TW202020703A - Transaction system, automated teller machine and method for card-less transaction - Google Patents

Abstract

A transaction system includes an automated teller machine, a face recognition server unit and a host unit. The transaction system is able to use face recognition technology and living body identification technology to implement an enable process and an execution process of card-less transaction. In the enable process of card-less transaction, the transaction system add relevant data which is for identifying user during card-less transaction. In the execution process of card-less transaction, upon receiving a card-less transaction starting command and determining a condition of living body is met, the automated teller machine shoots and generates image data and transfers the image data to the face recognition server unit, so that the face recognition server unit transfers a face recognition result to the automated teller machine. After determining the face recognition result indicates verification passed, the automated teller machine obtains a client account of specific client data from the host unit. Upon receiving a card-less transaction command entered by user, the automated teller machine executes the card-less transaction.

Description

Transaction system, automatic teller machine and card-free transaction method

The invention relates to a transaction system, an automatic teller machine and a cardless transaction method, in particular to a transaction system, an automatic teller machine and a cardless transaction method using face recognition and living body identification to verify identity.

In the existing cardless withdrawal system, the drawer can withdraw money at the automated teller machine (ATM) without holding a financial card, which is very convenient. In order to improve the security of the cardless withdrawal system, an existing authentication method is that the withdrawer must cooperate with an application of a mobile device to withdraw money from the ATM. For the withdrawer, the operation process is lengthy and complicated Sacrificing convenience. Another way to improve security is to use finger vein identification technology for identity verification. However, ATMs using this method must be equipped with a vein sensing device, which requires higher hardware costs.

Therefore, one of the objects of the present invention is to provide a transaction system that uses living body recognition and face recognition to verify the identity and overcome the inconvenience of the prior art.

Therefore, the transaction system of the present invention includes an ATM, a face recognition server unit, and a host unit. The automatic teller machine includes an image capturing module and a processing unit electrically connected to the image capturing module. The image capturing module can be photographed under the control of the processing unit. The face recognition server unit communicates with the ATM. The host unit communicates with the ATM, and stores multiple pieces of customer data corresponding to multiple customers, each piece of customer data includes an identification code corresponding to the corresponding customer and at least one customer account. The ATM transmits the user identification code to the host unit when receiving a cardless transaction initiation command and a user identification code input by the user, so that the host unit selects from the customer data A specific customer profile of the ID code that matches the user ID. The automatic teller machine generates and sends a face recognition request to the face recognition server unit when a living body recognition model determines that a living body condition is met, wherein the living body condition includes the existence of living bodies in the shooting range of the image capturing module, And the face recognition request includes an image data generated by shooting by the shooting module and the user identification code. When receiving the face recognition request, the face recognition server unit generates a face recognition verification based on the image data of the face recognition request and a face feature information stored in advance and related to the user identification code As a result, the face recognition verification result is transmitted to the ATM. After judging that the face recognition verification result indicates that the verification is passed, the ATM obtains the at least one client account number of the specific client data from the host unit, and upon receiving a user input and indicating the specific client data When one of the at least one customer account is related to a cardless transaction instruction of a cardless transaction, the cardless transaction is executed.

In some implementation aspects of the trading system of the present invention, the living body recognition model is pre-trained in a deep learning manner.

In some embodiments of the transaction system of the present invention, when the ATM determines that the face recognition verification result indicates that the verification is passed and also receives a user password input by the user, the user password is transmitted to The host unit. When the host unit receives the user password and determines that the user password matches a cardless transaction password included in the specific customer data, the at least one customer account number of the specific customer data is transmitted to the ATM.

In some implementation aspects of the transaction system of the present invention, the face recognition server unit not only transmits the face recognition verification result, but also encrypts a verification data to the ATM, so that the ATM will receive The verification data of the face recognition server unit is transmitted to the host unit, and the host unit stores the verification data when successfully decrypting the verification data received from the ATM, wherein the verification The verification data includes the identification code of the specific customer data, a determination date when the face recognition verification result is generated, and a pair of determination serial numbers corresponding to the face recognition verification result.

In some embodiments of the transaction system of the present invention, the ATM further includes a data reading unit electrically connected to the processing unit, and before the ATM receives the cardless transaction initiation instruction: the ATM is When it is detected that the data reading unit is electrically connected to a data carrier and receives an input password input by the user, it is determined whether the input password is consistent with a card password stored in the data carrier; the ATM determines After the input password matches the card password, use the living body identification model to judge whether the living body conditions are met, and send an image data generated by the shooting module to the face when the judgment result is yes Recognition server unit; the face recognition server unit obtains and stores the identification code of the specific customer data and a facial feature information related to the image data, and creates the identification code of the specific customer data and the Correspondence between facial feature information.

In some implementations of the transaction system of the present invention, each customer profile also includes a communication profile, and after the ATM determines that the input password matches the card password, the host unit determines the specific customer profile based on the The communication data transmits a one-time password to a pair of portable electronic devices corresponding to the communication data, and a verification password input by the user at the ATM and consistent with the one-time password is received from the ATM At this time, a password is correctly notified to the ATM, and the ATM only judges whether the living condition is met when receiving the notification that the password is correct from the host unit.

In some implementations of the transaction system of the present invention, when the ATM determines that the input password matches the card password and receives a string input by the user, it generates and transmits a string containing the string and indicating The cardless transaction password establishment request of the specific customer data is sent to the host unit, and the host unit adds the string as a cardless transaction password to the specific customer information when receiving the cardless transaction password establishment request.

In some embodiments of the transaction system of the present invention, the ATM sends multiple pieces of image data to the face recognition server unit when it judges that the living body condition is met; the face recognition server unit receives the images When data is selected, a target image data that meets an applicable condition is selected from the image data, and the target image data is transmitted to the ATM, wherein the applicable condition relates to the availability of each image data in face recognition. Recognition; when the ATM receives the target image data from the face recognition server unit, the target image data is displayed, and upon receiving a confirmation command input by the user and corresponding to the target image data, Causing the identification code of the specific customer data to be provided to the face recognition server unit and sending a file creation request to the face recognition server unit; when the face recognition server unit receives the specific customer data The ID code and the file creation request, the ID code of the specific customer data and a facial feature information related to the target image data are stored.

Another object of the present invention is to provide an automatic teller machine that uses living body recognition and face recognition technology.

The automatic teller machine of the present invention is suitable for communicating with a face recognition server unit and a host unit. The host unit stores multiple pieces of customer data corresponding to multiple customers, and each piece of customer data includes a corresponding one corresponding to the customer. Identification code and at least one customer account. The ATM includes a processing unit and an image shooting module. The image shooting module is electrically connected to the processing unit and can be photographed under the control of the processing unit. The processing unit transmits the user identification code to the host unit when receiving a card-less transaction initiation instruction and a user identification code input by the user, so that the host unit selects from the customer data A specific customer profile of the ID code that matches the user ID. The processing unit generates and sends a face recognition request to the face recognition server unit when using a living body recognition model to determine that a living body condition is met, wherein the living body condition includes the existence of a living body in the shooting range of the image capturing module, And the face recognition request includes an image data generated by shooting by the shooting module and the user identification code. The face recognition request is used for the face recognition server unit to generate a face recognition verification result. After receiving the face recognition verification result and judging that the face recognition verification result indicates that the verification is passed, the processing unit obtains the at least one client account of the specific client information from the host unit, and receives a When a person enters and indicates one of the at least one customer account of the specific customer profile and is related to a cardless transaction instruction of a cardless transaction, the cardless transaction is executed.

In some embodiments of the ATM of the present invention, the living body recognition model is pre-trained in a deep learning manner.

In some implementations of the ATM of the present invention, the processing unit transmits the user password to the user when the face recognition verification result indicates that the verification is passed and a user password input by the user is received The host unit, so that the host unit transmits the at least one client account number of the specific client information to the ATM when it determines that the user password matches a cardless transaction password included in the specific client information, and, The processing unit executes the card-free transaction with the host unit when the at least one client account number of the specific client data is received and the card-free transaction instruction is also received.

Another object of the present invention is to provide a cardless transaction method.

The cardless transaction method of the present invention is suitable for execution by a transaction system including an ATM, a face recognition server unit and a host unit. The ATM includes an image capturing module and an electrical connection to the image capturing module The processing unit of the group, the image shooting module can be controlled by the processing unit to shoot, the host unit stores multiple pieces of customer data corresponding to multiple customers, and each piece of customer data includes a corresponding one corresponding to the customer An identification code and at least one customer account, the cardless transaction method includes: (A) the ATM receives the user identification code when receiving a cardless transaction initiation instruction and a user identification code input by the user Sent to the host unit, so that the host unit selects from the customer data a specific customer data containing the identification code that matches the user identification code; (B) the ATM uses a living body identification model to determine A face recognition request is generated and sent to the face recognition server unit when a living body condition is met, wherein the living body condition includes the presence of a living body in the shooting range of the image capture module, and the face recognition request includes a Image data generated by the shooting module and the user identification code; (C) when the face recognition server unit receives the face recognition request, according to the image data of the face recognition request and a A face feature information pre-stored and related to the user identification code generates a face recognition verification result, and transmits the face recognition verification result to the ATM; (D) the ATM determines the face recognition The verification result indicates that after verification is passed, the at least one client account number of the specific client information is obtained from the host unit, and one of the at least one client account number input by the user and indicating the specific client information is received And when related to a cardless transaction instruction of a cardless transaction, the cardless transaction is executed.

In some implementation aspects of the cardless transaction method of the present invention, in step (B), the living body recognition model is pre-trained in a deep learning manner.

In some embodiments of the cardless transaction method of the present invention, step (D) includes: (D1) the ATM determines that the face recognition verification result indicates that the verification is passed and also receives a usage input by the user User password, the user password is transmitted to the host unit; (D2) when the host unit receives the user password and determines that the user password matches a cardless transaction password contained in the specific customer data , Transmitting the at least one customer account number of the specific customer information to the ATM; (D3) the ATM receives the at least one customer account number of the specific customer information and also receives the card-free transaction instruction, and The host unit executes the card-free transaction.

In some implementation aspects of the cardless transaction method of the present invention, in step (C), the face recognition server unit not only transmits the face recognition verification result, but also encrypts the verification data to the ATM , So that the ATM transmits the verification data received from the face recognition server unit to the host unit, and causes the host unit to decrypt the verification data received from the ATM The verification data is stored, wherein the verification data includes the identification code of the specific customer data, a determination date when the face recognition verification result is generated, and a pair of determination serial numbers corresponding to the face recognition verification result.

In some embodiments of the cardless transaction method of the present invention, the ATM further includes a data reading unit electrically connected to the processing unit; the cardless transaction method further includes: (E) the When the ATM detects that the data reading unit is electrically connected to a data carrier and receives an input password input by the user, it determines whether the input password is consistent with a card password stored in the data carrier; (F ) After judging that the input password is consistent with the card password, the ATM uses the living body identification model to judge whether the living body conditions are met, and transmits a shot generated by the shooting module when the judgment result is yes Image data to the face recognition server unit; (G) the face recognition server unit obtains and stores the identification code of the specific customer data and a facial feature information related to the image data, and creates the specific The correspondence between the identification code of the customer data and the facial feature information.

In some embodiments of the cardless transaction method of the present invention, each customer profile also includes a communication profile, and the cardless transaction method also includes between steps (E) and (F): (H) the The host unit transmits a one-time password to a pair of portable electronic devices corresponding to the communication data according to the communication data of the specific customer data, and receives from the ATM a user input from the ATM and the When the verification password matches the one-time password, a correct password is notified to the ATM; in step (F), the ATM only judges the living condition when it receives the correct notification from the host unit Whether it meets.

In some embodiments of the cardless transaction method of the present invention, the cardless transaction method further includes between steps (E) and (F): (I) when the ATM determines the input password and the card After the passwords match and a string entered by the user is received, a cardless transaction password creation request containing the string and indicating the specific customer information is generated and sent to the host unit; (J) the host unit receives When the request for establishing a cardless transaction password is established, the string is added to the specific customer information as a cardless transaction password.

In some embodiments of the cardless transaction method of the present invention, in step (F), the ATM transmits multiple pieces of image data to the face recognition server unit, and the cardless transaction method further includes steps Between (F) and (G): (K) The face recognition server unit selects a target image data that meets an applicable condition from the image data, and transmits the target image data to the ATM, Wherein, the applicable condition is related to the recognizability of each image data in face recognition; (L) when the ATM receives the target image data from the face recognition server unit, the target image data is displayed, and Upon receiving a confirmation command input by the user and corresponding to the target image data, the identification code of the specific customer data is provided to the face recognition server unit, and a file creation request is sent to the Face recognition server unit; (M) when the face recognition server unit receives the identification code of the specific customer data and the file creation request, the identification code of the specific customer data is stored and a related Facial feature information of the target image data.

Another cardless transaction method of the present invention is suitable for being executed by an ATM, which is suitable for communicating with a face recognition server unit and a host unit. The ATM includes an image capturing module and an electrical connection A processing unit of an image shooting module, which can be controlled by the processing unit to shoot, and the host unit stores a plurality of pieces of customer data corresponding to a plurality of customers, and each piece of customer data includes a corresponding customer Corresponding to an identification code and at least one customer account, the cardless transaction method includes: (A) the ATM will use the ATM when receiving a cardless transaction initiation instruction and a user identification code input by the user The identification code is sent to the host unit, so that the host unit selects from the customer data a specific customer data containing the identification code that matches the user identification code; (B) the ATM is using a living body The recognition model generates and sends a face recognition request to the face recognition server unit when it determines that a living body condition is met, wherein the living body condition includes the existence of a living body in the shooting range of the image capture module, and the face recognition request includes An image data generated by the shooting module and the user identification code, the face recognition request is used for the face recognition server unit to generate a face recognition verification result; (C) the processing unit is After receiving the face recognition verification result and judging that the face recognition verification result indicates that the verification is passed, the at least one client account number that obtains the specific client information from the host unit, and upon receiving a user input and instruction When one of the at least one customer account of the specific customer profile is associated with a cardless transaction instruction related to a cardless transaction, the cardless transaction is executed.

In some implementation aspects of another cardless transaction method of the present invention, in step (B), the living body identification model is pre-trained in a deep learning manner.

In some implementation aspects of another cardless transaction method of the present invention, step (C) includes: (C1) the ATM determines that the face recognition verification result indicates that the verification is passed and receives a user input User password, the user password is transmitted to the host unit, so that the host unit determines the specific customer information when it determines that the user password matches a cardless transaction password included in the specific customer information The at least one customer account number is transmitted to the ATM; (C2) the ATM machine executes the cardless operation with the host unit when receiving the at least one customer account number of the specific customer information and also receiving the cardless transaction instruction transaction.

The effect of the present invention is that when the ATM determines that the living conditions are met and the face recognition verification result indicates that the verification is passed, the identity verification of the user is completed, and the subsequent cardless transaction is performed, compared with the current The identity verification method of the card transaction process has the advantage of reducing the user's operating burden. Moreover, the ATM does not need to be equipped with a finger vein recognition device or perform face recognition, which is beneficial to the hardware construction and maintenance costs .

Before the present invention is described in detail, it should be noted that in the following description, similar elements are denoted by the same number.

Referring to FIG. 1, an embodiment of the transaction system 100 of the present invention belongs to a bank, for example, and includes an ATM 1, a host unit 2, and a face recognition server unit 3. The ATM 1, the host unit 2 and the face recognition server unit 3 can communicate with each other through a communication network 4. In this embodiment, the communication network 4 is, for example, an internal network belonging to the bank, but In other embodiments, and the communication network 4 may also be the Internet, for example, and is not limited to this embodiment.

The ATM 1 includes an image capturing module 11, a storage unit 12, an input unit 13 for user operation, an output unit 14, and a data carrier (such as a chip-type financial card or credit card) ) A data reading unit 15 inserted and capable of reading the data carrier, and a processing unit electrically connected to the image capturing module 11, the storage unit 12, the input unit 13, the output unit 14 and the data reading unit 15 16.

In this embodiment, the image capturing module 11 is implemented as a camera, for example, and can be controlled by the processing unit 16 to generate images, and the image capturing module 11 is configured to be When the input unit 13 operates the ATM 1, it is possible to photograph the user's face.

The storage unit stores a living body identification model 121 that is implemented using a living body detection (English for liveness detection) technology. In this embodiment, the living body identification model 121 is pre-trained in a deep learning manner, for example, and the living body The recognition model 121 can be used by the processing unit 16 to execute a living body recognition program. Specifically, in the living body recognition process of this embodiment, the processing unit 16 will first control the image capturing module 11 to continuously perform multiple shootings to generate multiple (for example, ten but not limited) image data Then, the living body recognition model 121 is used to analyze the image data to generate a living body recognition result, and the living body recognition result indicates whether the shooting range of the image shooting module 11 includes living bodies, but not limited to this .

In this embodiment, the input unit 13 is, for example, implemented as a keyboard, the output unit 14 has, for example, a display screen and a speaker (not shown), and the data reading unit 15 is, for example, implemented as a reader Card machine. However, in other embodiments, the input unit 13 and the display screen of the output unit 14 may be jointly implemented as a touch screen, and not limited to this embodiment.

In this embodiment, the processing unit 16 is, for example, a central processing unit, and the central processing unit can read the living body recognition model 121 in the storage unit 12 and execute a living body recognition algorithm in the living body recognition model 121 law. However, in other embodiments, the living body recognition model 121 may also be implemented as a hardware circuit inside the central processor. More specifically, in other embodiments, the processing unit 16 may include the The central processing unit and the living body recognition model 121 electrically connected to the central processing unit, and the living body recognition model 121 is implemented as a living body recognition circuit (not shown in the figure) implemented by hardware, instead of this embodiment Limited. In addition, the living body identification model 121 can also be implemented as a combination of hardware circuits and algorithms (that is, software), and is not limited to this embodiment.

In this embodiment, the host unit 2 stores a plurality of customer data 21 corresponding to a plurality of customers, and each customer corresponding to the customer data 21, for example, holds one or more financial institutions opened by the bank Account, and each client profile 21 includes an identification code 211, one or more client accounts 212 (two are exemplarily shown in FIG. 1), and a communication profile 213. In this embodiment, for each customer profile 21 and its corresponding customer, the identification code 211 is, for example, the customer's ID number, and each customer account 212 is, for example, a financial account belonging to the customer (For example, a living account or a foreign currency account), and the communication data 213 is, for example, a mobile phone number of the customer, but not limited to this.

In this embodiment, the host unit 2 includes, for example, a front-end system (FEP) (not shown) electrically connected to the ATM 1 through the communication network 4, and an electric connection to the front-end system to store the customers The data 21 is also used as a host (not shown) for executing the bank's electronic financial transactions, but it is not limited to this.

In this embodiment, the face recognition server unit 3 stores a plurality of identification codes 31 and a plurality of facial feature information 32 corresponding to the identification codes 31, respectively. In particular, the identification codes 31 correspond to many of the customers who already hold the bank’s financial account and have also opened a face recognition cardless transaction service. Taking a specific customer A among these customers as an example, if the specific customer A holds a financial account opened by the bank but does not open the face recognition cardless transaction service, the host unit 2’s There will be a specific customer data 21' corresponding to the specific customer A in the customer data 21 (as shown in FIG. 1), but there will be no identity recognition corresponding to the specific customer A in the face recognition server unit 3 Code 31 and facial feature information 32. In this embodiment, each identity code 31 in the face recognition server unit 3 is, for example, the ID number of the corresponding customer, and it will be the same as the identity code 211 corresponding to the same customer in the host unit 2 Match,.

It is added that, in this embodiment, the face recognition server unit 3 includes, for example, a recognition server (not shown) built with a face recognition algorithm and used to perform face recognition, and a A database server (not shown) electrically connected to the identification server and used for storing the identification codes 31 and the facial feature information 32, but not limited to this.

Referring to FIGS. 1 and 2 (FIGS. 2A and 2B) at the same time, the following exemplarily explains in detail how the transaction system 100 of the present embodiment implements a cardless transaction method using face recognition and living body identification to verify the identity of the present invention. In this embodiment, the cardless transaction method includes a cardless transaction opening process, and a cardless transaction execution process that can be implemented after the cardless transaction opening process is completed. The card transaction opening process will be described.

For ease of understanding, the specific customer A is used as a user in the following description, and it is assumed that the specific customer A is operating the ATM 1 in order to activate the face recognition cardless transaction service.

First, in step S101, when the processing unit 16 of the ATM 1 receives a cardless transaction opening instruction input by the user via the input unit 13, the processing unit 16 controls the display screen of the output unit 14 Display a financial card request message. In this embodiment, the cardless transaction opening instruction is generated, for example, by the user operating the input unit 13 to select a cardless transaction opening option, and the financial card request message may indicate, for example, "please insert a financial card", But not limited to this. Next, step S102 is performed.

In step S102, when the processing unit 16 detects that the data reading unit 15 is inserted by a chip financial card as a data carrier and is electrically connected to a chip on the chip financial card, the data reading unit 15 is controlled Read the wafer. Moreover, when the processing unit 16 also receives an input password input by the user via the input unit 13, the processing unit 16 determines whether the input password matches a card password stored in the chip. In addition, if the processing unit 16 determines that the input password matches the card password (that is, the user has input the correct card password), then step S103 is performed.

It is added that in this embodiment, if the processing unit 16 determines that the input password does not match the card password (that is, the user enters the wrong card password), and also determines that the input password and the card When the number of first consecutive errors whose passwords do not match is within three times, the processing unit 16 controls the display screen of the output unit 14, for example, to display "The chip financial card password is incorrect, please re-enter" to instruct the user to re-enter the card password . On the other hand, when the processing unit 16 determines that the input password does not match the card password, and also determines that the number of the first consecutive errors has exceeded three times (that is, the number of times the user enters the wrong card password exceeds three times) , The processing unit 16 controls, for example, the display screen of the output unit 14 to display “the chip financial card password is incorrect, please re-operate”, and controls the data reading unit 15 to withdraw the financial card, and the cardless transaction opening process ends , But not limited to this.

In step S103, when the processing unit 16 determines that the input password matches the card password, the display screen of the output unit 14 is controlled to display a first password setting message, and the first password setting message may be, for example, " Please enter the cardless transaction password you want to set", but not limited to this. Then, when the processing unit 16 receives a first character string input by the user and corresponds to the first password setting message through the input unit 14, the display screen of the output unit 14 is controlled to display a second password setting Message, and the second password setting message may be, for example, "please enter the cardless transaction password to be set again", but not limited to this. Then, when the processing unit 16 receives a second character string input by the user and corresponds to the second password setting message through the input unit 14, and also determines that the second character string and the first character string are mutually When they match, the processing unit 16 generates a cardless transaction password establishment request containing the first character string (equivalent to including the second character string) and indicating the specific customer A, and passes the cardless transaction password establishment request via The communication network 4 is sent to the host unit 2. In this embodiment, the cardless transaction password creation request, for example, further includes a card number or a customer account stored in the chip and corresponding to the specific customer information 21', thereby indicating the specific customer information 21', but Not limited to this.

After the processing unit 16 transmits the cardless transaction password establishment request to the host unit 2, the process proceeds to step S104.

In step S104, when the host unit 2 receives the cardless transaction password creation request from the ATM 1 via the communication network 4, the host unit 2 first stores from the cardless transaction password creation request The specific customer data 21' corresponding to the specific customer A is selected from the customer data 21 of, and the first character string (equivalent to the second character string) is added to the specific customer data as a cardless transaction password 214 21' (as shown in FIG. 3), so that the cardless transaction password 214 is included in the specific customer profile 21', and the specific customer profile 21' is set to a state in which face recognition cardless transaction is enabled. Next, step S105 is performed.

In step S105, the host unit 2 generates a one-time password (One Time Password in English, referred to as OTP), and transmits the one-time password in the form of a short message according to the communication data 213 of the specific customer data 21' To a pair of portable electronic devices (not shown) corresponding to the communication data 213. The portable electronic device may be, for example, a mobile phone held by the specific customer A, but it is not limited thereto. Next, step S106 is performed.

In step S106, when the processing unit 16 of the ATM 1 receives a verification password input by the user via the input unit 13, the verification password is transmitted to the host unit 2 via the communication network 4. Next, step S107 is performed.

In step S107, when the host unit 2 receives the verification password from the ATM 1 via the communication network 4 within a predetermined period, it is determined whether the verification password is consistent with the one-time password. When the host unit 2 judges that the verification password is consistent with the one-time password (that is, the user enters the correct one-time password on the ATM 1), a password correct notification is generated, and via the communication network 4 Transmit the correct notification of the password to the processing unit 16 of the ATM 1.

In this embodiment, the predetermined period may be, for example, one minute, for example, when the host unit 2 transmits the one-time password to the portable electronic device, but it is not limited thereto. Moreover, it is added that if the host unit 2 determines that the verification password does not match the one-time password (that is, the user enters the wrong one-time password at the ATM 1), and also determines When the number of second consecutive errors where the verification password does not match the one-time password is within three times, the host unit 2 transmits a first password error notification to the processing unit 16 of the ATM 1 via the communication network 4 , So that the processing unit 16 controls the display screen to display, for example, "single password error, please re-receive the SMS and re-enter the one-time password" according to the first password error notification, and then proceed again from step S105. On the other hand, if the host unit 2 determines that the verification password does not match the one-time password, and the second consecutive error has exceeded three times (that is, the user has entered the wrong one-time password more than three times in a row) At this time, the host unit 2 sends a second password error notification to the processing unit 16 via the communication network 4, so that the processing unit 16 controls the display screen according to the second password error notification such as "Failed to open, please re-operate Once, and control the data reading unit 15 to withdraw the financial card, and the card-free transaction opening process is ended, but not limited to this.

After the host unit 2 transmits the correct notification of the password to the processing unit 16 of the ATM 1, step S108 follows.

In step S108, when the processing unit 16 of the ATM 1 receives the correct password notification from the host unit 2, the processing unit 16 controls the output unit 14 to output a shooting reminder notification to prompt the user to take a picture Then, the processing unit 16 determines whether a living body condition is met, and then proceeds to step S109 if the result of the determination is yes.

In this embodiment, the processing unit 16 determines whether the living body condition is met by first performing the living body recognition process with the living body recognition model 121 in the storage unit 12 to generate the living body recognition result, and then according to the living body recognition The result judges whether the living condition is met. Specifically, when the living body recognition result indicates that the shooting range of the image capturing module 11 includes a living body, the processing unit 16 determines that the living body condition is met, on the other hand, when the living body recognition result indicates the image shooting When the imaging range of the module 11 does not include a living body, the processing unit 16 determines that the living body condition is not met. It is added that if the processing unit 16 determines that the living body condition is not met, the processing unit 16 will control the display screen to display, for example, "Failed to open" and control the data reading unit 15 to withdraw the financial card, and the The card transaction opening process ends, but not limited to this. In addition, the shooting prompt notification includes, for example, a shooting prompt voice output by the speaker of the output unit 14 and a shooting prompt message displayed by the display screen of the output unit 14, the shooting prompt voice indicating, for example, " Ready to take photos, please look directly at the lens", and the shooting prompts indicate the precautions that should be taken when shooting, such as "Do not cover eyebrows, facial features and facial contours", "Do not wear sunglasses" and other messages, but do not take This is limited.

In step S109, the processing unit 16 controls the image capturing module 11 to automatically capture three times at intervals of, for example, 0.5 seconds to generate three pieces of image data and transmit the three pieces of image data to the network via the communication network 4 Face recognition server unit 3. Next, step S110 is performed.

In step S110, when the face recognition server unit 3 receives the three image data from the ATM 1, it performs face recognition on the three image data and generates three images corresponding to the three images, respectively The three recognizable score results of the data. For each recognizable score result, the level of the value is positively related to the corresponding number of facial features of the user’s face in the corresponding image data, the recognizable degree and sharpness of facial features and contours, etc., but Not limited to this. Since the technical details of the face recognition server unit 3 performing face recognition on the three strokes of image data are not the focus of the present invention, they will not be described here. Next, step S111 is performed.

In step S111, the face recognition server unit 3 determines whether there is a piece of target image data that meets an applicable condition among the three pieces of image data. In this embodiment, for each piece of image data, the applicable condition includes that the recognizability score corresponding to the image data is greater than a predetermined threshold, and the recognizability score corresponding to the image data is The highest of all such recognizable scores, but not limited to this. In a case where the face recognition server unit 3 determines that the target image data exists in the three strokes of image data, then step S112 is performed.

It is added that if the face recognition server unit 3 determines that the target image data does not exist in the image data, the face recognition server unit 3 will send a shooting failure notification to the The processing unit 16 of the ATM 1. In this embodiment, when the processing unit 16 receives the shooting failure notification for the first time, it will first control the display screen of the output unit 14 to display a shooting failure message and a re-shooting option, and when the processing unit 16 passes When the input unit 13 receives a re-shooting command input by the user and corresponds to the re-shooting option, it will start again from the aforementioned step S109, so that even if the first shooting fails, the user does not need to Re-inserting the financial card can directly take the second shot, so the operation is more time-saving and convenient. However, when the processing unit 16 receives the shooting failure notification for the second time, it will no longer provide the re-shooting option for the user to select, and the cardless transaction opening process ends.

In step S112, the face recognition server unit 3 transmits the target image data to the processing unit 16 of the ATM 1 via the communication network 4, and then proceeds to step S113.

In step S113, when the processing unit 16 of the ATM 1 receives the target image data from the face recognition server unit 3, the display screen of the output unit 14 is controlled to display the target image data for use Reference. Next, step S114 is performed.

In step S114, when the processing unit 16 receives a confirmation command input by the user and corresponds to the target image data through the input unit 13, the processing unit 16 enables the identification of the specific customer data 21' The code 211 (in this case, the ID number of the specific customer A is taken as an example) is provided to the face recognition server unit 3, and a file creation request is sent to the face recognition server unit 3 via the communication network 4. Next, step S115 is performed.

In particular, in this embodiment, the processing unit 16 may, for example, first obtain the identification code 211 of the specific customer profile 21' from the host unit 2, and then recognize the identity of the specific client profile 21' The code 211 is sent to the face recognition server unit 3 via the communication network 4, or the processing unit 16 can also send commands to the host unit 2 so that the host unit 2 directly sends the specific customer data The identification code 211 of 21' is transmitted to the face recognition server unit 3 via the communication network 4, so that the face recognition server unit 3 obtains the identification code 211 of the specific customer data 21', but not This is the limit.

In addition, in this embodiment, if the processing unit 16 receives the target image data from the face recognition server unit 3 for the first time in step S113, the processing unit 16 controls the display screen to display the target image data At the same time, the display screen is also controlled to display the re-shooting option for the user to select, and if the processing unit 16 receives the re-shooting command corresponding to the re-shooting option via the input unit 13 after step S113, It starts again from the aforementioned step S109. As such, in step S113, if the user is not satisfied with the result of the first shooting, as long as the input unit 13 is operated to select the re-shooting option displayed on the display screen, the second shooting can be directly performed Without having to do it again. However, if the processing unit 16 receives the target image data from the face recognition server unit 3 for the second time in step S113 (that is, the user has retaken it once), it will no longer control the display The screen displays the re-shooting option for the user to select.

In step S115, referring to FIG. 4, when the face recognition server unit 3 receives the identification code 211 of the specific customer data 21' and the file creation request from the ATM 1, the face recognition The server unit 3 performs the following operations according to the file creation request: 1. The identification code 211 of the specific customer data 21' is stored as an identification code 31' corresponding to the user (that is, the specific customer A) . 2. Generate and store a facial feature information 32' corresponding to the target image data. In this embodiment, the face recognition server unit 3 may directly store the target image data as the face feature information 32', however, in other embodiments, the face recognition server unit 3 also The target image data may be subjected to image analysis and/or image processing, thereby generating the facial feature information 32' corresponding to the target image data, and is not limited to this embodiment. 3. Establish a unique correspondence between the identification code 31' and the facial feature information 32'.

In this way, as shown in FIG. 5, through the above steps S101 to S115, the host unit 2 has added the specific customer A to the ATM 1 in the specific customer data 21 ′. The cardless transaction password 214, and the face recognition server unit 3 has also stored the identification code 31' and the facial feature information 32' corresponding to the specific customer A, and completed this embodiment of the Cardless transaction opening process.

Referring to FIGS. 5 and 6 (FIG. 6A and FIG. 6B) at the same time, the following describes the execution flow of the cardless transaction in this embodiment. And, following the previous example, the following continues to use the specific customer A as a user, and it is assumed that the specific customer A is operating the ATM 1 and wants to use the face recognition cardless transaction service for withdrawal.

First, in step S201, when the processing unit 16 of the ATM 1 receives a cardless transaction initiation instruction and a user identification code input by the user via the input unit 13, the user identification code is transmitted To the host unit 2. In this embodiment, the cardless transaction initiation command is generated by the user operating the input unit 13 to select a cardless transaction option, and the user identification code is, for example, the user (here the specific customer A is an example) operate the ID number of the ID card input by the input unit 13, but not limited to this. For ease of explanation, it is assumed here that the user identification code received by the processing unit 16 and the identification code 211 in the specific customer profile 21' and the identification code 31' in the face recognition server unit 3 Agree with each other. Next, step S202 is performed.

In step S202, when the host unit 2 receives the user identification code from the ATM 1, the specific customer data 21' is selected from the customer data 21 according to the user identification code, and the specific Whether the customer information 21' has been set to the status of face recognition without card transaction. In a case where the judgment result of the host unit 2 is yes, the host unit 2 transmits a notification of activation confirmation to the ATM 1, and then proceeds to step S203.

In step S203, when the processing unit 16 of the ATM 1 receives the activation confirmation notification from the host unit 2, the processing unit 16 controls the output unit 14 to output a shooting reminder notification to prompt the user to take a picture, Next, the processing unit 16 determines whether the living body condition is met, and proceeds to step S204 if the result of the determination is yes. In this step, the processing unit 16 determines whether the living condition is met in the same manner as described above, and the shooting prompt notification described in this step is also the same as the shooting prompt notification described in step S108 in the cardless transaction opening process , So I won’t repeat it here. In addition, in this embodiment, if the processing unit 16 determines that the living body condition is not met, the processing unit 16 controls the display screen to display, for example, "transaction failed", and the cardless transaction execution process ends, but not as limit.

In step S204, the processing unit 16 controls the image capturing module 11 to perform a single shot to generate image data for face recognition. Next, step S205 is performed.

In step S205, the processing unit 16 generates a face recognition request including the user identification code (that is, the ID number entered by the specific customer A) and the image data, and sends the request through the communication network 4 The face recognition request is sent to the face recognition server unit 3. Next, step S206 is performed.

In step S206, when the face recognition server unit 3 receives the face recognition request from the ATM 1, the face recognition server unit 3 first stores the The ID code 31' matching the user ID is selected from the ID codes 31, and the image data is compared with the facial feature information 32' corresponding to the ID code 31' to generate One indicates the face recognition verification result of passing or failing verification. In this embodiment, the face recognition server unit 3 first extracts the facial features of the specific customer A included in the image data from the image data requested by the face recognition, and then extracts The facial feature is compared with the facial feature information 32', and when a similarity between the two is greater than or equal to a preset threshold, the face recognition verification result indicates that the verification is passed, otherwise, when When the degree of similarity is less than the preset threshold, the face recognition verification result indicates that the verification fails.

After the face recognition server unit 3 generates the face recognition verification result, it proceeds to step S207.

In step S207, the face recognition server unit 3 not only encrypts the face recognition verification result, but also encrypts the verification data to the processing unit 16 of the ATM 1 via the communication network 4. In this embodiment, the verification data includes, for example, the identification code 31', a determination date when the face recognition verification result is generated this time, and a pair of determination serial numbers corresponding to the face recognition verification result. Moreover, in this embodiment, when the processing unit 16 receives the verification data from the face recognition server unit 3, it will transmit the verification data to the host unit 2, and the host unit 2 When the verification data from the ATM 1 is received, it will be decrypted, and upon successful decryption, the decrypted verification data will be stored as a face recognition verification record that has passed verification for subsequent verification For the purpose of nuclear inspection, but not limited to this.

The face recognition server unit 3 transmits the face recognition verification result and the verification data to the ATM 1, and then proceeds to step S208.

In step S208, when the processing unit 16 of the ATM 1 receives the face recognition verification result, it is determined whether the face recognition verification result indicates that the verification is passed, and if the judgment result is yes, then Go to step S209.

It is added that in the case where the processing unit 16 determines that the face recognition verification result indicates that the verification is not passed, if the processing unit 16 also determines that it is the first time to receive the face recognition verification result indicating that the verification fails Then, the processing unit 16 will first control the display screen of the output unit 14 to display a face recognition failure message, and then start again from step S204 to provide the user with a second chance of face recognition. However, if the processing unit 16 also determines that it is the second time to receive the face recognition verification result indicating that the verification fails, then the processing unit 16 controls the display screen to display the face recognition failure message, the cardless transaction The execution process ends.

In step S209, the processing unit 16 controls the display screen of the output unit 14 to display a cardless transaction password request message, and when the processing unit 16 also receives a user input by the user via the input unit 13 At the time of password, the user password is transmitted to the host unit 2. Next, step S210 is performed.

In step S210, when the host unit 2 receives the user password, it is determined whether the user password matches the cardless transaction password 214 in the specific customer profile 21'. When the host unit 2 determines that the user password matches the cardless transaction password 214 of the specific customer profile 21' (that is, the specific customer A enters the correct cardless transaction password at the ATM 1) The host unit 2 transmits the customer account numbers 212 of the specific customer profile 21' (two in this embodiment are taken as examples) to the processing unit 16 of the ATM 1, and then proceeds to step S211.

It is added that in this embodiment, if the host unit 2 determines that the user password does not match the cardless transaction password 214 of the specific customer profile 21' (that is, the specific customer A is on the ATM 1 An incorrect cardless transaction password is entered), the host unit 2 sends a cardless transaction password error notification to the processing unit 16 of the ATM 1 via the communication network 4, so that the processing unit 16 controls the output unit 14 by Instruct the user to re-enter the cardless transaction password, and the ATM 1 allows the user to re-enter the cardless transaction password for up to two times, but not limited to this.

In step S211, when the processing unit 16 of the ATM 1 receives the two customer account numbers 212 of the specific customer profile 21', the display screen of the output unit 14 is controlled to indicate the two customer account numbers 212 respectively. Account option, and when the processing unit 16 also receives a cardless transaction instruction input by the user and indicating one of the two customer account numbers 212 and a cardless transaction via the input unit 13, the ATM 1 The processing unit 16 and the host unit 2 jointly execute the card-free transaction. Specifically, the cardless transaction is a withdrawal transaction in this embodiment, but in another embodiment, it may also be, for example, a transfer transaction or a deposit transaction, without being limited to this embodiment. The cardless transaction instruction is generated by the user operating the input unit 3 to select one of the two account options, for example. In addition, the ATM 1 and the host unit 2 execute the cardless transaction according to the financial account corresponding to the customer account 212 indicated by the cardless transaction instruction, for example, thereby completing the execution of the cardless transaction in this embodiment Process.

It is added that in the cardless transaction execution flow of this embodiment, the ATM 1 first displays the cardless transaction password request message (step S209), and the host unit 2 determines that the user input After the user's password matches the card-free transaction password 214 (step S210), the ATM 1 will display the two account options (step S211) for the user to select the desired financial method for card-less transaction Account, in other words, the user must enter the correct cardless transaction password before making a financial account selection. However, in the cardless transaction execution flow of another embodiment, the ATM 1 may also obtain the specific from the host unit 2 after determining that the face recognition verification result indicates that the verification is passed (step S208) The two client account numbers 212 of the client profile 21' also display the two account options, and the cardless transaction password request message is displayed only when the account selection instruction is received, and then, the user password is sent when the user password is received To the host unit 2 for the host unit 2 to compare the user password with the cardless transaction password 214. In other words, in the cardless transaction execution process of another embodiment, the user may first Select the financial account you want to use, and then enter the cardless transaction password, but not limited to this embodiment.

In summary, when the ATM 1 in this embodiment judges that the living conditions are met and the face recognition verification result indicates that the verification is passed, the identity verification of the user is completed, and the subsequent follow-up with the host unit 2 is performed. Compared with the current cardless transaction authentication method, the cardless transaction has the advantage of reducing the user's operation burden. Moreover, the ATM 1 in this embodiment does not need to be equipped with a finger vein identification device nor Performing face recognition is beneficial to hardware construction and maintenance costs, so it can indeed achieve the purpose of cost invention.

However, the above are only examples of the present invention, and should not be used to limit the scope of implementation of the present invention, any simple equivalent changes and modifications made according to the scope of the patent application of the present invention and the content of the patent specification still belong to This invention covers the patent.

100: Transaction system 1: ATM 11: Image shooting module 12: Storage unit 121: Living body recognition model 13: Input unit 14: Output unit 15: Data reading unit 16: Processing unit 2: Host unit 21: Customer data 21 ': Specific customer information 211: Identification code 212: Customer account 213: Communication data 214: Cardless transaction password 3: Face recognition server unit 31: Identification code 31': Identification code 32: Facial feature information 32 ': Facial feature information 4: Communication network S101~S114: Step S201~S211: Step

Other features and functions of the present invention will be clearly presented in the embodiment with reference to the drawings, in which: FIG. 1 is a block diagram exemplarily illustrating an embodiment of the trading system of the present invention; FIG. 2 (by the diagram 2A and 2B) is a flowchart illustrating how the embodiment implements a cardless transaction opening process in a cardless transaction method; FIG. 3 is a schematic diagram showing an example of a cardless transaction password It is added to a piece of specific customer data by this embodiment; FIG. 4 is a schematic diagram exemplarily showing the identification code and facial feature information stored in a face recognition server unit; FIG. 5 is a block diagram, exemplary Shows the embodiment after the cardless transaction opening process is completed; and FIG. 6 (composed of FIGS. 6A and 6B) is a flowchart illustrating an example of how the embodiment implements the cardless transaction method. Cardless transaction execution process.

100: trading system

1: ATM

11: Image shooting module

12: storage unit

121: Living body recognition model

13: input unit

14: output unit

15: Data reading unit

16: Processing unit

2: Host unit

21: Customer Information

21’: Customer-specific information

211: Identification code

212: Customer account

213: Communication materials

214: Cardless transaction password

3: Face recognition server unit

31: ID code

31’: Identification code

32: Facial feature information

32’: Facial feature information

4: Communication network

Claims (22)

A transaction system, including: an automatic teller machine, including an image shooting module and a processing unit electrically connected to the image shooting module, the image shooting module can be photographed under the control of the processing unit; a face recognition server unit , To communicate with the ATM; and a host unit to communicate with the ATM, and store multiple pieces of customer data corresponding to multiple customers, each piece of customer data includes a corresponding identification corresponding to the customer Code and at least one customer account number; wherein: when the ATM receives a card-less transaction initiation instruction and a user identification code input by the user, the user identification code is transmitted to the host unit to enable the host The unit selects from the customer data a specific customer data containing the identification code that matches the user identification code; the ATM generates and sends a face recognition request when a living body identification model determines that a living body condition is met To the face recognition server unit, wherein the living condition includes the existence of a living body in the shooting range of the image shooting module, and the face recognition request includes an image data generated by shooting by the shooting module and the User identification code; when the face recognition server unit receives the face recognition request, according to the image data of the face recognition request and a pre-stored face feature information related to the user identification code Generating a face recognition verification result, and transmitting the face recognition verification result to the ATM; after determining that the face recognition verification result indicates that the verification is passed, the ATM obtains the specific customer information from the host unit At least one customer account, and when receiving one of the at least one customer account input by the user and indicating the specific customer information and related to a card-free transaction instruction, execute the card-free transaction . The transaction system according to claim 1, wherein the living body recognition model is pre-trained in a deep learning manner. The transaction system according to claim 1, wherein: when the ATM determines that the face recognition verification result indicates that the verification is passed and also receives a user password input by the user, the user password is transmitted to The host unit; when the host unit receives the user password and determines that the user password matches a cardless transaction password included in the specific customer data, the at least one customer account of the specific customer data is sent to The ATM. The transaction system according to claim 1, wherein the face recognition server unit not only transmits the face recognition verification result, but also encrypts a verification data to the ATM, so that the ATM will receive The verification data of the face recognition server unit is transmitted to the host unit, and the host unit stores the verification data when successfully decrypting the verification data received from the ATM, wherein the verification The verification data includes the identification code of the specific customer data, a determination date when the face recognition verification result is generated, and a pair of determination serial numbers corresponding to the face recognition verification result. The transaction system according to claim 1, wherein the ATM further includes a data reading unit electrically connected to the processing unit, and before the ATM receives the cardless transaction initiation instruction: When it is detected that the data reading unit is electrically connected to a data carrier and receives an input password input by the user, it is determined whether the input password is consistent with a card password stored in the data carrier; the ATM determines After the input password matches the card password, use the living body identification model to judge whether the living body conditions are met, and send an image data generated by the shooting module to the face when the judgment result is yes Recognition server unit; the face recognition server unit obtains and stores the identification code of the specific customer data and a facial feature information related to the image data, and creates the identification code of the specific customer data and the Correspondence between facial feature information. The transaction system according to claim 5, wherein each customer data further includes a communication data, and after the ATM determines that the input password matches the card password, the host unit according to the specific customer data The communication data transmits a one-time password to a pair of portable electronic devices corresponding to the communication data, and a verification password input by the user at the ATM and consistent with the one-time password is received from the ATM At this time, a password is correctly notified to the ATM, and the ATM only judges whether the living condition is met when receiving the notification that the password is correct from the host unit. The transaction system according to claim 5, wherein, when the ATM determines that the input password matches the card password and receives a character string input by the user, generates and transmits a string containing the character string and indicating The cardless transaction password establishment request of the specific customer data is sent to the host unit, and the host unit adds the string as a cardless transaction password to the specific customer information when receiving the cardless transaction password establishment request. The transaction system according to claim 5, wherein: the ATM transmits multiple pieces of image data to the face recognition server unit when it judges that the living body condition is met; the face recognition server unit receives the images When data is selected, a target image data that meets an applicable condition is selected from the image data, and the target image data is transmitted to the ATM, wherein the applicable condition relates to the availability of each image data in face recognition. Recognition; when the ATM receives the target image data from the face recognition server unit, the target image data is displayed, and upon receiving a confirmation command input by the user and corresponding to the target image data, Causing the identification code of the specific customer data to be provided to the face recognition server unit and sending a file creation request to the face recognition server unit; when the face recognition server unit receives the specific customer data The identification code and the file creation request, the identification code of the specific customer data and a facial feature information related to the target image data are stored. An automatic teller machine suitable for communicating with a face recognition server unit and a host unit, the host unit stores multiple pieces of customer data corresponding to multiple customers, and each piece of customer data includes a corresponding one corresponding to the customer An identification code and at least one customer account, the ATM includes: a processing unit; and an image capturing module, electrically connected to the processing unit and capable of shooting under the control of the processing unit; wherein, the processing unit receives A cardless transaction initiation command and a user identification code input by the user send the user identification code to the host unit, so that the host unit selects from the customer data to include the user ID A specific customer data of the identification code; the processing unit generates and sends a face recognition request to the face recognition server unit when a living body recognition model determines that a living body condition is met, wherein the living body condition includes the There is a living body in the shooting range of the image shooting module, and the face recognition request includes an image data generated by shooting by the shooting module and the user identification code, and the face recognition request is used for the face The recognition server unit generates a face recognition verification result; after receiving the face recognition verification result and determining that the face recognition verification result indicates that the verification is passed, the processing unit obtains the at least the specific customer data from the host unit A customer account, and upon receiving a cardless transaction instruction related to a cardless transaction, one of the at least one customer account input by the user and indicating the specific customer information, executes the cardless transaction. The ATM according to claim 9, wherein the living body recognition model is pre-trained in a deep learning manner. The ATM as described in claim 9, wherein the processing unit transmits the user password to the user when the face recognition verification result indicates that the verification is passed and a user password input by the user is received The host unit, so that the host unit transmits the at least one client account number of the specific client information to the ATM when it determines that the user password matches a cardless transaction password included in the specific client information, and, The processing unit executes the card-free transaction with the host unit when the at least one client account number of the specific client data is received and the card-free transaction instruction is also received. A cardless transaction method suitable for execution by a transaction system including an automated teller machine, a face recognition server unit and a host unit, the automated teller machine includes an image capture module and an electrical connection to the image capture module The processing unit of the group, the image shooting module can be controlled by the processing unit to shoot, the host unit stores multiple pieces of customer data corresponding to multiple customers, and each piece of customer data includes a corresponding one corresponding to the customer An identification code and at least one customer account, the cardless transaction method includes: (A) the ATM receives the user identification code when receiving a cardless transaction initiation instruction and a user identification code input by the user Sent to the host unit, so that the host unit selects from the customer data a specific customer data containing the identification code that matches the user identification code; (B) the ATM uses a living body identification model to determine A face recognition request is generated and sent to the face recognition server unit when a living body condition is met, wherein the living body condition includes the presence of a living body in the shooting range of the image capture module, and the face recognition request includes a Image data generated by the shooting module and the user identification code; (C) when the face recognition server unit receives the face recognition request, according to the image data of the face recognition request and a A face feature information stored in advance and related to the user identification code generates a face recognition verification result, and transmits the face recognition verification result to the ATM; and (D) the ATM determines the face The identification verification result indicates that after verification is passed, the at least one client account number of the specific client data is obtained from the host unit, and one of the at least one client account number input by the user and indicating the specific client data is received And related to a cardless transaction instruction of a cardless transaction, the cardless transaction is executed. The cardless transaction method according to claim 12, wherein, in step (B), the living body identification model is pre-trained in a deep learning manner. The cardless transaction method as described in claim 12, wherein step (D) includes: (D1) the ATM determines that the face recognition verification result indicates that the verification is passed and also receives a usage input by the user User password, send the user password to the host unit; (D2) when the host unit receives the user password and determines that the user password matches a cardless transaction password included in the specific customer data , Transmitting the at least one customer account number of the specific customer profile to the ATM; and (D3) the ATM receives the at least one customer account number of the specific customer profile and also receives the cardless transaction instruction, and The host unit executes the card-free transaction. The cardless transaction method according to claim 12, wherein in step (C), the face recognition server unit not only transmits the face recognition verification result, but also encrypts the verification data to the ATM , So that the ATM transmits the verification data received from the face recognition server unit to the host unit, and causes the host unit to decrypt the verification data received from the ATM The verification data is stored, wherein the verification data includes the identification code of the specific customer data, a determination date when the face recognition verification result is generated, and a pair of determination serial numbers corresponding to the face recognition verification result. The cardless transaction method according to claim 12, wherein the ATM further includes a data reading unit electrically connected to the processing unit; the cardless transaction method further includes before step (A): (E) the When the ATM detects that the data reading unit is electrically connected to a data carrier and receives an input password input by the user, it determines whether the input password is consistent with a card password stored in the data carrier; (F ) After judging that the input password is consistent with the card password, the ATM uses the living body identification model to judge whether the living body conditions are met, and transmits a shot generated by the shooting module when the judgment result is yes Image data to the face recognition server unit; (G) the face recognition server unit obtains and stores the identification code of the specific customer data and a facial feature information related to the image data, and creates the specific The correspondence between the identification code of the customer data and the facial feature information. The cardless transaction method according to claim 16, wherein each customer profile further includes a communication profile, and the cardless transaction method further includes between steps (E) and (F): (H) the The host unit transmits a one-time password to a pair of portable electronic devices corresponding to the communication data according to the communication data of the specific customer data, and receives from the ATM a user input from the ATM and the When the verification password matches the one-time password, a correct password is notified to the ATM; in step (F), the ATM only judges the living condition when it receives the correct notification from the host unit Whether it meets. The cardless transaction method as described in claim 16, further comprising between steps (E) and (F): (I) when the ATM determines that the input password matches the card password and receives a message After the character string input by the user, generate and send a cardless transaction password creation request containing the string and indicating the specific customer data to the host unit; and (J) the host unit receives the cardless transaction password When creating the request, the string is added to the specific customer information as a cardless transaction password. The cardless transaction method according to claim 16, wherein, in step (F), the ATM transmits multiple pieces of image data to the face recognition server unit, and the cardless transaction method further includes steps Between (F) and (G): (K) The face recognition server unit selects a target image data that meets an applicable condition from the image data, and transmits the target image data to the ATM, Wherein, the applicable condition relates to the recognizability of each image data in face recognition; (L) when the ATM receives the target image data from the face recognition server unit, the target image data is displayed, and Upon receiving a confirmation command input by the user and corresponding to the target image data, the identification code of the specific customer data is provided to the face recognition server unit, and a file creation request is sent to the Face recognition server unit; (M) when the face recognition server unit receives the identification code of the specific customer data and the file creation request, the identification code of the specific customer data is stored and a Facial feature information of the target image data. A cardless transaction method suitable for execution by an ATM, which is suitable for communicating with a face recognition server unit and a host unit. The ATM includes an image capturing module and an electrical connection to the image capturing module The processing unit of the group, the image shooting module can be controlled by the processing unit to shoot, the host unit stores multiple pieces of customer data corresponding to multiple customers, and each piece of customer data includes a corresponding one corresponding to the customer An identification code and at least one customer account, the cardless transaction method includes: (A) the ATM receives the user identification code when receiving a cardless transaction initiation instruction and a user identification code input by the user Sent to the host unit, so that the host unit selects from the customer data a specific customer data containing the identification code that matches the user identification code; (B) the ATM uses a living body identification model to determine A face recognition request is generated and sent to the face recognition server unit when a living body condition is met, wherein the living body condition includes the presence of a living body in the shooting range of the image capture module, and the face recognition request includes a The image data generated by the shooting module and the user identification code, the face recognition request is used for the face recognition server unit to generate a face recognition verification result; and (C) the ATM receives the After the face recognition verification result and it is determined that the face recognition verification result indicates that the verification is passed, the at least one client account that obtains the specific client information from the host unit, and upon receiving a user input and indicating the When one of the at least one customer account of the specific customer profile is related to a cardless transaction instruction of a cardless transaction, the cardless transaction is executed. The cardless transaction method according to claim 20, wherein, in step (B), the living body identification model is pre-trained in a deep learning manner. The cardless transaction method of claim 20, wherein step (C) includes: (C1) the ATM determines that the face recognition verification result indicates that the verification is passed and receives a user input by the user Password, the user password is transmitted to the host unit, so that the host unit determines the at least the specific customer data when it determines that the user password matches a cardless transaction password included in the specific customer data A customer account number is transmitted to the ATM; and (C2) the ATM performs the card-less transaction with the host unit when receiving the at least one customer account of the specific customer information and also receiving the card-free transaction instruction.

Images