Light counter service method based on deep learning and intelligent security terminal

Abstract

The invention discloses a light counter service method based on deep learning and an intelligent security terminal. The light counter service method specifically comprises the steps of building a light counter service high-integration-level terminal; and in combination with remote teller business, full-delegation remote business is handled. In the aspect of business efficiency, the business process is simplified, and the time and energy cost of business handling are saved for customers; based on intelligent equipment and advanced technologies, the service processing speed and the overall service handling efficiency are greatly improved.

Description

Technical Field

[0001] This invention relates to the field of bank counter services, and more particularly to a lightweight counter service method based on deep learning and intelligent security terminals. Background Technology

[0002] Due to the current limitations of physical branches, elderly people, those in areas without branches, and other special customer groups who cannot reach branches find it inconvenient to go to bank branches and encounter certain difficulties in the process of handling business.

[0003] The physical equipment has a low level of intelligence, lacks the ability to automatically adjust and capture facial features, and is difficult to use without the guidance of bank staff to manually adjust the angle.

[0004] Traditional equipment lacks a device for automatically destroying vouchers. When a business transaction malfunctions, the vouchers cannot be automatically destroyed, compromising security and compliance, and making it unsuitable for unattended deployment of lightweight equipment.

[0005] Traditional equipment lacks a data self-destruct mechanism, meaning data security cannot be guaranteed in the event of equipment loss.

[0006] The existing system lacks electronic fence functionality, allowing business to continue even when devices are moved out of control, and data cannot be destroyed in a timely manner. This makes it unsuitable for the security and compliance of unattended deployment.

[0007] The existing system's remote whiteboard function makes it difficult for remote tellers to guide customers through transactions by writing and marking on the screen, thus hindering remote collaborative transactions.

[0008] The existing system lacks CV (computer vision) technology, making it unable to perform dynamic detection of real people and unable to prevent personnel changes during business processing.

[0009] The existing system lacks NLP (Natural Language Processing) technology and cannot support business processing such as dual recording, wealth management, and face-to-face signing for loan approval. Summary of the Invention

[0010] In view of the above problems, the present invention is proposed to provide a lightweight counter service method based on deep learning and intelligent security terminals to overcome or at least partially solve the above problems.

[0011] According to one aspect of the present invention, a lightweight counter service method based on deep learning and intelligent security terminals is provided, the lightweight counter service method specifically comprising:

[0012] Build a highly integrated, lightweight counter service terminal;

[0013] Combine remote teller services to handle fully entrusted remote business.

[0014] Optionally, the construction of the lightweight counter service highly integrated terminal specifically includes:

[0015] Using audio and video technology, a remote, face-to-face business processing model can be established between tellers and customers;

[0016] Both parties establish a video connection using camera and video streaming technology, and a remote voice connection using microphone and audio streaming technology.

[0017] Optionally, the construction of the lightweight counter service highly integrated terminal also includes:

[0018] When audio and video communication is obstructed, tellers can use a mouse to write, circle, and point on the customer's screen via remote whiteboard technology. Customers can then learn how to operate the system, write unfamiliar characters, and understand ambiguities by following the teller's instructions.

[0019] Optionally, the construction of the lightweight counter service highly integrated terminal also includes:

[0020] When a customer conducts a transaction, the camera captures the customer's current image and compares it with the archived photo or ID card photo for verification. If the comparison fails to match, the transaction is terminated.

[0021] If the consistency check passes, the CV technology is used to extract customer profile features. If the current person leaves midway and another person continues to process the business that the previous customer did not complete, photos are captured from the video stream.

[0022] The system uses CV technology to check whether the current person handling the transaction matches the profile features recorded at the start of the transaction. If they do not match, the transaction is terminated. If someone continues to handle the transaction, identity verification is performed again.

[0023] If someone uses a photo to impersonate another person to conduct business, CV technology is used to detect whether the person conducting the business in front of the camera is a real person. If the person is not a real person, a message will be displayed saying "Not a real person conducting business" and the business will be terminated.

[0024] Optionally, the construction of the lightweight counter service highly integrated terminal also includes:

[0025] In business transactions involving compliance with process evidence retention, customer verbal expressions are collected via microphone and transmitted to the back office.

[0026] The back-end uses NLP technology to process the customer's spoken voice and recognize the customer's semantics. If the customer's true intention is agreed and confirmed, the business will continue to be processed; otherwise, the business will be terminated.

[0027] Optionally, the construction of the lightweight counter service highly integrated terminal also includes:

[0028] When a user is involved in a core transaction, the system uses a binocular camera to perform biometric liveness detection on the entity and captures a photo, which is then sent to the backend for facial consistency comparison with the archived photo. If either biometric liveness detection or facial consistency comparison fails, the transaction is terminated.

[0029] Optionally, the construction of the lightweight counter service highly integrated terminal also includes:

[0030] In business guidance, processing, consultation, and recommendation activities, we rely on natural language processing, big data knowledge base, machine deep learning, and machine speech expression technologies to achieve human-computer interaction.

[0031] The terminal receives customer instructions and understands their intent; the transaction processing layer then translates the intent into specific transaction operations.

[0032] Optionally, the process of handling fully entrusted remote business in conjunction with remote teller services specifically includes:

[0033] Re-control credentials are automatically destroyed;

[0034] Before using the Beidou electronic fence, each device must be registered. During registration, the administrator sets the usable geographical area in the background and installs the Beidou positioning device. When the device leaves the geographical area set by the administrator, it will automatically change to an unusable state.

[0035] When the terminal communicates with the backend, sensitive data is encrypted using an encryption algorithm before transmission. The receiving party decrypts the data into plaintext using the agreed algorithm and public key.

[0036] This invention provides a lightweight counter service method based on deep learning and intelligent security terminals. The method specifically includes: building a highly integrated lightweight counter service terminal; and handling fully entrusted remote business in conjunction with remote teller services. In terms of business efficiency, it simplifies business processes, saving customers time and effort; and, relying on intelligent devices and advanced technologies, significantly improves business processing speed and overall business efficiency.

[0037] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0038] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0039] Figure 1 A flowchart illustrating a lightweight counter service method based on deep learning and intelligent security terminals, provided as an embodiment of the present invention. Detailed Implementation

[0040] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0041] The terms "comprising" and "having," and any variations thereof, in the specification, embodiments, claims, and drawings of this invention are intended to cover non-exclusive inclusion, such as including a series of steps or units.

[0042] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0043] like Figure 1 As shown, the present invention provides a lightweight counter service method based on deep learning and intelligent security terminals. The lightweight counter service method includes: building a highly integrated lightweight counter service terminal; and combining it with remote teller services to support fully delegated remote business processing.

[0044] This invention comprehensively utilizes technologies such as deep learning, liveness detection, BeiDou positioning, and data encryption to build a lightweight, highly integrated counter terminal device. Combined with remote teller services, it supports fully delegated remote business processing, meets complex business needs, and promotes the transformation of branches towards lightweight, interactive, and digital solutions. Simultaneously, it caters to the needs of the elderly by adding a magnifying glass, an eye tracker, and handrail functionality, thus constructing an age-friendly service ecosystem.

[0045] This invention innovatively applies digital technology to rural revitalization financial services, providing convenient, multi-dimensional, and cross-industry financial services. It reduces the cost of physical branches, extends financial services, further frees up branch manpower, and facilitates efficient operation and customer acquisition. While improving customer experience, it also reduces costs and increases efficiency.

[0046] Regarding service methods:

[0047] Application of audio and video technology: The traditional face-to-face business handling mode between tellers and customers is transformed into a remote face-to-face business handling mode. Both parties establish a video connection through camera and video streaming technology, and then conduct remote calls and exchanges through microphone and audio streaming technology, so that tellers can assist customers in handling business or entrust customers to handle business from thousands of miles away.

[0048] Application of remote whiteboard projection technology: During business transactions, when audio and video communication is hindered, remote tellers can use a mouse to write, draw, and point on the customer's screen via remote whiteboard projection technology. This allows customers to learn system operation, write unfamiliar characters, and understand ambiguous language based on the teller's prompts. This service model utilizes remote whiteboard projection technology.

[0049] CV (Verification by Image) technology is used when a customer begins a transaction. The system automatically captures the customer's current profile using the camera and compares it with a saved photo or ID photo for verification. If the comparison fails, the system automatically terminates the transaction. If the verification passes, the system uses CV technology to extract the customer's profile features. If the current customer leaves and someone else attempts to continue the transaction, the system captures a photo from the video stream within a very short time interval and uses CV technology to check if the current customer's profile features match the profile recorded at the start of the transaction. If they don't match, the transaction is automatically terminated. If someone continues the transaction, identity verification is required again. If someone uses someone else's photo to impersonate the customer, the system uses CV technology to check if the person in front of the camera is a real person. If not, the system displays "Not a real person conducting the transaction" and automatically terminates the transaction.

[0050] Application of NLP (Natural Language Processing) technology: In business transactions involving compliance with evidence retention, such as dual recording of financial records, the customer's spoken words are recorded by the microphone integrated into the device and transmitted to the back-end system. The back-end system then uses NLP technology to process the customer's spoken words and identify their semantics. If the customer's true consent and confirmation are obtained, the transaction can continue; otherwise, the transaction will be terminated. NLP ensures the confirmation of the user's true intention.

[0051] Application of liveness detection and facial recognition technology: When a user uses the system or is involved in a core transaction, the system calls a binocular camera to perform biometric liveness detection on the entity in front of the device and captures a photo, which is then sent to the backend system for facial consistency comparison with the archived photo. If either biometric liveness detection or facial consistency comparison fails, the business transaction is automatically terminated.

[0052] Application of instant messaging technology in web applications: During business processing, the information transmission between the remote teller system and the customer application system uses instant messaging technology (WebSocket) for push. For example, the customer application will push collected identity information, photos, bank card data, etc. to the remote teller. If either end pushes information and there is no response, it will push continuously. If there is still no response after two consecutive pushes, the system will consider this an anomaly as a network error and automatically terminate the business processing.

[0053] Application of AI digital human technology: In activities such as business guidance, processing, consultation, and recommendation, human-computer interaction is achieved by relying on technologies such as natural language processing, big data knowledge base, machine deep learning, and machine speech expression. The device receives customer instructions and understands their intentions, and the transaction processing layer transforms the intentions into specific transaction operations. For example: (1) Intelligent business consultation: First, the machine collects the customer's voice and identifies the semantics, searches for matching answers in the big data knowledge base, and then answers the customer's questions through machine speech expression. When the machine cannot answer the customer's questions, it will automatically learn the correct process of the customer's business, record and upgrade the knowledge base for subsequent guidance and guidance of other customers to handle similar businesses.

[0054] Automatic personalized service: The system calculates the customer's age based on the ID card information and the customer's archived identity information. For customers of different age groups, the system interface automatically adapts the system font, color, brightness, and volume.

[0055] Regarding security:

[0056] Automatic destruction of control vouchers: The device integrates a destruction unit. When a customer applies for a card, as the card is transferred from the card box to the device's read / write position, the system retrieves the card data from the backend system and performs read / write operations on the card. If the data writing fails or only part of the data is written, the system will automatically verify the integrity of the card data and the integrity of the business process. If an anomaly is found in the verification, the system will automatically transfer the card to the destruction unit and send an instruction to physically drill holes in the card in the current business to destroy the chip.

[0057] Beidou electronic fence: Each device requires registration before use. During registration, the device administrator sets the usable geographical area in the backend system. A Beidou positioning device is installed in the device. When the device leaves the geographical area set by the administrator, the system on the device automatically changes to an unavailable state. If the device leaves the designated geographical area for the set time limit, the system considers the device lost and will automatically initiate a full data self-destruction, destroying all data on the device.

[0058] Data encryption: When the device system and the backend system communicate, both parties use encryption algorithms to encrypt sensitive data before transmission. The receiving party decrypts the data into plaintext using the agreed algorithm and public key. If the system-generated data needs to be archived, it is encrypted using an encryption algorithm. The archived data can only be viewed using a specific tool with decryption capabilities.

[0059] Privacy Protection: The business operation screen configured on the equipment is a privacy screen; the screen content is not visible from the side of the equipment or from a non-effective distance, effectively protecting customer privacy. Password input devices are equipped with privacy shields to allow customers to input their passwords discreetly.

[0060] In summary, the system integrates CV (human detection, face-swapping detection), NLP (natural language processing), liveness detection and facial recognition, big data knowledge bases, machine deep learning, instant messaging technology, AI digital humans, lightweight devices, smart tellers, remote video, and fully delegated business processing models to achieve comprehensive application of diversified service channels. The lightweight devices can be deployed in nursing homes, communities, and industrial parks, enabling cross-ecosystem cooperation and expanding the service scope.

[0061] In terms of business efficiency, we simplify business processes to save customers time and effort in handling business; and rely on smart devices and advanced technologies to significantly improve business processing speed and overall business efficiency.

[0062] In terms of service scope, the focus is on the elderly and special customer groups, precisely meeting their specific needs and effectively improving the accessibility of financial services. At the same time, it covers areas without branches and customer groups inaccessible by branch network, contributing to rural revitalization and inclusive finance.

[0063] In terms of security, a security system is built from multiple dimensions, including physical security, information security, network security and identity security, to ensure that secure terminals are not damaged or illegally intruded upon from the outside, and data encryption technology is used to strengthen information security protection to prevent customer information from being leaked and data from being tampered with.

[0064] Beneficial effects:

[0065] 1. Utilize technologies such as computer vision (CV), natural language processing (NLP), large-scale language modeling (LLM), augmented reality (AR), and deep learning to achieve rapid response to customer needs, enhance international service capabilities, and improve the accessibility of financial services across various financial business scenarios, including customer identity verification, financial consulting services, transaction processing, customer service, risk assessment, cross-language services, and barrier-free services.

[0066] 2. By using liveness detection technology, we can accurately capture facial images, extract and analyze facial biometric features, determine whether specific actions performed by users meet the requirements, ensure that financial transactions are conducted in accordance with customer wishes, prevent the risk of illegal operations, and improve the identification capabilities of tellers and customers.

[0067] 3. Utilizing technologies such as BeiDou positioning, time synchronization, differential positioning, and multi-mode reception, it achieves real-time uninterrupted positioning and trajectory monitoring of the equipment. Even after the equipment is powered off, it supports continuous uploading of location coordinate information for up to 24 hours. The hardware module uploads BeiDou coordinates, and the background monitoring system can provide functions such as electronic fences. It has a wider positioning area coverage, stronger anti-interference capabilities, and includes rescue communication functions, thereby improving the level of intelligent equipment management.

[0068] 4. Employing data encryption technology, based on the onboard TCM security chip, it provides trusted secure boot and data integrity verification, supports encrypted communication, anti-tampering protection, and privacy protection, ensuring that the entire process from system startup to operating system loading is not tampered with, increasing the confidentiality of counter terminal data, and comprehensively improving the security and reliability of the equipment.

[0069] The above specific embodiments further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A lightweight counter service method based on deep learning and intelligent security terminals, characterized in that, The lightweight counter service method specifically includes: Build a highly integrated, lightweight counter service terminal; Combine remote teller services to handle fully entrusted remote business.

2. The lightweight counter service method based on deep learning and intelligent security terminal according to claim 1, characterized in that, The construction of a highly integrated, lightweight counter service terminal specifically includes: Using audio and video technology, a remote, face-to-face business processing model can be established between tellers and customers; Both parties establish a video connection using camera and video streaming technology, and a remote voice connection using microphone and audio streaming technology.

3. The lightweight counter service method based on deep learning and intelligent security terminal according to claim 1, characterized in that, The construction of a highly integrated, lightweight counter service terminal also includes: When audio and video communication is obstructed, tellers can use a mouse to write, circle, and point on the customer's screen via remote whiteboard technology. Customers can then learn how to operate the system, write unfamiliar characters, and understand ambiguities by following the teller's instructions.

4. The lightweight counter service method based on deep learning and intelligent security terminal according to claim 1, characterized in that, The construction of a highly integrated, lightweight counter service terminal also includes: When a customer conducts a transaction, the camera captures the customer's current image and compares it with the archived photo or ID card photo for verification. If the comparison fails to match, the transaction is terminated. If the consistency check passes, the CV technology is used to extract customer profile features. If the current person leaves midway and another person continues to process the business that the previous customer did not complete, photos are captured from the video stream. The system uses CV technology to check whether the current person handling the transaction matches the profile features recorded at the start of the transaction. If they do not match, the transaction is terminated. If someone continues to handle the transaction, identity verification is performed again. If someone uses a photo to impersonate another person to conduct business, CV technology is used to detect whether the person conducting the business in front of the camera is a real person. If the person is not a real person, a message will be displayed saying "Not a real person conducting business" and the business will be terminated.

5. A lightweight counter service method based on deep learning and intelligent security terminals according to claim 1, characterized in that, The construction of a highly integrated, lightweight counter service terminal also includes: In business transactions involving compliance with process evidence retention, customer verbal expressions are collected via microphone and transmitted to the back office. The back-end uses NLP technology to process the customer's spoken voice and recognize the customer's semantics. If the customer's true intention is agreed and confirmed, the business will continue to be processed; otherwise, the business will be terminated.

6. The lightweight counter service method based on deep learning and intelligent security terminal according to claim 1, characterized in that, The construction of a highly integrated, lightweight counter service terminal also includes: When a user is involved in a core transaction, the system uses a binocular camera to perform biometric liveness detection on the entity and captures a photo, which is then sent to the backend for facial consistency comparison with the archived photo. If either biometric liveness detection or facial consistency comparison fails, the transaction is terminated.

7. A lightweight counter service method based on deep learning and intelligent security terminals according to claim 1, characterized in that, The construction of a highly integrated, lightweight counter service terminal also includes: In business guidance, processing, consultation, and recommendation activities, we rely on natural language processing, big data knowledge base, machine deep learning, and machine speech expression technologies to achieve human-computer interaction. The terminal receives customer instructions and understands their intent; the transaction processing layer then translates the intent into specific transaction operations.

8. A lightweight counter service method based on deep learning and intelligent security terminals according to claim 1, characterized in that, The aforementioned integration of remote teller services with fully entrusted remote services specifically includes: Re-control credentials are automatically destroyed; Before using the Beidou electronic fence, each device must be registered. During registration, the administrator sets the usable geographical area in the background and installs the Beidou positioning device. When the device leaves the geographical area set by the administrator, it will automatically change to an unusable state. When the terminal communicates with the backend, sensitive data is encrypted using an encryption algorithm before transmission. The receiving party decrypts the data into plaintext using the agreed algorithm and public key.

Images