TWI752312B - Financial transaction fraud detection and prevention system and method - Google Patents

Abstract

The present invention provides a financial transaction fraud detection and prevention system and a method thereof. A fingerprint scanning module is set in the touch panel. When the user inputs the transaction password on the touch panel, the user can scan his finger and obtain the fingerprint. An image capture device is used to capture the video of the user, an analysis module is used to determine whether there are multiple withdrawal transactions from different accounts with the same fingerprint, and the user's face is determined according to the video captured by the image capture device. Whether the image and body movements conform to at least one characteristic of fraudulent behavior, and whether the sound contains fraudulent words or instruction words, etc., so as to use a warning module to issue a warning notification to inform the financial institution personnel that there are fraudsters or victims. payment.

Description

Financial transaction fraud detection and prevention system and method

The invention relates to a financial transaction system, in particular to a financial transaction fraud detection and prevention system and a method thereof.

At present, fraud cases are emerging in an endless stream. People often make wrong and regrettable decisions due to factors such as nervousness, fear, and worry, and then give their hard-earned money to the unlawful gainer. There is no clear basis for judging whether the withdrawal or remittance has been defrauded or not. It can only be asked verbally, which may make the customer impatient, or point the finger at the banker after the customer is defrauded.

In addition, in the ATM part, there is currently a limit system for non-agreed transfers and withdrawals. However, fraudsters use multiple illegally obtained financial cards to withdraw cash. As long as the limit is not exceeded, they cannot be prevented from withdrawing cash or transferring funds. It is also impossible to warn the same person to operate the ATM at the same time. Even if you switch to another ATM or even change to another branch to withdraw, it is extremely difficult to detect.

Therefore, the present invention proposes a financial transaction fraud detection and prevention system and method thereof, which effectively solve the above-mentioned problems. The specific structure and its implementation will be described in detail below:

The main purpose of the present invention is to provide a financial transaction fraud detection and prevention system and method thereof. A fingerprint scanning module is set in a touch panel for inputting transaction passwords to capture the user's fingerprint to determine whether the user is not. Make multiple withdrawals from different accounts to detect fraudsters early and prevent crimes from continuing.

Another object of the present invention is to provide a financial transaction fraud detection and prevention system and method thereof, which captures the user's transaction audio and video, analyzes and determines whether the user's facial expressions and body movements conform to a number of common behavioral characteristics of being defrauded, and if so, It can prevent the victim from being defrauded to transfer money.

Another object of the present invention is to provide a financial transaction fraud detection and prevention system and method thereof, wherein a pulse detection module is further arranged on the touch panel to obtain the user's pulse rate from the finger to determine the user's pulse rate. of tension.

In order to achieve the above object, the present invention provides a financial transaction fraud detection and prevention system, comprising: a touch panel for a user to input a transaction password for a withdrawal transaction, and the touch panel further includes a fingerprint scanner a module, which scans the user's finger when pressing the touch panel to obtain a fingerprint; an image capture device captures the user's video, including facial images and body movements; an analysis module, It is electrically connected with the touch panel and the image capture device, and includes a fingerprint comparison module and an image analysis module. The fingerprint comparison module determines that when there are withdrawal transactions of a plurality of different accounts, the fingerprint Whether the fingerprints scanned by the scanning module are the same, the image analysis module determines whether the facial image and the body movements of the user conform to at least one characteristic of fraudulent behavior; and a warning module is electrically connected to the analysis module , when the judgment result of the analysis module is that the same fingerprints are scanned in the withdrawal transactions of a plurality of different accounts, a first warning notification is issued, and when the analysis module judges that the facial image or the body movements of the user conform to When the fraudulent behavior is characterized, a second warning notice is issued.

According to an embodiment of the present invention, the touch panel further includes a pulse detection module for detecting a pulse rate from the finger of the user pressing the touch panel.

Continuing from the above, the analysis module further includes a pulse judging module for judging whether the pulse frequency exceeds a preset value, and if it exceeds, the warning module is required to issue a fourth warning notification. The preset value is pulse rate greater than 100 beats per minute.

According to an embodiment of the present invention, the characteristics of the fraudulent behavior include frowning, eyes widening, looking around, wiping sweat, breathing deeply, acting in a hurry, talking on the phone frequently, operating a machine while using a mobile phone, and the like.

According to an embodiment of the present invention, it further includes a storage module, which is electrically connected to the touch panel, the image capturing device and the analysis module, and temporarily stores the fingerprint and the video and audio.

The present invention further provides a financial transaction fraud detection and prevention method, comprising the following steps: a user inputs a transaction password for a withdrawal transaction on a touch panel; when the user's finger presses the touch panel, A fingerprint scanning module in the touch panel scans the finger and obtains a fingerprint, and an image capture device captures the user's video, including the user's facial image and body movements; using an analysis model The group judges whether the fingerprints scanned by the fingerprint scanning module are the same when there are multiple withdrawal transactions of different accounts, and judges whether the facial image and the body movements of the user conform to at least one characteristic of fraudulent behavior according to the video and audio. ; and when the judgment result of the analysis module is that the same fingerprints are scanned in the withdrawal transactions of a plurality of different accounts, use a warning module to issue a first warning notification, and when the analysis module judges the face of the user When the image or the body movement matches the characteristics of the fraudulent behavior, the warning module sends a second warning notice.

The present invention provides a financial transaction fraud detection and prevention system and a method thereof, which utilizes multiple judgment mechanisms to analyze the physiological information and facial expressions of users, which can not only prevent the unscrupulous and unscrupulous profit of fraudsters, but also reduce the incidence rate of customers being defrauded. On the other hand, financial institution personnel can also increase the basis for judging whether customers are fraudsters or defrauded, so as to avoid unreasonable treatment by customers.

Please refer to FIG. 1 , which is a block diagram of the first embodiment of the financial transaction fraud detection and prevention system of the present invention. The present invention includes a touch panel 10 , an image capture device 12 , an analysis module 14 and a warning The module 16, wherein the touch panel 10 is used for a user to input a transaction password for a cash withdrawal transaction. Generally speaking, the cash withdrawal transaction includes cash withdrawal at the counter and cash withdrawal through an ATM. If the cash withdrawal is at the counter The input transaction password is the deposit book password or the password set by other customers. If the ATM is used to withdraw money, the financial card password is input. The touch panel 10 further includes a fingerprint scanning module 102. When the user's finger presses When the transaction password is input on the touch panel, the fingerprint scanning module 102 scans the finger and obtains a fingerprint; the image capture device 12 is used to capture the user's video, including facial images, body movements and voices; the analysis module 14 is electrically connected to the touch panel 10 and the image capture device 12, the analysis module 14 can be set in the back-end server (not shown in the figure) of the financial institution, and includes a fingerprint comparison module 142, an image The analysis module 144 and a voice recognition module 148, the fingerprint comparison module 142 determines whether the fingerprints scanned by the fingerprint scanning module 102 are the same when there are multiple withdrawal transactions of different accounts, and the image analysis module 144 determines whether to use Whether the person’s facial image and body movements conform to at least one of the characteristics of the fraudulent behavior, such as frowning, wide-eyed, looking around, wiping sweat, taking a deep breath, acting in a hurry, talking on the phone frequently, operating the machine while using a mobile phone, etc. micro-expressions and body language; the voice recognition module 148 performs voice recognition on the voice of the user's audio and video, and analyzes whether the audio and video contains at least one word related to fraud or instructed operation words, such as canceling settings, installment payment, etc.; warning module The group 16 is electrically connected to the analysis module 14 for issuing a warning notice. For example, when the analysis module 14 determines that the same fingerprint is scanned for the withdrawal transaction of multiple different accounts, it means that a fraudulent driver is using multiple different financial cards When withdrawing money, a first warning notification will be issued. When the analysis module 14 determines that the user's facial image or body movements conform to the characteristics of the fraudulent behavior, a second warning notification will be issued to inform the financial institution personnel that the user is suspected of being defrauded. , will then carry out fraud-related publicity, remind users and counter staff and other measures to delay the withdrawal and remittance time, or immediately notify relevant units to assist in processing, when the analysis module 14 analyzes that the video contains fraud-related words or instructions to operate When the word is displayed, a third warning notice is issued.

FIG. 2 is a block diagram of the second embodiment of the financial transaction fraud detection and prevention system of the present invention. The financial transaction fraud detection and prevention system of the present invention can further be provided with a pulse detection module 104 in the touch panel 10 . A pulse rate can be detected from the user's finger pressing the touch panel 10, because from the pulse rate, it can be inferred whether the user's psychological state is tense, and the heart rate will increase when tense, so the analysis module 14 further includes a pulse rate The determination module 146 determines whether the user's pulse rate exceeds a preset value. For example, if the pulse rate per minute is greater than 100 beats, the user is likely to be a fraudster or a victim of fraud. Analyze Module 14 requests alert module 16 to issue a fourth alert notification to financial institution personnel.

In addition, the present invention further includes a body temperature sensor 13 , which can be installed on the image capture device 12 or can be installed independently, aiming at the direction of the user to detect the body temperature of the user, and setting a body temperature in the analysis module 14 The determination module 149 is used to determine whether the body temperature detected by the body temperature sensor 13 is higher than a predetermined body temperature, and the predetermined body temperature is the body temperature of ordinary people under normal conditions, such as 36-37.5 degrees Celsius, because people are nervous When using a mobile phone, the body temperature will increase, or the ear temperature will also increase due to the heat generated by the mobile phone when using the mobile phone, or the body temperature sensor 13 may detect that a high temperature object (such as a mobile phone) is placed on the ear Therefore, if the body temperature judging module 149 judges that the detected body temperature is higher than the preset body temperature, it means that a fraudster or a fraudulent driver is nervously operating the ATM or withdrawing money at the counter, the warning module 16 will issue a first Five warning notices.

The present invention further includes a storage module 18, which is electrically connected to the touch panel 10, the image capture device 12 and the analysis module 14 for temporarily storing fingerprints and video and audio. Whether the user has withdrawn from different accounts is determined from the fingerprints temporarily stored in the storage module 18, or whether the user is a fraudulent driver is determined based on the temporarily stored video and audio.

Next, please refer to FIG. 3 , which is a flow chart of the first embodiment of the method for detecting and preventing financial transaction fraud according to the present invention. First, in step S10, the user inputs a transaction password for a withdrawal transaction on a touch panel; in step S12, when the user's finger presses the touch panel, the fingerprint scanning module in the touch panel scans Finger and obtain a fingerprint, and the image capture device captures the user's video, the video includes the user's face image and body movements; then in step S14, the image analysis module in the analysis module is used to determine the usage Whether the facial image and the body movement of the user conform to at least one characteristic of fraudulent behavior, and if so, it means that the user may be defrauded, then a second warning notification is sent to the financial institution personnel in step S16 to prevent the user from continuing to withdraw. If not, proceed to step S18, use the speech recognition module to perform speech recognition on the sound part of the video and audio, and analyze whether the audio and video contains at least one word related to fraud or instruction operation words, if yes, it represents the user It may be the victim who has been defrauded, and a third warning notice is issued in step S20; if not, continue to step S22 and use the fingerprint matching module in the analysis module to determine whether it is a withdrawal transaction of multiple different accounts but scans If the fingerprints are the same, if so, a first warning notification will be sent to the financial institution personnel in step S24, informing that it may be a fraudulent driver. On the contrary, if not, the user can continue the transaction, withdraw cash or transfer funds, as shown in step S26 described.

FIG. 4 is a flow chart of the second embodiment of the method for detecting and preventing financial transaction fraud according to the present invention. In step S30, the user inputs a transaction password for the withdrawal transaction on a touch panel; in step S32, when the user's finger presses the touch panel, the fingerprint scanning module in the touch panel scans the finger A fingerprint is obtained, and the image capture device captures the user's video, which includes the user's facial image and body movements; then in step S34, it is determined whether the user's pulse rate has reached a preset value, and if so , then in step S36, a fourth warning notification is issued to the financial institution personnel, informing the person that the person is in a state of tension and may be a fraud driver or a fraud victim. If the pulse is normal, then step S38 detects whether the user's body temperature exceeds a preset body temperature. If so, it means that the user is in a state of tension and may be a person who has been defrauded. Therefore, step S40 issues a fifth warning notification. If not, then in step S42, the image analysis module in the analysis module is used to determine whether the facial image and the body movements of the user conform to at least one characteristic of fraudulent behavior. If so, it means that the user may be defrauded. Then, in step S44, a second warning notification is sent to the financial institution personnel to prevent the user from continuing to withdraw or transfer funds; Fraud-related words or instruction operation words, the former means that the user may be a defrauded person, and the latter means that the user may be a fraudulent driver, so a third warning notification will be issued in step S48; if the voice recognition result is normal, then step S50 In the process, the fingerprint comparison module in the analysis module is used to determine whether it is a withdrawal transaction of a plurality of different accounts but the same fingerprint is scanned. It is a fraudulent driver, otherwise, if not, the user can continue the transaction, withdraw cash or transfer money, as described in step S54.

For example, if a user withdraws money at the first ATM in the same branch, the fingerprint is temporarily stored in the storage module, and then the user switches to other ATMs and uses multiple different debit cards to withdraw money or For money transfer, since the touch panel of all ATMs is connected to the storage module and analysis module of the back-end server, even if the user changes the machine without continuous operation, the analysis module can still analyze multiple different transactions. Withdrawals or transfers from accounts are performed by people with the same fingerprints, which in turn detect the fraudulent driver and prevent him from continuing his wrongdoing.

For another example, if the user not only has a fast pulse rate when withdrawing money, but also has a tense facial expression, a physical expression of bewilderment, and even a rise in body temperature, it means that he is very likely to be deceived to transfer money to the bank. fraudsters, the analysis module can also prevent them from continuing to trade before they are transferred out.

The present invention can also incorporate the above detection and analysis data into formulas, for example, the total judgment score L is divided into 7 scoring criteria, including behavior a, expression b, heartbeat c, fingerprint d, body temperature e, fraudulent word f and the amount to be claimed g, and add a weight M to each scoring standard, and the sum of the weights is 1, and the fraud score of the operator can be calculated, as shown in the following formula (1): L = M1a+ M2b+ M3c+ M4d+ M5e+ M6f+ M7g Among them, M1+ M2+ M3+ M4+ M5+ M6+ M7 = 1, 0 < M1 < 1 (1) Among them, each standard has detailed items. For example, behavior a is subdivided into a fixed number of parameters such as talking on the phone a1, taking a deep breath a2, wiping sweat a3, and talking to others a4. According to different scoring standards, different numbers of parameters are set. The detected scoring standard score is 100 points, as shown in the following formula (2): 0 ≦ a = (a1)+(a2)+(a3)+(a4) ≦ 100 (2)

Then, the judgment score L is imported into the warning level table, and different warning measures are given according to different levels, such as promoting fraudulent videos, enhancing the attention of the staff, operating the screen prompt button to confirm, or immediately notifying the relevant units to assist in processing, etc.

To sum up, the financial transaction fraud detection and prevention system and method of the present invention captures the fingerprints of the user to determine whether the user has made multiple withdrawals from different accounts, so that fraudsters can be detected early and the crime can be prevented from continuing. occur. In addition, the present invention captures the user's transaction audio and video, analyzes and determines whether the user's facial expressions and body movements conform to a number of common behavior characteristics of being defrauded, and combines voice recognition to determine whether the user has spoken words related to being defrauded, and if so, then It can prevent the victim from being fraudulently transferred out, and conversely, it can strengthen the identification of whether the same person is constantly operating the transaction. Furthermore, the present invention captures the user's pulse rate on the touch panel, and provides a body temperature sensor to detect the user's body temperature, both of which can determine the user's nervousness, combined with the above fingerprint recognition. , image recognition, voice recognition, etc., with a high probability to determine whether the user has been defrauded or defrauded the driver, so as to reduce the incidence of illegal transactions.

Only the above descriptions are only preferred embodiments of the present invention, and are not intended to limit the scope of the present invention. Therefore, all equivalent changes or modifications made according to the features and spirits described in the scope of the application of the present invention shall be included in the scope of the application for patent of the present invention.

10: Touch panel 102: Fingerprint Scanning Module 104: Pulse detection module 12: Image capture device 13: Body temperature sensor 14: Analysis module 142: Fingerprint comparison module 144: Image Analysis Module 146: Pulse judgment module 148: Speech recognition module 149: Body temperature judgment module 16: Warning module 18: Storage Module

FIG. 1 is a block diagram of the first embodiment of the financial transaction fraud detection and prevention system of the present invention. FIG. 2 is a block diagram of the second first embodiment of the financial transaction fraud detection and prevention system of the present invention. FIG. 3 is a flow chart of the first embodiment of the method for detecting and preventing financial transaction fraud according to the present invention. FIG. 4 is a flow chart of the second embodiment of the method for detecting and preventing financial transaction fraud according to the present invention.

10: Touch panel

102: Fingerprint Scanning Module

12: Image capture device

14: Analysis module

142: Fingerprint comparison module

144: Image Analysis Module

148: Speech recognition module

16: Warning module

Claims (10)

A financial transaction fraud detection and prevention system, comprising: a touch panel for a user to input a transaction password for a withdrawal transaction, the touch panel further includes a fingerprint scanning module, which scans the finger when the user's finger presses the touch panel and get a fingerprint; an image capture device to capture the video and audio of the user, including facial images, body movements and voices; An analysis module, electrically connected to the touch panel and the image capture device, includes: A fingerprint comparison module to determine whether the fingerprints scanned by the fingerprint scanning module are the same when there are multiple withdrawal transactions from different accounts; an image analysis module to determine whether the facial image and the body movements of the user conform to at least one characteristic of fraudulent behavior; a voice recognition module, which performs voice recognition on the user's voice in the audio and video, and analyzes whether the audio and video contains at least one fraudulent-related word or instruction word; and a warning module, electrically connected to the analysis module, when the analysis module determines that the same fingerprint is scanned for withdrawal transactions of a plurality of different accounts, a first warning notification is issued, and when the analysis module determines that the When the facial image or the body movement of the user matches the characteristics of the fraudulent behavior, a second warning notification is issued, and when the voice recognition module analyzes that the video and audio contains the fraudulent related words or instruction operation words, A third warning notice is issued. The financial transaction fraud detection and prevention system as claimed in claim 1, wherein the touch panel further includes a pulse detection module for detecting a pulse from the finger of the user pressing the touch panel frequency, the analysis module further includes a pulse judging module to determine whether the pulse frequency exceeds a preset value, and if it exceeds, the warning module is required to issue a fourth warning notification, and the preset value is the pulse rate per minute more than 100 times. The financial transaction fraud detection and prevention system as claimed in claim 1, wherein the characteristics of the fraudulent behavior include frowning, wide-eyed, looking around, wiping sweat, taking a deep breath, acting in a hurry, talking on the phone frequently, using a mobile phone while operating a machine Taiwan and so on. The financial transaction fraud detection and prevention system as described in claim 1 further includes a body temperature sensor to detect the user's body temperature, and the analysis module further includes a body temperature judgment module to judge the body temperature Whether the body temperature detected by the detector is higher than a preset body temperature, the preset body temperature is the body temperature of ordinary people under normal conditions, if so, the warning module sends a fifth warning notification, and the body temperature sensor updates The ear temperature of the user is detected, so that the analysis module can determine whether there is heat energy generated by using the mobile phone. The financial transaction fraud detection and prevention system as claimed in claim 1 further includes a storage module electrically connected to the touch panel, the image capture device and the analysis module to temporarily store the fingerprint and the video. A financial transaction fraud detection and prevention method, comprising: A user inputs a transaction password for the withdrawal transaction on a touch panel; When the user's finger presses the touch panel, a fingerprint scanning module in the touch panel scans the finger and obtains a fingerprint, and an image capture device captures the user's video, including the User's facial images, body movements and voices; An analysis module is used to determine whether the fingerprints scanned by the fingerprint scanning module are the same when there are multiple withdrawal transactions from different accounts, and based on the video and audio to determine whether the user's facial image and the body movements conform to at least one of The characteristics of the fraudulent behavior, and perform speech recognition according to the voice in the video to analyze whether the video contains at least one fraudulent-related word or instruction word; and When the determination result of the analysis module is that the same fingerprints are scanned in the withdrawal transactions of a plurality of different accounts, a warning module is used to issue a first warning notification, and when the analysis module determines that the user's face image or When the physical action conforms to the characteristics of the fraudulent behavior, the warning module sends a second warning notice, and when the voice recognition module analyzes that the video and audio contains the fraudulent-related words or instruction words, it sends a third warning Alert notification. The method for detecting and preventing financial transaction fraud according to claim 6, wherein the touch panel further includes a pulse detection module for detecting a pulse from the finger of the user pressing the touch panel frequency, the analysis module is used to determine whether the pulse rate exceeds a preset value, and if it exceeds, the warning module is required to issue a fourth alarm notification, and the preset value of the pulse rate per minute is greater than 100 beats per minute. The method for detecting and preventing financial transaction fraud according to claim 6, wherein the characteristics of the fraudulent behavior include frowning, wide-eyed, looking around, wiping sweat, taking a deep breath, acting in a hurry, talking on the phone frequently, using a mobile phone while operating a machine Taiwan and so on. The method for detecting and preventing financial transaction fraud according to claim 6, further utilizes a body temperature sensor to detect the body temperature of the user, and the analysis module further includes a body temperature judgment module to judge the body temperature sensing Whether the body temperature detected by the device is higher than a preset body temperature, the preset body temperature is the body temperature of ordinary people under normal conditions, if so, the warning module sends a fifth warning notification, and the body temperature sensor further detects The ear temperature of the user is used by the analysis module to determine whether there is thermal energy generated by using the mobile phone. The method for detecting and preventing financial transaction fraud according to claim 6, further comprising a storage module electrically connected to the touch panel, the image capture device and the analysis module to temporarily store the fingerprint and the video.

Images