sound or vibration is detected when an operator is crouching in front of the ATM 10, it can be determined that there is a possibility that the operator is damaging the coin storage section 34.

[0026] The command unit 21c, based on the determination result of the determination unit 21b, refers to the command pattern data 22d and generates a command indicating how to respond to the operator and surrounding people. The command unit 21c can generate commands corresponding to the operator's actions and the actions of people in the vicinity by referring to the command pattern data 22d.

[0027] Furthermore, the command unit 21c can generate commands for situations identified by a combination of the operator's actions and the actions of surrounding individuals. For example, if an operator is crouching in front of ATM 10 and people nearby are waiting in line, it is highly likely that the operator is performing a legitimate action, such as picking up a dropped item. On the other hand, if the operator is crouching in front of ATM 10 and people nearby are acting cautiously, it is possible that the operator and those nearby are working together to commit fraud.

[0028] Command unit 21c can generate a command to report, warn, detain, or draw attention to the person's actions if they are inappropriate. Furthermore, the command unit 21c can generate commands to assist the operator's actions if the operator's actions deviate from the proper operating procedure. Furthermore, the command unit 21c can generate commands to deactivate and activate the ATM 10's power-saving mode based on the operator's approach and departure. Specific examples of these instructions will be discussed later.

[0029] Figure 3 is a flowchart showing the processing procedure of ATM 10. ATM 10 repeatedly executes the processes from steps S101 to S109. When processing begins, the imaging unit 23 first takes an image (step S101). The posture detection unit 21a detects the skeleton from the imaging results of the imaging unit 23 (step S102) and identifies the operator and surrounding people (step S103). The posture detection unit 21a detects the postures of the operator and surrounding people based on the positional relationships of the joints (step S104).

[0030] The determination unit 21b identifies the status of a transaction by obtaining it from the transaction control unit 31 or by referring to the transaction operation procedure data 22c (step S105). The determination unit 21b determines the actions of the operator and surrounding persons using the operator's posture, the posture of surrounding persons, the status of the transaction, etc. (step S106).

[0031] The command unit 21c refers to the command pattern data 22d based on the actions of the operator and surrounding persons, and generates a command if necessary (step S107). If a command has been generated, it is necessary to output the command (step S108; Yes), so the generated command is output (step S109) and the process ends. If no command has been generated (step S108; No), the process ends as is.

[0032] Figure 4 is an explanatory diagram detailing the command pattern data. As shown in Figure 4, the command pattern data 22d indicates that if the operator's action is "in the middle of a transaction" and the surrounding person's action is "peeking," a command to alert the operator is generated. Specifically, a command to output a message to the display operation unit 32 indicating that someone is peeking from behind is preferred. On the other hand, command pattern data 22d indicates no command if the operator is "in the middle of a transaction" and a nearby person is "on a mobile phone call".

[0033] Furthermore, the command pattern data 22d indicates that if the operator's actions result in a "delay or deviation from the operation procedure," a command to assist the operation will be generated. For example, if the operator gets stuck in the middle of the procedure shown in the transaction operation procedure data 22c, the display operation unit 32 will output a message indicating the next step.

[0034] Furthermore, the command pattern data 22d indicates that if the operator's action is "making a transfer while on a phone call," a command to prevent the operator from making the transfer will be generated. For example, by displaying a message such as "Please wait a moment. An officer will be on your way," it is possible to prevent transfer fraud. In addition, a report may be made. Alternatively, the operator may be directly informed that there is suspicion of transfer fraud.

[0035] Furthermore, the command pattern data 22d indicates that if the operator's actions are "disengaged and an item has been left behind," a warning command will be generated. For example, if the operator inserts a card and then leaves the ATM 10 without removing the card, it can be assumed that the card has been left behind. In this case, a command such as "You have left your card behind" is suitable for outputting a voice message from the speaker 33.

[0036] Furthermore, command pattern data 22d indicates no command if the operator's action is merely a short crouching position, as they may be picking up a dropped item.

[0037] Furthermore, the command pattern data 22d indicates that if the operator's action is to crouch for an extended period, different responses will be taken depending on the accuracy of the crouching detection. A prolonged crouching motion may indicate an abnormal act, such as vandalism against the coin storage compartment 34. Therefore, the command pattern data 22d associates a warning command (for example, an alarm sound output from the speaker 33) when the accuracy is low, and a command to notify a security guard when the accuracy is high.

[0038] When the operator's action is crouching, different responses may be taken depending on the actions of surrounding people. Command pattern data 22d assigns no command if the operator's action is "crouching" and the surrounding person's action is "waiting in line". On the other hand, command pattern data 22d assigns "notify security guard" if the operator's action is "crouching" and the surrounding person's action is "keeping watch on the surroundings".

[0039] The operator's actions may include unauthorized manipulation. Unauthorized manipulation is, for example, the act of attaching a machine to the ATM 10 that reads cards illegally. The command pattern data 22d indicates that a security guard should be notified when it is determined with a high degree of certainty that the operator is performing unauthorized manipulation. It also indicates that a security guard should be notified when it is determined with a low degree of certainty that the operator is performing unauthorized manipulation, or when the operator is occupying the area in front of the ATM 10 for an extended period of time and nearby persons are on alert.

[0040] Deterrence or tampering with the coin storage compartment 34 can also be detected using the output of the vibration sensor 24 and microphone 25. If vibrations or sounds caused by acts of destruction or tampering are detected, the accuracy of the determination that acts of destruction or tampering have occurred can be increased.

[0041] Furthermore, the command pattern data 22d indicates that it generates commands to deactivate and activate the ATM 10's power-saving mode based on the operator's approach and departure. The power-saving mode is a mode that suppresses the power consumption of unnecessary functions when the operator is not present. Specifically, the command pattern data 22d generates a command to deactivate the power-saving mode when it detects the approach of an operator. It also generates a command to activate the power-saving mode when the operator has finished a transaction and left, and there are no other people waiting in line nearby.

[0042] Next, we will explain specific examples of images and actions, referring to Figures 5 to 9. In Figure 5, the posture detection unit 21a detects the skeletal position of each person displayed on the screen, including the head, left and right shoulders, and the joints of the left and right arms. The determination unit 21b determines, for example, whether each person is the person facing the ATM (the ATM operator) or another person in the vicinity, based on the size of the triangle formed by the head and both shoulders, and its position on the screen.

[0043] The determination unit 21b further compares the skeletal information detected by the detection unit with predefined candidates for specific postures or actions for each person, and calculates the degree of match with those candidates. For the candidate with the highest degree of match, if the degree of match is above a certain level, that candidate is determined to be the posture of that person. If the degree of match is below a certain level, it is determined that there are no candidates or the posture is unknown.

[0044] For example, the determination unit 21b determines that the person in the foreground is an operator facing the ATM based on their size and position on the screen, and that they are performing a card insertion operation into the ATM 10 based on their skeletal information. The determination unit 21b determines that the person in the background is a person in the vicinity of the ATM 10 and that they are looking at the screen of the ATM 10.

[0045] The command unit 21c issues commands to the transaction control unit 31 for predefined actions corresponding to each piece of information reported by the determination unit 21b. For example, since inserting a card by an ATM 10 operator is within the scope of normal ATM operation, no command is generated in this case. Also, since a person nearby peering at the screen of the ATM 10 is an abnormal act, a warning is displayed on the ATM 10 screen. Based on the predefined priority of the definition information for these multiple persons, in this case, for example, the command for the actions of the person nearby is given priority, and the command unit instructs the higher-level device to display a warning on the ATM 10 screen. The priority can be set in the order of, for example, notification, warning, deterrence, attention, operation support, and no command, and if there are multiple commands corresponding to the determination result, the command with the highest priority should be generated.

[0046] In Figure 6, the determination unit 21b determines, for example, that the person in the foreground is an ATM operator based on their size and position on the screen, and that they are touching a button on the ATM 10 screen based on their skeletal information. Similarly, the determination unit determines that the person in the background is a person in the vicinity of the ATM 10 and is making a call on their mobile phone. The command unit 21c, for example, determines that if the ATM 10 operator is touching a button on the screen, this is within the scope of normal ATM operation and therefore no command is issued. Similarly, if a person near the ATM 10 is making a phone call, this is also a normal action and therefore no command is issued. Based on this definition information, the command unit 21c does not output any commands to the transaction control unit 31.

[0047] In Figure 7, the determination unit 21b determines that the person in the image is the operator of the ATM 10 and that the person is turning around and about to leave. The command unit 21c obtains information from, for example, the transaction control unit 31 regarding the progress of the operation being performed by the operator. For example, if it obtains information that the operator has completed a transaction, it does not need to output a command to the transaction control unit 31. Alternatively, if it obtains information that the operator has not yet taken out the cash that was withdrawn, it commands the speaker 33 to emit a warning sound or similar alert.

[0048] In Figure 8, the determination unit determines, based on the size and position on the screen, that the person in the image is a person facing the ATM 10, i.e., a person acting as an operator for convenience, and that the person is crouching in front of the ATM 10. The command unit 21c determines that the operator crouching down is an abnormal act such as damaging the safe, and for example, commands the transaction control unit 31 to call a security guard. The command unit 21c can also change the content of its commands depending on the circumstances, such as the duration of the action. For example, if the time spent crouching is short, it may consider it as a normal action such as picking up something that has been dropped and instruct the person to do nothing. If the time is longer than a certain duration, it may consider it as a potential act of destruction such as cutting the safe and instruct the security guard to be notified. The command unit 21c can change the content of its commands depending on the degree of match, for example, by issuing a warning or crime deterrent, such as emitting an alarm from an ATM, if the degree of match is low, as it does not necessarily mean that the person is in a position to commit vandalism. If the degree of match is high, it can consider the person to be in a position to commit vandalism and issue a warning to security guards.

[0049] In Figure 9, the determination unit 21b determines that the person in the image is an operator facing the ATM 10 and that, based on skeletal information, which part of the ATM 10 the person is operating. In Figure 9, the areas enclosed by dashed lines include the touch panel display, switches, card reader, coin deposit / withdrawal slot, and receipt printer. If the position of the operator's hand matches any of these areas enclosed by dashed lines, it is possible to determine which operation the operator performed. If the determined operation differs from the order shown in the transaction operation procedure data 22c, there is a possibility of an operational error, and the command unit 21c instructs the operator to perform the correct operation. [Examples]

[0050] Figure 10 is a configuration diagram showing the ATM configuration of Embodiment 2. The ATM 10 shown in Figure 10 has the same configuration as shown in Figure 2, but with the addition of a behavior estimation unit 21d, a facial expression detection unit 21e, an object detection unit 21f, and a heart rate detection sensor 26. Furthermore, the ATM 10 shown in Figure 10 can communicate with a facial recognition server 41 via a network.

[0051] The behavior estimation unit 21d, the facial expression detection unit 21e, and the object detection unit 21f are included in the monitoring control unit 21. The heart rate detection sensor 26 is a sensor that detects the operator's heart rate.

[0052] The action estimation unit 21d estimates the smoothness of the operator's operation. The command unit 21c generates commands to support the operator's operation if the smoothness of the operation is insufficient. The behavior estimation unit 21d estimates whether the operation is smooth or not by using factors such as delays in the operation, the operator's facial expressions, and the operator's heart rate.

[0053] A delay in operations can be determined by obtaining the transaction status in the transaction control unit 31 and comparing the estimated time required until the next operation with the actual time required for the operation. The operator's facial expression can be detected by the facial expression detection unit 21e performing image processing on the image of the operator's face. The operator's heart rate can be detected by the heart rate detection sensor 26.

[0054] The behavior estimation unit 21d estimates that the operation is not smooth if there is a delay in the operation. The behavior estimation unit 21d estimates that the smoothness of the operation is low if the operator's facial expression shows confusion. The behavior estimation unit 21d estimates that smoothness is low if the operator's heart rate is high. In addition to these, the behavior estimation unit 21d can estimate the smoothness of the operation using arbitrary indicators such as when the operator is about to perform an operation different from the procedure, when an operation cancellation is accepted, or when hesitant movements are detected.

[0055] The behavior estimation unit 21d stores information about operations that it estimates to have been performed insufficiently in a predetermined memory unit. For example, by accumulating information about operations that the operator was unable to perform smoothly in the memory unit 22, it can be used to improve the interface.

[0056] The command unit 21c associates information indicating the physical characteristics of the person who committed the fraudulent act (e.g., a facial image or facial image features) with information indicating the fraudulent act (e.g., the type of fraud), and registers this information with the facial recognition server 41, which is an external device. As a result, the facial recognition server 41 accumulates a history of fraudulent activities. The command unit 21c can query the facial recognition server 41 about the person captured by the imaging unit 23. If the query reveals a history of fraud, it links the information obtained from the new transaction and registers it with the facial recognition server 41. For example, if a person who committed fraud uses their own account, the fraudulent activity can be linked to the person's identity information. Furthermore, if the query reveals a history of fraud, the command unit 21c may generate a command for at least one of the following: reporting, warning, deterrence, or caution. In other words, for individuals who have committed fraud in the past, reporting, etc., can be done without needing to determine if fraud has occurred again.

[0057] The object detection unit 21f processes the image captured by the imaging unit 23 to detect an object held in a person's hand and identifies the detected object. Command unit 21c generates a command for notification and / or warning if the object is an object related to fraudulent activity. Objects related to fraudulent activity include items unnecessary for transaction operations and items that could be used to destroy ATM 10.

[0058] The ATM 10 may be equipped with a display unit capable of displaying advertisements. The display unit capable of displaying advertisements may be part of the display operation unit 32, or a separate display unit may be provided. The determination unit 21b can determine a person's gaze. The command unit 21c can switch the advertisement display when a person's gaze is directed towards the display unit for advertisements. Furthermore, the determination unit 21b can measure and record the amount of time a person's gaze was directed towards the display unit for advertising.

[0059] Figures 11 to 13 show specific examples of operation support. The transaction illustrated in Figure 11 shall be performed using the following steps: "Select transaction," "Insert card or passbook," "Enter PIN," "On-screen guidance (e.g., confirm fee)," "Enter amount," "Confirm amount," and "Receive banknotes, card, and receipt."

[0060] In Figure 11, if the transaction status is "inserting card or passbook" and the skeletal information indicates that the operator is confused, the command unit 21c will output the guidance "Please insert your card or passbook" to assist the operator. Also, if the transaction status is "amount input screen" and the skeletal information indicates that the operator is confused, the command unit 21c will output the guidance "Please enter the amount you wish to withdraw" to assist the operator.

[0061] In this way, by comparing the actions of the ATM 10 operator with the stage of the transaction procedure in which those actions are performed, the ATM 10 can provide appropriate guidance. The guidance can be displayed, audible, or controlled in a way that highlights the area that needs to be operated, such as by making the card slot flash more rapidly.

[0062] In addition to providing operational guidance, ATM10 also remembers at which stage of the transaction procedure the user became confused. This allows for adjustments to the guidance when the same person uses the ATM again, making it easier to understand the same situation. Furthermore, the points where many users become confused can be collected as big data, which can be used to improve the guidance.

[0063] Figure 12 shows that the transaction status is "inserting card or passbook," and facial recognition indicates that the user is confused. Therefore, the command unit 21c assists the user by outputting the guidance "Please insert card or passbook." The ATM 10 also remembers at which stage of the transaction procedure the user became confused.

[0064] In this way, by combining facial recognition data with transaction account information, it is possible to remember at what stage of the transaction a user became confused and how frequently that confusion occurred. This allows for measures such as providing operational assistance earlier for users or procedures that frequently cause confusion. A confused or troubled facial expression can be identified by detecting and determining facial expressions and actions such as furrowing the brow, drooping the corners of the eyes, or tilting the head.

[0065] In Figure 13, the transaction status is "inserting card or passbook," and it is presumed that the user is experiencing difficulty due to an increased heart rate. Therefore, the command unit 21c assists the user by outputting the guidance "Please insert your card or passbook." The ATM 10 also remembers at which stage of the transaction procedure the user became confused. Heart rate can be measured non-contact by detecting minute vibrations on the body surface using millimeter-wave radar. When the behavior estimation unit 21d detects an increase in heart rate, it estimates that the person is in a state of anxiety, such as not knowing the procedure, or a state of agitation, such as being a victim of a wire fraud.

[0066] Figure 14 is an explanatory diagram regarding the sharing of fraud history. In Figure 14, user A performs fraudulent activity at ATM 10a and then operates ATM 10b, which is located in a different location.

[0067] ATM10a registers and shares information about individuals who have committed fraudulent acts with an external device. Subsequently, if any ATM detects that individual, it links that individual's information. This allows for the accumulation of information that can be used for identity verification, such as the account information of individuals who have committed fraudulent acts.

[0068] Figure 15 is an explanatory diagram of object detection. ATM 10 detects objects held by a person nearby and issues an alert if the detected object is unnecessary for operating ATM 10 or could be used to damage ATM 10. The alert may be a notification to the security room or a warning to the person nearby. For example, a loud voice message such as "An attendant will be on your way. Please wait a moment" may be played. In this way, fraudulent activity can be deterred by issuing an alert while a suspicious object is in possession, before any fraudulent activity occurs.

[0069] Figure 16 is an explanatory diagram of the advertising display. The ATM 10 can display advertisements and promotions. For example, a portion of the display area of ​​the display operation unit 32 used for transactions may be used to display advertisements to the operator. Alternatively, a separate display unit for advertisements may be provided on the top of the ATM 10 so that it is visible to people nearby. ATM10 detects the gaze of the operator and surrounding individuals using image processing. When it detects that the gaze is directed towards an advertising display area, it switches the display content to show the advertisement. Furthermore, by measuring and recording the time the gaze is directed towards the display area, it can accumulate data for evaluating the content of the advertisement. In addition, the content of the advertisement may be determined in conjunction with user information and transaction details. For example, conditions such as "showing loan advertisements when the account balance is low and withdrawals or transfers are being made" or "offering investment information when the account balance is high and deposits are being made" could be used.

[0070] As described above, the ATM 10, which incorporates the function of a monitoring device, includes an imaging unit 23 whose imaging range includes the operator of the currency handling transaction device and the surrounding area, a posture detection unit 21a that detects joint positions related to the skeleton of the person's image included in the imaging result of the imaging unit 23, a determination unit 21b that determines the person's movements based on the operator detection result of the posture detection unit 21a, and a command unit 21c that generates a command indicating the response to the person based on the determination result of the determination unit 21b. Furthermore, the determination unit 21b determines the person's movements by comparing the detection result of the posture detection unit 21a with operator movement pattern data 22a that has been defined in advance for the operator's movements, and the command unit 21c generates the commands by referring to command pattern data 22d that associates the movements with the commands. With this configuration and operation, the monitoring device can achieve highly functional monitoring of the operation of the currency handling transaction device. Specifically, the monitoring device can deter criminal acts such as vandalism of ATM 10 and the theft of user information, and improve the usability of ATM 10 for users.

[0071] Furthermore, the determination unit 21b can determine the actions of surrounding persons located around the operator by using predefined surrounding person action pattern data 22b, and the command unit 21c can generate the command using the actions of the operator and the actions of the surrounding persons. Furthermore, the command unit 21c is capable of generating commands for situations identified by a combination of the operator's actions and the actions of the surrounding persons. Therefore, by taking advantage of the fact that people around the trading device may perform specific actions such as waiting in line, it is possible to comprehensively judge the actions of the operator and the people around them and generate appropriate commands.

[0072] Furthermore, the determination unit 21b can determine the actions of the person by further using the transaction status obtained from the transaction device. Furthermore, the determination unit 21b can identify the transaction status by comparing the detection result of the posture detection unit 21a with the transaction operation procedure data 22c which shows the operation procedure of the transaction device, and can further determine the actions of the person using the transaction status. In this way, by obtaining the transaction status, it is possible to determine the operator's actions in more detail.

[0073] Furthermore, the command unit 21c generates commands to support the operator's actions if the operator's actions deviate from the proper operating procedure. Therefore, monitoring devices can contribute to improving the convenience of the operator.

[0074] Furthermore, the command unit 21c generates a command for at least one of the following: reporting, warning, detaining, or alerting, if the person's actions are deemed to be illegal. Therefore, monitoring devices can effectively deter fraudulent activities.

[0075] The monitoring device further includes a microphone and / or a vibration sensor, and the determination unit 21b further uses the output of the microphone and / or vibration sensor to determine the person's movements. Therefore, acts of vandalism and unauthorized processing can be detected with high accuracy.

[0076] Furthermore, the command unit 21c generates commands to deactivate and activate the power-saving mode of the trading device based on the approach and departure of the operator. Therefore, the power saving mode can be started and stopped with high precision.

[0077] Furthermore, the monitoring device includes an action estimation unit 21d as an estimation unit for estimating the smoothness of the operator's operation, and the command unit 21c generates commands to support the operator's operation when the smoothness of the operation is insufficient. Therefore, it is possible to provide users with timely and appropriate guidance, thereby improving usability.

[0078] As an example, the monitoring device is characterized in that the estimation unit determines whether the operation is stalled based on the transaction status obtained from the trading device and the time required for the operation, and estimates the smoothness of the operation using the result of the stall determination. As an example, the monitoring device further includes a facial expression detection unit 21e that detects the operator's facial expression from the imaging results of the imaging unit, and the estimation unit determines whether the operator is confused or not based on the operator's facial expression, and uses the determination result to estimate the smoothness of the operation. As an example, the monitoring device further includes a heart rate detection sensor 26 that detects the operator's heart rate, and the estimation unit uses the heart rate to estimate the smoothness of the operation. In this way, the smoothness of the operation can be comprehensively estimated using factors such as the transaction status, the time taken for the operation, facial expressions, and heart rate.

[0079] Furthermore, the estimation unit stores information regarding the operation that it estimates to be insufficiently smooth in a predetermined storage unit. Therefore, it can be used to improve usability, such as changing the guidance for the next time the same person performs an action, or improving operations that are difficult for many people to understand.

[0080] Furthermore, if the actions of the person are fraudulent, the command unit 21c registers a history of fraud in an external device, associating information indicating the person's physical characteristics with information indicating the fraudulent act. The command unit 21c also registers information obtained from the person's new transactions in the external device, as a result of querying the external device for information about the person captured by the imaging unit 23, if a history of fraud exists. In this way, information about individuals who engage in fraudulent activities can be collected by accumulating information obtained from subsequent transactions.

[0081] The monitoring device further includes an object detection unit 21f that detects an object held in the person's hand and identifies the detected object, and the command unit 21c generates a command for notification and / or warning if the object is related to an illegal act. Therefore, it can deter fraudulent activities.

[0082] Furthermore, the monitoring device is further equipped with a display unit capable of displaying advertisements, the determination unit 21b determines the person's gaze, and the command unit 21c switches the advertisement display when the person's gaze is directed toward the display unit. Furthermore, the determination unit 21b measures and records the amount of time the person's gaze was directed towards the display unit. Therefore, it is possible to provide effective advertising to people who use the trading device.

[0083] It should be noted that the present invention is not limited to the embodiments described above, and various modifications are included. For example, the embodiments described above are explained in detail to make the present invention easier to understand, and are not necessarily limited to those having all the configurations described. Furthermore, it is possible to replace or add configurations, not just delete them. For example, in the above embodiment, a configuration in which the monitoring device function is built into the ATM, which is the transaction device, was used as an example for explanation. However, the monitoring device may be a separate device from the transaction device. If it is a separate device, it may be connectable to the transaction device. If it is not connected, instead of outputting various commands to the transaction device, the monitoring device will generate commands for its own speaker and communication functions.

[0084] Furthermore, although the above embodiment illustrates the application of the present invention to an ATM, it can be applied to any transaction device, such as ticket vending machines, automatic vending machines, and foreign currency exchange machines. [Explanation of Symbols]

[0085] 10: ATM, 21: Monitoring and Control Unit, 21a: Posture Detection Unit, 21b: Judgment Unit, 21c: Command Unit, 21d: Action Estimation Unit, 21e: Facial Expression Detection Unit, 21f: Object Detection Unit, 22: Memory Unit, 22a: Operator Motion Pattern Data, 22b: Surrounding Person Motion Pattern Data, 22c: Transaction Operation Procedure Data, 22d: Command Pattern Data, 23: Imaging Unit, 24: Vibration Sensor, 25: Microphone, 26: Heart Rate Detection Sensor, 31: Transaction Control Unit, 32: Display and Operation Unit, 33: Speaker, 34: Coin Storage Unit, 35: Deposit and Withdrawal Unit, 41: Face Recognition Server

Claims

1. An imaging unit whose imaging range includes the operator of a currency handling device and the area surrounding the operator, A posture detection unit detects the joint positions related to the skeleton of the human image included in the imaging results of the imaging unit, A determination unit determines the movement of the person based on the detection result of the posture detection unit, The system includes a command unit that generates a command indicating how to respond to the person based on the determination result of the determination unit, The determination unit determines the operator's actions by comparing the detection result of the posture detection unit with operator action pattern data that has been defined in advance for the operator's actions. The command unit generates the command by referring to command pattern data that associates the operation with the command. The determination unit further uses the detection result of the posture detection unit and predefined surrounding person movement pattern data for the movements of surrounding people located around the operator to determine the movements of the surrounding people. The command unit is capable of generating commands for situations identified by a combination of the operator's actions and the actions of surrounding persons. The aforementioned instructions include reporting when the operator is engaging in fraudulent activity, alerting the operator if a person in the vicinity is behaving suspiciously, and providing instructions to the operator on how to operate the system. A monitoring device characterized by the following features.

2. A monitoring device according to claim 1, The monitoring device is characterized in that the determination unit further determines the operator's actions using the transaction status obtained from the trading device.

3. A monitoring device according to claim 1, The monitoring device is characterized in that the determination unit identifies the status of a transaction by comparing the detection result of the posture detection unit with transaction operation procedure data indicating the operation procedure of the transaction device, and further determines the operator's actions using the status of the transaction.

4. A monitoring device according to claim 1, The monitoring device is characterized in that the command unit generates commands to support the operator's operations when the operator's operations deviate from the proper operating procedure.

5. A monitoring device according to claim 1, The monitoring device is characterized in that the command unit generates a command for at least one of the following: reporting, warning, detaining, or drawing attention, when the actions of the operator and / or the actions of the person in the vicinity are deemed to be illegal.

6. A monitoring device according to claim 1, A monitoring device further comprising a microphone and / or a vibration sensor, wherein the output of the microphone and / or vibration sensor is used to determine the movements of the person.

7. A monitoring device according to claim 1, The monitoring device is characterized in that the command unit generates commands to deactivate and activate the power-saving mode of the trading device based on the approach and departure of the operator.

8. The monitoring device, An imaging step of imaging the operator of a currency handling device and the area around the operator, A posture detection step for detecting joint positions related to the skeleton in the image of a person included in the imaging results of the aforementioned imaging step, A determination step is performed to determine the movement of the person based on the detection result of the posture detection step, The command step includes generating a command indicating how to respond to the person based on the determination result of the determination step, The determination step determines the operator's actions by comparing the detection result of the posture detection step with operator action pattern data that has been defined in advance for the operator's actions. The command step involves generating the command by referring to command pattern data that associates the operation with the command, The determination step further uses the detection result of the posture detection step and predefined surrounding person movement pattern data for the movements of surrounding people located around the operator to determine the movements of the surrounding people. The command step can generate commands for situations identified by a combination of the operator's actions and the actions of surrounding persons. The aforementioned instructions include reporting when the operator is engaging in fraudulent activity, alerting the operator if a person in the vicinity is behaving suspiciously, and providing instructions to the operator on how to operate the system. A monitoring method characterized by the following features.

9. A monitoring device according to claim 1, The system further includes an estimation unit that estimates the smoothness of the operator's operation, The monitoring device is characterized in that the command unit generates commands to assist the operator's operation when the smoothness of the operation is insufficient.

10. A monitoring device according to claim 9, The monitoring device is characterized in that the estimation unit determines whether the operation is stalled based on the transaction status obtained from the trading device and the time required for the operation, and estimates the smoothness of the operation using the result of the stall determination.

11. A monitoring device according to claim 9, The system further includes a facial expression detection unit that detects the operator's facial expression from the imaging results of the imaging unit, The monitoring device is characterized in that the estimation unit determines whether or not the operator is confused based on the operator's facial expression, and estimates the smoothness of the operation using the determination result.

12. A monitoring device according to claim 9, The system further includes a heart rate detection sensor that detects the operator's heart rate, The monitoring device is characterized in that the estimation unit estimates the smoothness of the operation using the heart rate.

13. A monitoring device according to claim 9, The monitoring device is characterized in that the estimation unit stores information regarding the operation that is estimated to be insufficiently smooth in a predetermined storage unit.

14. A monitoring device according to claim 1, The command unit, when it determines that the person's actions are fraudulent, registers a history of the fraud in an external device, which associates information indicating the person's physical characteristics with information indicating the fraudulent act. The monitoring device is characterized in that, if the command unit queries the external device for information about a person captured by the imaging unit and finds a history of fraud, it links the information obtained from the person's new transactions to the external device and registers it there.

15. A monitoring device according to claim 1, The system further includes an object detection unit that detects an object held in the hand of the person and identifies the detected object. The monitoring device is characterized in that the command unit generates a command for notification and / or warning when the object is an object related to illegal activity.

16. A monitoring device according to claim 1, It also features a display area capable of displaying advertisements, The determination unit determines the gaze of the person, The monitoring device is characterized in that the command unit switches the advertisement display when the person's gaze is directed toward the display unit.

17. A monitoring device according to claim 16, The monitoring device is characterized in that the determination unit measures and records the amount of time the person's gaze was directed toward the display unit.

Images