An address authentication and document verification system

Abstract

The present invention provides a system and method for customer identification and address verification during onboarding, integrating multi-layered authentication mechanisms to enhance security and prevent fraud. The method for customer identification and address verification during onboarding involves a multi-step authentication process integrating geolocation validation, triggering the document verification, and facial recognition. Additionally, the system includes client executive devices, field devices, and a server unit that work in coordination to verify customer identity and location in real time. This AI-driven, automated identity verification system and method ensures fast, accurate, and highly secure customer onboarding, making it ideal for banks, financial institutions, and fintech companies.

Description

[0001] Title -AN ADDRESS AUTHENTICATION AND DOCUMENT VERIFICATION SYSTEM

[0002] FIELD OF INVENTION

[0003] The present invention relates to a system and method for customer identification and address verification. It enables secure and accurate authentication by integrating digital document verification, geolocation-based address validation, facial recognition, and OTP-based authentication. The invention enhances verification efficiency, reduces fraud, and ensures compliance with regulatory requirements.

[0004] BACKGROUND OF THE INVENTION

[0005] Various industries, including banking, finance, and e-commerce, require identity verification and document authentication during customer onboarding to ensure transaction legitimacy and prevent fraud. Document verification involves reviewing identity documents to confirm an individual’s identity, ensuring compliance with legal and regulatory requirements across industries like banking, finance, logistics, and e-commerce.

[0006] The different types of Verification services such as Know Your Customer (KYC), Intelligent Document Recognization (IDR), BULK verification, Customer Due Diligence (CDD), DOCUMENT FRAUD CHECK, E-Residence Physical Verification (ERPV) etc. KYC is crucial in the financial industry to prevent illegal activities, protect customer data, and build trust. The main purpose of KYC is to prevent fraudulent activities, such as money laundering, identity theft, terrorist financing and other forms of financial crimes. KYC verifies the customer related documents. Intelligent Document Recognization (IDR) consists of documents to be validated like driving licence, voter id, etc. The service itself is intelligent as it uses a combination of data available from the authenticated sources plus document intelligence such as compression artefact technology to verify if the document itself is genuine or fraudulent. Bulk verification of documents refers to the process of verifying a large number of documents or records simultaneously or in bulk, as opposed to verifying them one by one. This process is commonly used in various industries, including finance, healthcare, and customer onboarding, to efficiently and accurately confirm the authenticity and validity of a significant volume of documents, such as identity documents, contracts, invoices, or any other paper work that requires verification.

[0007] Customer Due Diligence (CDD) assesses customer risk based on location, occupation, and transaction patterns, helping financial institutions understand their background and activities. DOCUMENT FRAUD CHECK consist of PANCARD or other such documents to be validated. Once this service is registered by the client executive then the client field executive can upload the PANCARD or other such documents to be validated by face matching, template matching, data matching etc., and the validated results will be available to download by the client field executive or any other authenticated source within the institution. This service ensures that document trust is validated and spoofing of documents by tools such as photoshop cannot take place. Lastly, the ERPV (Enhanced Real-time Physical Verification) is a process used to validate physical presence, identity, or information in real-time through automated data collection and verification methods. It is applied across various industries to enhance accuracy, security, and operational efficiency by reducing reliance on manual verification.

[0008] Other verification services includes the documents like E-Bill, Voter, AadharCard, Driving License, Corporate Verification, MSME, Shop and establishment verification, UAN, etc.,

[0009] Ongoing Monitoring involves continuously tracking customer activities to detect suspicious behavior and ensure compliance with regulations. Document verification is the process of validating identity or official documents to confirm their authenticity and accuracy. It is widely used across industries to ensure compliance, prevent fraud, and verify customer information. This process may involve manual inspection, database crossreferencing, or automated technologies for enhanced security and efficiency.

[0010] The general process for document verification includes verifying authenticity, checking security features such as watermarks and holograms, and detecting tampering or forgery. Key details such as name, date of birth, and photographs are validated against the provided information.

[0011] The process includes checking authenticity, security features, watermarks, holograms, and detecting tampering or forgery. Details like name, date of birth, and photographs are matched against provided data, and expired documents are flagged. Verification is further strengthened by cross-referencing with government databases or other trusted sources to prevent identity theft, fraud, and money laundering. It ensures compliance with legal and regulatory requirements during customer onboarding. The prior art IN202021006129A discloses a method for digitally onboarding customers by processing user details and verifying them before granting access. It focuses on decisionmaking for onboarding. However, it does not disclose registering customer details from an authorized client executive device for initiating a verification request and uploading documents. It does not describe storing these details for authentication, transmitting verification requests to a field device, or capturing and submitting images of physical surroundings along with geographical location data for address validation. Additionally, it does not include capturing and submitting a live facial image, and compiling a final verification report integrating all verification results.

[0012] The prior art WO2023119318A2 describes a financial verification method that enables digital onboarding using authentication techniques. It employs multi-factor authentication and document cross-referencing but does not incorporate detailed address validation through geolocation. However, it does not disclose registering customer details from an authorized client executive device for verification request initiation and document uploads. It does not describe storing these details for authentication, triggering verification through a field device, or capturing images of physical objects along with location data to confirm an address. Additionally, it does not describe capturing and submitting a live facial image, and generating a final verification report consolidating location, document, and facial verification results.

[0013] The prior art IN202111024004A explains a method for user onboarding based on electronic document verification. It validates identity through user-submitted documents and OTP-based authentication. However, it does not disclose registering customer details from an authorized client executive device to trigger a verification request and allow document submission. It does not describe storing registered details for authentication retrieval, sending verification requests to a field device, or capturing and submitting images of surrounding objects with geographical location data for address confirmation.

[0014] The prior art US20150341370A1 details a method for verifying identity by processing user- submitted documents and cross-referencing them with official databases. It relies on document attributes for identity verification. However, it does not disclose registering customer details from an authorized client executive device to generate a verification request and enable document uploads. It does not describe storing registered details for authentication retrieval, transmitting verification requests to a field device, or submitting images of physical objects and associated location data for address validation. The method lacks matching submitted location data with stored addresses, converting address components into geographical coordinates, refining geolocation accuracy. Additionally, it does not include capturing and submitting a live facial image, and generating a final verification report consolidating geolocation, document validation, and facial authentication.

[0015] The prior art US20210075788A1 discloses a method for verifying identity through document processing and external database validation. It extracts document details for authentication. However, it does not disclose registering customer details from an authorized client executive device to generate a verification request and enable document uploads. It does not describe storing registered details for authentication, sending verification requests to a field device, or capturing and submitting images of surrounding objects with location data to validate addresses. Additionally, it does not describe capturing and submitting a live facial image, and generating a final verification report consolidating geolocation, document validation, and facial recognition results.

[0016] Currently, document validation relies on manual processes, making customer onboarding timeconsuming and prone to errors. Existing methods do not incorporate real-time geolocationbased address verification, facial data authentication, and document validation using Optical Character Recognition (OCR), leading to inefficiencies and security vulnerabilities. The invention addresses these challenges based on E-Residence Physical Verification (ERPV) which is a physical residential address verification service offered in the present invention. It verifies the address of the customer based on geolocation-based address validation, with OCR- driven document authentication. The solution also perform facial comparison of the customer between Official IDs, and live picture of the customer. This multi-layered customer identification and address verification solution enhances accuracy, security, and efficiency in customer onboarding.

[0017] OBJECTS OF THE INVENTION

[0018] The principal objective of the invention is to provide a geolocation-based address authentication and document verification system. More specifically, the invention relates to a system and method that integrates geolocation validation for address verification, document verification using Optical Character Recognition (OCR), facial data extraction and comparison, and OTP-based authentication to enhance security, accuracy, and efficiency in customer onboarding.

[0019] Another objective of the invention is to assist organizations, banks, insurance companies, NBFCs, fintech firms, and other allied industries in controlling fraud during customer onboarding, investigations, money lending, and service provision by integrating geolocationbased address verification, OCR-based document verification, facial data-based authentication, and OTP-based authentication for enhanced security, accuracy, and compliance.

[0020] Said and other objects of the present disclosure will be apparent to a person skilled in the art after consideration of the following summary of subject matter as claimed, detailed description taken into consideration with accompanying drawings in which preferred embodiments of the present disclosure are illustrated.

[0021] SUMMARY OF THE INVENTION

[0022] It is therefore a general aspect of the embodiments to provide a system and method for customer identification and address verification during onboarding. More specifically, the embodiments provide a structured verification framework integrating geolocation-based address validation, OCR-driven document authentication, biometric facial comparison, and OTP-based authentication. The system enhances fraud prevention, regulatory compliance, and onboarding efficiency by automating verification using real-time customer location analysis, identity document validation, and multi-factor authentication mechanisms.

[0023] A method for customer identification and address verification during onboarding is provided, comprising registering customer details from an authorized client executive device to a server unit, storing the details in a memory unit for retrieval and authentication, generating and transmitting a verification request to at least one field device, submitting an image of physical objects along with geographical location data, verifying the customer’s geographical location by comparing the submitted data with a registered address, converting the address into longitude and latitude coordinates, segmenting the textual address into multiple components, iteratively removing each segment from narrowest to broadest, identifying multiple possible location coordinates, performing a coordinate similarity check, capturing and submitting a live facial image, comparing the submitted geographical location coordinates with the derived coordinates from an external database, matching the facial image with facial templates extracted from identification documents, and generating a verification report based on geolocation verification, document validation, and facial recognition to establish customer identity and address authenticity.

[0024] According to an embodiment, the method further comprises: computing the shortest distance between the location of the customer device and the registered customer address; comparing the computed distance against a predefined proximity threshold for geographical location validation; and incorporating the geographical location verification results into the verification report.

[0025] According to an embodiment, the method further comprises: receiving and extracting textual and image-based data from customer-uploaded identification documents and geographical coordinates using an Optical Character Recognition (OCR) model selected from Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), or Transformer-Based (LLM) Models; extracting key document attributes, including name, date of birth, address, and document number, from the uploaded identification documents and geographical coordinates; comparing the extracted OCR text with official documents stored in an external database to confirm customer authenticity and address validation; and generating a validation report incorporating the document verification result and address validation results.

[0026] According to an embodiment, the method further comprises: extracting facial features from a live image using a Convolutional Neural Network (CNN); converting the extracted facial features into numerical vector representations using a neural network; comparing the generated numerical vectors with facial templates stored in an external database by applying one or more similarity distance metrics; assigning a face match score based on the computed similarity metrics and evaluating the score against a predefined threshold value; and triangulating the face match score by integrating geographical location validation results and document verification results in the verification report generated by the server unit.

[0027] According to an embodiment, the method further comprises: comparing the generated numerical vectors with stored facial templates in the external database using one or more similarity distance metrics, including cosine similarity to measure the angle between multiple numerical vectors to determine similarity, Euclidean distance to measure the straight-line distance between multiple numerical vectors, and L2-Norm to normalize the vector magnitude and measure similarity.

[0028] According to an embodiment, the method further comprises: transmitting the verification request from the authorized client executive device to the server unit to initiate customer authentication; generating a verification link by the server unit and transmitting it to at least one field device; accessing the verification link by at least one field device to establish communication with the server unit; generating a One-Time Password (OTP) from the server unit and sending it to the customer device upon accessing the verification link by at least one field device; and submitting the OTP received by the customer device, along with geographical location data, to the server unit for customer authentication.

[0029] According to an embodiment, a system for customer identification and validation for onboarding a customer is provided, comprising: a client executive device configured to register customer details, including name, contact number, and address, to a server unit for triggering a verification request and enabling customer document uploading, and transmitting the registered customer details to the server unit for authentication and retrieval; a server unit configured to store the registered customer details in a memory unit for identity verification, generate and transmit a verification request to at least one field device, and receive and process an image of physical objects provided by the customer or present in the vicinity of the customer's location, along with geographical location data, including a live image of customer identification documents for verifying the geographical location of the customer; a geographical location verification module executed by the server unit, configured to compare the submitted geographical location data with the address registered in an external database, convert the address into longitude and latitude coordinates, segment the textual address into multiple location components, remove each segment iteratively, identify multiple possible location coordinates, and perform a coordinate similarity check to determine the most probable address location; a facial verification module executed by the server unit, configured to capture and submit a live facial image from the customer device for face verification of the customer; and an identity and address verification module executed by the server unit, configured to compare the submitted geographical location coordinates from the customer device with the derived geographical location coordinates obtained from the external database, match the customer identification documents and facial image with facial templates extracted from stored identification documents, and generate a verification report incorporating results from geolocation verification, document validation, and facial recognition to establish the authenticity of the customer's identity and address verification.

[0030] According to an embodiment, the system further comprises: a customer device configured to capture and upload one or more customer identification documents, and transmit the captured document images and geographical coordinates to the server unit for processing; a server unit communicatively connected to an external database and configured to execute a document verification module, including receiving and processing customer- uploaded identification documents and geographical coordinates, extracting textual and image-based data from the uploaded documents using an Optical Character Recognition (OCR) model selected from Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), or Transformer- Based (LLM) Models, identifying and extracting key document attributes, including name, date of birth, address, and document number, from the customer-uploaded identification documents and geographical coordinates, comparing the extracted OCR text with official documents stored in the external database to confirm customer authenticity and validate the customer’s residential address, and generating a verification report incorporating the document verification and address validation results.

[0031] According to an embodiment, the system further comprises: a customer device configured to capture a live facial image of the customer for identity authentication and transmit the captured facial image to a server unit for processing; a server unit, communicatively connected to an external database, configured to execute a facial verification module, including extracting facial features from the received image using a Convolutional Neural Network (CNN) model, converting the extracted facial features into numerical vector representations using a neural network-based facial recognition model, comparing the generated numerical vectors with facial templates stored in the external database by applying at least one similarity distance metric, assigning a face match score based on the computed similarity metrics and evaluating the score against a predefined threshold value, and triangulating the face match score by integrating geolocation validation and document verification in the verification report generated by the server unit.

[0032] According to an embodiment, the system further comprises: an OTP-based authentication module, including a client executive device configured to transmit a verification request to a server unit to initiate customer authentication; a server unit, communicatively connected to a memory unit, configured to execute the OTP-based authentication module, including generating a verification link upon receiving the verification request, transmitting the generated verification link to at least one field device for initiating authentication, enabling at least one field device to access the verification link and start communication with the server unit, generating a One-Time Password (OTP) from the server unit and sending it to the customer device upon accessing the verification link by at least one field device, and submitting the OTP received by the customer device, along with geographical location data, to the server unit for customer authentication.

[0033] BRIEF DESCRIPTION OF DRAWINGS

[0034] It is to be noted, however, that the appended drawings illustrate only typical embodiments of the present subject matter and are therefore not to be considered for limiting its scope, for the invention may admit to other equally effective embodiments. The detailed description is described with reference to the accompanying figures. In the figures, a reference number identifies the figure in the reference number first appears. The same numbers are used throughout the figures to reference features and components. Some embodiments of system, method, or structure in accordance with embodiments of the present subject matter are now described, by way of example, and with reference to the accompanying figures, in which:

[0035] Figure 1 illustrates the customer identification and address verification process for onboarding a customer, according to the present invention.

[0036] Figure 2 provides a step-by-step representation of the customer identification document verification process using Optical Character Recognition (OCR).

[0037] Figure 3 illustrates the facial verification process for authenticating customer identity, according to the present invention.

[0038] Figure 4 demonstrates the OTP-based customer authentication process used for verifying customer identity and preventing fraud, according to the present invention.

[0039] Figure 5 provides an overview of the system architecture for customer Identification and address verification, according to the present invention.

[0040] The figures depict embodiments of the present subject matter for the purposes of illustration only. A person skilled in the art will easily recognize from the following description that alternative embodiments of the device and process illustrated herein may be employed without departing from the principles of the disclosure described herein. DETAILED DESCRIPTION

[0041] The best and other modes for carrying out the present invention are presented in terms of the embodiments, herein depicted in drawings provided. The embodiments are described herein for illustrative purposes and are subject to many variations. It is understood that various omissions and substitutions of equivalents are contemplated as circumstances may suggest or render expedient, but are intended to cover the application or implementation without departing from the spirit or scope of the present invention. Further, it is to be understood that the phraseology and terminology employed herein are for the purpose of the description and should not be regarded as limiting. Any heading utilized within this description is for convenience only and has no legal or limiting effect.

[0042] The terms “a” and “an” herein do not denote a limitation of quantity, but rather denote the presence of at least one of the referenced items.

[0043] The terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process or method that comprises a list of steps does not include only those steps but may include other steps not expressly listed or inherent to such a process or method. Similarly, one or more sub-systems or elements or structures or components preceded by "comprises... a" does not, without more constraints, preclude the existence of other, sub-systems, elements, structures, components, additional sub-systems, additional elements, additional structures or additional components. Appearances of the phrase "in an embodiment", "in another embodiment" and similar language throughout this specification may, but not necessarily do, all refer to the same embodiment.

[0044] For the purpose of promoting an understanding of the principles of the invention, reference will now be made to the embodiment illustrated in the figures and specific language will be used to describe them. It will nevertheless be understood that no limitation of the scope of the invention is thereby intended. Such alterations and further modifications in the illustrated system, and such further applications of the principles of the invention as would normally occur to those skilled in the art are to be construed as being within the scope of the present invention.

[0045] It will be understood by those skilled in the art that the foregoing general description and the following detailed description are exemplary and explanatory of the invention and are not intended to be restrictive thereof. Unless otherwise defined, all technical and scientific terms used herein have the same meanings as commonly understood by those skilled in the art to which this invention belongs.

[0046] The system, method and examples provided herein are only illustrative and not intended to be limiting. Embodiments of the present invention will be described below in detail with reference to the accompanying figures.

[0047] The present invention provides a method (100) and system (500) for customer identification and address verification during customer onboarding. The invention focuses on enhancing authentication and verification processes using geolocation validation, document verification, facial recognition, and OTP-based authentication. The solution ensures secure, automated, and real-time verification of a customer’s identity and residential address.

[0048] The invention is a machine-intelligence driven risk-assessment tool useful for banks, financial institutions, insurance companies, fintech firms, and other organizations that require a reliable identity verification mechanism for their customers. Through the present invention, it ensure that the customer data generally-accepted, and documents can be verified securely and seamlessly. However, the instant solution itself need not be limited to documents from a particular geography / country.

[0049] Fig. 1 illustrate a multi-layered verification process (100) for onboarding a customer. The solution integrate multiple verification layers — including document authentication, geolocation matching, and face verification, to minimizes fraud risks while ensuring compliance with regulatory requirements. This multi-layered verification process (100) significantly improves the accuracy and reliability of the onboarding process, reducing reliance on manual verification while increasing the speed and efficiency of customer authentication.

[0050] Fig. 5 shows a system (500) having an admin device (501), one or more client executive devices (502), client field executive devices (504) and customer devices (505) for multi-layered approach during customer onboarding. The admin device (501) allows the client executive devices (502) to access the platform. The client executive device (502) is operated by a bank official or an equivalent institution representative, responsible for initiating and overseeing the customer onboarding process. The customer device (505) is a mobile or computing device held by the customer undergoing verification. This device is essential in the identity authentication and address verification process, as it enables the customer to actively participate in the onboarding process. The client field executive device (504) is operated by a field official of the bank responsible for conducting on-site verification of the customer’s identity and address. This device plays a crucial role in establishing communication with the customer device and the server unit. A field device (504) refers to any device used for conducting identity and address verification in real-time. Both customer device (505) and client field executive device (504) act as field device (503). By categorizing both customer device (505) and client field executive device (504) under field device (503), the system (500) ensures flexibility in authentication, allowing either the customer to verify remotely or the field executive to perform an in-person verification, thereby enhancing security, accuracy, and fraud prevention in the customer onboarding process.

[0051] The process (100) begins with registering a login authorization user list via the admin device, allowing client executive devices to access the verification platform. Once the login authorization is established for the client executive device, this device initiates the registration

[0052] (101) of customer details, including name, contact number, and residential address with the type of documents required for customer registration (e.g., PAN card, Aadhar card, Electricity Bill, Passport, Driving License). Physical proof of documents are required for verification. The device transmit the data to a server unit for customer registration. The customer data is stored

[0053] (102) in a memory unit for retrieval and authentication. This ensure that the registered details are available for cross-verification at later stages. Once the customer details are registered, the server unit generates a verification request (103), which is transmitted to a field device (either a client field executive device or a customer device) to initiate the authentication process.

[0054] As shown in fig. 4, upon registering the verification request (103), the server unit generates and transmits a verification link (401) to at least one of the field device, to initiate customer authentication. Upon accessing the verification link (402) by the field device and start communicating with the server unit. Here, the field device can be client field executive device or the customer device. Then, the server unit sends an One-Time Password (OTP) (404) to only customer device for authentication. The OTP is submitted (405) to the server, enabling login access to the platform for the field device, provided that the device’s location tracking is set to high accuracy for validating the geographical location relative to the customer's registered address.

[0055] Following successful customer authentication, the customer device captures and submits (104) images of physical objects in and around the vicinity of the customer, such as government- issued identification documents, house exterior images, utility meters, street signs, or property- related objects etc. along with real-time geographical location data. These elements are used to verify if the customer is present at the registered address. The submitted geographical location data is compared (105) with the registered address stored in the external database. If a match is found, the system converts the address into latitude and longitude coordinates for further validation.

[0056] To refine address verification, the system segments (106) the textual address into multiple location components, including house number, street name, locality, district, state, and postal code. It then performs iterative removal (107) of address components, starting from the most specific (house number) to the most general (state or country), to identify possible geolocation coordinates and validate address proximity. A coordinate similarity check is performed to determine the most probable address location (108), ensuring accurate verification.

[0057] In an embodiment, the method includes computing the shortest distance between the customer's actual location and the officially registered address to determine the great-circle distance between two points on the Earth’s surface based on their latitude and longitude values. Once the distance is computed, the system compares the result against a predefined proximity threshold to assess whether the customer is physically present at or near the registered address. If the computed distance falls within the acceptable range, the address is considered valid; however, if the deviation exceeds the threshold, the system may flag the verification attempt as potentially fraudulent or request additional proof of residence.

[0058] For example: if the given address is "Flat No. 12B, Green Heights Apartments, Sector 45, Gurgaon, Haryana, India - 122003," the system first checks for an exact match in an external database. If found, it retrieves the latitude and longitude coordinates for verification. If no exact match is available, the system iteratively removes address components, starting with the most specific details. Initially, it removes the flat number (12B) and checks if the apartment complex (Green Heights Apartments, Sector 45, Gurgaon) has a valid location. If verification at this level fails, the system further removes the apartment name and attempts to match at the sector level within Gurgaon, using the postal code (122003) to refine accuracy. If multiple locations exist for the same sector name, the system prioritizes matches with the registered city (Gurgaon) and state (Haryana). Further, the system removes Sector 45, attempting to validate the broader city-level location (Gurgaon, Haryana, India - 122003). If verification remains uncertain, it checks if the state and postal code combination can refine the results, ensuring that the location is mapped correctly. After identifying multiple possible geolocation coordinates, the system conducts a coordinate similarity check, comparing the customer's submitted latitude / longitude data with the database-derived coordinates. A distance deviation determines whether the submitted location falls within an acceptable threshold. If the deviation is within limits, the system validates the address; otherwise, it flags potential fraud. This iterative refinement determines the most probable address location and ensures greater accuracy in geolocation validation, minimizing fraud risks.

[0059] In an embodiment, the acceptable threshold limit in present solution is 30 km.

[0060] In addition to address validation, facial recognition (as shown in Fig. 3) is employed for identity authentication. The customer is also required to capture and submit a live facial image (109), which is analyzed using OCR (Optical Character Recognition) works based on machine learning (ML) and deep learning (DL) techniques. The extracted facial features (301) are compared with pre-stored official documents in external database, ensuring that the person undergoing verification is the same as the one in the official records. The machine learning based OCR uses feature extraction with classical ML algorithms (SVM, KNN, Random Forest, etc.). Further, Deep learning based OCR Uses neural networks for automatic feature extraction and recognition. Here the popular architecture are CNN (Convolutional Neural Networks) - For feature extraction from text images, RNN (Recurrent Neural Networks) / LSTMs - For sequence prediction and context understanding and Transformer-Based (LLM) Models used for text-processing. Lastly, End-to-End OCR with Deep Learning can also be used. This approach uses CNN + RNN + CTC Loss (Connectionist Temporal Classification) with Lightweight LLMs to recognize text directly from images.

[0061] Further, in face recognition process of present solution, feature extraction (301) and embedding comparison is using a Convolutional Neural Networks (CNNs) in deep learning techniques. Here, the captured live images first processed for face detection and alignment using models such as MTCNN (Multi-task Cascaded Convolutional Networks) or RetinaFace. Once the face is detected, a Convolutional Neural Network (CNN) extracts unique facial features, which are then converted into numerical vector representations (302) known as embeddings. These embeddings serve as unique identifiers for a person's face. The system then compares (303) the extracted embeddings with pre-stored facial templates obtained from government-issued IDs in external databases for hygiene check.

[0062] In an embodiment, the comparison (303) between numerical vectors of the customer’s live facial image with pre-stored facial templates is performed, using one or more similarity distance metrics, including Cosine Similarity, Euclidean Distance, and L2-Norm-based Matching, each offering distinct ways to measure the similarity between numerical facial embeddings. Cosine Similarity measures the angular difference between two numerical vectors (facial embeddings) rather than their absolute distance. It evaluates how closely two vectors align in the multidimensional space. A higher similarity score indicates a stronger match, whereas a lower score suggests dissimilarity. Euclidean Distance measures the direct spatial difference between two numerical vectors by calculating the straight-line distance in n- dimensional space. A lower Euclidean Distance value indicates a higher similarity, meaning the facial images belong to the same person. The L2-Norm (also known as the Euclidean Norm) normalizes numerical vectors before measuring similarity. By normalizing vectors before applying distance metrics, L2-Norm ensures better consistency in similarity comparisons, improving accuracy in real- world facial recognition applications.

[0063] In facial recognition (109), the distance between facial embeddings determines the strength of a match which is determined using distance metrics. A lower numerical distance between two embeddings signifies higher similarity, meaning the faces belong to the same individual. If the distance is above a predefined threshold (304), it is classifying the face as a non-match. Also, the angular difference is calculated using distance metrics, between two embeddings. A higher similarity score of face indicates a stronger match, confirming that the facial features are highly correlated. Lastly, a thresholding mechanism ensures high verification accuracy. If the similarity score exceeds the threshold (or the distance falls below a defined limit), the solution validates the customer’s identity and triangulate (305) the face match score by integrating geographical location validation result and document verification results in verification report. Otherwise, additional authentication steps, such as manual review or alternate verification methods, may be triggered or verification may be considered as fraud. Further, image of physical identification document (as shown in fig. 2) of customer is submitted along with real-time geographical location data. The process (104) begins with the receipt of one or more identification documents, such as a passport, driver's license, Aadhaar card, or utility bill, uploaded by the customer through the customer device. Along with these documents, the geographical coordinates of the customer’s location are also collected to support address validation.

[0064] The customer device captures and uploads images of physical identification documents, such as government-issued ID cards, passports, Aadhaar cards, driving licenses, or utility bills, which are later analyzed using Optical Character Recognition (OCR). OCR model extract and compare textual and image-based data from customer-uploaded documents. This ensures identity authentication and address verification before customer onboarding is approved. The triggering to upload identification document may be performed via client executive device, which is an extra safeguard.

[0065] The system utilizes an OCR model (as shown in fig. 2), selected from Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), or Transformer-Based Large Language Models (LLMs), to extract (201) both textual and image-based features from the uploaded documents. CNN-based OCR models are effective in detecting and classifying text in scanned images, while RNNs and LLM-based OCR models provide better sequence-based text extraction and context-aware text interpretation, improving the accuracy of recognition, even in cases of blurred or skewed documents.

[0066] Once the text is extracted, the system identifies key document attributes (202), including the customer’s name, date of birth, document number, and registered address. These attributes are cross-referenced (203) with official records stored in an external database, such as government identity verification systems, financial institution databases, or credit bureau records. Based on comparison the validation report (204) is generated, which ensures that the document details match official records, and minimizing the risk of identity fraud or document forgery.

[0067] In a real-world verification scenario, a customer undergoing identity and address verification submits an electricity bill as an identification document, along with a live image of their house and an electricity meter reading, to establish geolocation accuracy. In this case, the customer uploads an electricity bill containing their name, registered address, electricity account number, and bill issue date, ensuring that the document is recent. Simultaneously, the customer captures and submits a house image showing key exterior details and an electricity meter image displaying the meter number and current reading. The customer device also captures and transmits real-time geolocation coordinates (latitude and longitude) to validate the physical location. The electricity meter number extracted via OCR is compared with the meter number listed on the electricity bill to prevent fraud. Additionally, the geographical location coordinates of the meter image and house image are analyzed using coordinate similarity checks, to determine whether the submitted location aligns with the registered address providing extra layer on address verification. If the geolocation deviation is within a predefined threshold, the system confirms the customer's address authenticity.

[0068] Finally, as per fig. 1, the submitted geolocation data is compared (110) with the derived location coordinates from external databases. The live facial image is matched with facial templates extracted from customer identification documents (such as passports or driver's licenses) stored in the external database. And the OCR-based document verification process extracts textual data from scanned documents and compares them with official records. The verification process integrates geolocation validation, document authentication, and facial recognition to establish the authenticity of the customer's identity and residential address. At last, the server unit generates a comprehensive verification report (111) by aggregating the results from geolocation validation, document verification, and facial recognition analysis. This report determines whether the customer's identity and address have been successfully authenticated. If all verification layers align, the system confirms the customer's authenticity, allowing them to proceed with onboarding. However, if discrepancies arise in any verification step, the system may declare the authentication fail and the person is fraud. By leveraging this multi-layered process (100) ensuring secure, accurate, and fraud-resistant customer verification, significantly improving the reliability of digital onboarding processes while ensuring compliance with regulatory standards.

[0069] The multi-layered process (100) is also designed with conditional dependencies, meaning that if a core verification step fails, the subsequent authentication processes may not proceed further. For example, if geolocation verification is not justified, indicating that the customer’s current GPS coordinates significantly deviate from the registered address, the system may not proceed with document authentication or facial recognition, as the customer’s physical presence at the claimed location is uncertain. Similarly, OTP-based authentication serves as an initial security gateway, and if the customer fails to verify the OTP, the solution halts the verification process immediately, preventing unauthorized access or fraudulent verification attempts. The OTP validation is essential for enabling subsequent verification steps, ensuring that only the intended customer can proceed.

[0070] The process (100) also considers different verification combinations to determine whether customer authentication should continue. If geolocation verification is successful, but document authentication fails due to a mismatch in name, date of birth, or address details, the system may still attempt facial recognition as an additional security measure. However, if both geolocation and document authentication fail, the system may immediately flag the customer for manual review without proceeding to facial recognition. If geolocation verification is justified, document authentication is successful, but facial recognition fails, the system may trigger an additional document-based verification process (such as requesting another government-issued ID) before rejecting the customer.

[0071] This hierarchical validation model (100) ensures that each verification step is interdependent, preventing unnecessary processing if a critical layer of authentication is compromised. The conditional execution flow minimizes security risks, reduces false approvals, and enhances fraud detection, ensuring that only genuine customers with verifiable identities and addresses are onboarded.

[0072] The system (500) (as shown in fig. 5) for customer identification and validation is designed to facilitate a multi-layered authentication process for onboarding a customer. The system consists of five key components: the client executive device (502), the server unit (510), a geographical location verification module (506), a document verification module (508), a facial verification module (507), OTP based authentication module (514) and a verification module (509), all of which work in coordination to provide a secure and automated verification process.

[0073] Initially, the admin device (501) registers a login authorization user list on the server unit (510) to manage authentication and access control. The client executive device (502) logs in the verification platform after authorization by the admin device (501). The client executive device (502) transmits the registered customer details to the server unit (510), which is responsible for storing, retrieving, and processing customer information for identity verification. The server unit (510) plays a central role in managing the verification workflow, generating and transmitting verification requests to a field device (503) (either a client field executive device (504) or a customer device (505) for authentication. Additionally, the server unit (510) processes images of physical objects, such as customer-uploaded identification documents, house images, and environmental markers, alongside real-time geographical location data to enhance address verification.

[0074] Additionally, the geographical location verification module (506), executed by the server unit (510), is responsible for validating the customer's physical location to confirm their residential or business address. The system (500) first extracts the customer’s real-time geolocation coordinates from the customer device (505) and compares them with the registered address obtained from an external database (512). If an exact match is found, the address is converted into latitude and longitude coordinates for further validation. The system (500) segments the textual address into multiple location components, including house number, street name, locality, district, state, and postal code, and iteratively removes individual address segments from narrowest to broadest to determine possible location coordinates. The system (500) then identifies multiple potential location matches to determine the most probable address location. If the computed distance between the submitted coordinates and the official address falls within an acceptable threshold, the address is considered verified; otherwise, further verification steps are triggered.

[0075] The Document verification module (508) plays a crucial role in validating the authenticity of customer-submitted identity documents, ensuring compliance with regulatory. It is executed by the server unit (510), leveraging Al-powered Optical Character Recognition (OCR) technology, machine learning algorithms, and external database (512) comparisons to authenticate documents submitted via the customer device (505) or the client field executive device (504). The document verification process begins when the customer uploads images of government-issued identification documents, such as a passport, Aadhaar card, PAN card, driver's license, electricity bill, or any other supporting proof of identity or residence. Physical proof of documents are required for verification. These documents are transmitted to the server unit (510), where they are processed by the document verification module (508). The system (500) first receives and extracts textual and image-based data from the uploaded documents using OCR technology. The extracted information, including the customer’s name, date of birth, address, and document number, is then cross-verified with external database (512) to confirm its legitimacy.

[0076] The document verification module (508) also integrates with the geolocation verification module (506), ensuring that the address extracted from the document matches the real-time geographical location coordinates of the customer device (505). If the verification is successful, the system (500) allows the customer to proceed with further authentication steps, such as facial verification. However, if discrepancies are detected, the system (500) may flag the verification attempt for manual review or request additional supporting documents.

[0077] Further, the facial verification module (507), also executed by the server unit (510), and performs identity verification by capturing a live facial image from the customer device (505) and comparing it with pre-stored facial templates extracted from official identification documents for hygiene check. If the face match score surpasses a predefined threshold, the customer’s identity is verified; otherwise, the system (500) may flag the verification attempt as potentially fraudulent or request additional verification.

[0078] Furthermore, OTP-based authentication module (514) enhances security in customer identification and onboarding by ensuring that only authorized individuals can proceed with identity verification. This module (514) functions as a multi-layered security mechanism, integrating OTP validation with real-time geographical location verification to minimize fraud risks and unauthorized access attempts. The authentication process begins when the client executive device (502), typically operated by a bank official or an authorized representative, transmits a verification request to the server unit (510) to initiate customer authentication. Upon receiving the request, the server unit (510) generates a unique verification link associated with the specific customer and authentication session. This verification link is then transmitted to at least one field device (503), which could be either the client field executive device (504) or the customer device (505), depending on the authentication workflow.

[0079] When the customer or field executive accesses the verification link, the server unit (510) establishes a secure communication channel with the device (503) and triggers the OTP generation process. The server unit (510) generates a unique OTP and sends it to the customer device (505) via SMS, email, or a secure mobile application notification. This time- sensitive OTP, usually valid for a few minutes, ensures that only the intended customer with access to the registered device can complete the authentication process. Upon receiving the OTP, the customer submits it through the verification portal, along with real-time geographical location data (latitude and longitude) from their device (505) and the authentication process proceeds to the next step.

[0080] To enhance security, the OTP-based authentication module (514) implements multiple fraud prevention mechanisms. The system limits the number of OTP attempts, ensuring that if a customer enters the wrong OTP multiple times, the session is locked to prevent brute-force attacks. The OTP also has a short validity period, reducing the risk of stolen OTPs being used later. Additionally, the system (500) tracks the IP address and device (505) details of the customer to prevent access from unauthorized locations. By integrating OTP authentication with real-time geographical location tracking, the server unit (510) ensures that only legitimate customers with verifiable identities and valid geographical presence can proceed with onboarding. If authentication is successful, the server unit (510) passes the verification status to other modules, including the document verification module for ID validation (508), the facial verification module (507) for facial matching, and the geographical location verification module (506) for address confirmation. If any of these subsequent verification steps fail, the system (500) may flag the authentication request for further investigation or reject the onboarding process to prevent fraudulent activities. This comprehensive approach to identity verification and fraud detection ensures a secure, efficient, and compliant customer onboarding process.

[0081] Lastly, the verification module (509), executed by the server unit (510), combines multiple layers of verification results to generate a final authentication report (513). This module (509) cross-verifies the submitted geographical location coordinates from the customer device (505) with the derived coordinates from the external database (512). Additionally, the module (509) compares the extracted textual and image-based data from customer identification documents with official records to validate customer authenticity and address legitimacy. The facial image verification results are also incorporated to confirm whether the customer undergoing verification matches the identity on official documents. Finally, the server unit (510) generates a comprehensive validation report (513), integrating geolocation verification, document authentication, and facial recognition results. If all parameters align, the customer's identity and address are successfully verified, allowing them to proceed with onboarding. If any inconsistencies are detected, the system may flag the case of fraud.

[0082] In an embodiment, the system (500) leverages intrinsic document knowledge combined with Generative Al to validate document authenticity with a high degree of accuracy and reliability. This advanced Al-driven validation process ensures that each submitted document is genuine and free from tampering or forgery before it is used in customer onboarding.

[0083] The system (500) employs document- specific encoded triggers and hygiene checks to perform multi-layered verification. These encoded triggers act as predefined security markers that help detect document inconsistencies, alterations, or fabrication attempts. Additionally, the system (500) implements hygiene checks to verify that the document structure, format, and content integrity align with expected standards.

[0084] Once the individual document verification is completed, the system (500) cross-references the extracted information across multiple submitted documents to establish coherence and consistency. For instance, if a passport, driver’s license, and electricity bill are submitted together, the system validates the name, address, and date of birth across all documents, ensuring that the details match. By using Generative Al, the system performs human-like verification by identifying anomalies, potential mismatches, and document relationships, reducing the chances of identity fraud.

[0085] This Al-powered validation process enhances the accuracy, efficiency, and security of the verification process (100). By incorporating Generative Al and cross-referencing capabilities, the system (500) delivers a robust, fraud-resistant document authentication mechanism, ensuring trust and transparency in the onboarding process.

[0086] Hence, the system (500) in Fig. 5 ensures a highly secure, automated, and real-time identity verification process, reducing fraud risks while ensuring regulatory compliance for financial institutions, fintech firms, and businesses requiring customer authentication. By leveraging geolocation tracking, document verification, Al-powered facial recognition, and OTP-based authentication, the system (500) enhances efficiency, accuracy, and security in the customer onboarding process.

Claims

I Claims:

1. A method (100) for customer identification and address verification for onboarding a customer, comprising:-registering (101) one or more customer details from an authorized client executive device to a server unit for triggering a verification request and enabling customer document uploading, wherein the customer details comprises customer’s name, contact number, address etc.;-storing (102) the registered customer details in a memory unit associated with the server unit for retrieval and authentication;-generating and transmitting (103) the verification request to at least one field device for initiating authentication, wherein the field device comprises a client field executive device or a customer device;-submitting (104) an image of one or more physical objects provided by the customer or placed in and around of the vicinity where the customer is located at that movement, along with geographical location data; wherein the physical objects includes a live image of one or more customer identification documents, for verifying the geographical location of the customer; and wherein the geographical location verification process comprises:-comparing (105) the submitted geographical location data with the address registered in an external database to check for an exact match; If a match is found, converting the address into longitude and latitude coordinates for geolocation comparison;-segmenting (106) the textual address into more than one location components, wherein location components includes house number, street name, locality, district, state, and postal code etc.;-iteratively removing (107) each segment of the address one by one to identify possible geographical coordinates by focusing on broader location components; wherein the location components segmented from narrowest to broadest side;-identifying (108) more than one possible location coordinates within the identified area and performing a coordinate similarity check to determine the most probable address location;-capturing and submitting (109) a live facial image for face verification of the customer;-comparing (110) the submitted geographical location coordinates from the customer device with the derived geographical location coordinates obtained from the external database and the facial image with facial templates extracted from identification documents, with official documents stored in the external database; and—combining and generating a verification report (111) by the server unit incorporating results from geographical location verification, document validation, and facial recognition establishing authenticity of the customer's identity and address verification.

2. The method (100) as claimed in claim 1, wherein the geographical location verification process (104) further comprises:-computing the shortest distance between the location of customer device and the registered customer address;-comparing the computed distance against a predefined proximity threshold for geographical location validation; and-incorporating the geographical location verification results into the validation report.

3. The method (100) as claimed in claim 1, wherein the customer identification document verification process (104) comprises:-receiving and extracting (201) textual and image-based data from one or more customer-uploaded identification documents and the geographical coordinates using Optical Character Recognition (OCR) model selected from Convolutional Neural Networks, Recurrent Neural Networks or Transformer-Based (LLM) Models;-extracting (202) more than one key document attributes including name, date of birth, address, document number etc., from customer-uploaded identification documents and the geographical coordinates;-comparing (203) the extracted OCR text with official documents stored in the external database to confirm customer authenticity and address validation; and- generating (204) the validation report incorporating the document verification result and address validation results.

4. The method (100) as claimed in claim 1, wherein the face verification process (109) for customer identity authentication, comprises:- extracting facial features (301) from the live image using a Convolutional Neural Networks (CNNs);-converting (302) the extracted facial features into numerical vector representations using neural network;-comparing (303) the generated numerical vectors with data with facial templates stored in the external database by applying one or more similarity distance metrics;- assigning (304) a face match score based on the computed similarity metrics and evaluating the score against a predefined threshold value; and-triangulating (305) the face match score by integrating geographical location validation result and document verification result in the validation report by the server unit.

5. The method (100) as claimed in claim 4, wherein comparing the generated numerical vectors with stored facial templates in the external database using one or more similarity distance metrics (302), comprising:-cosine similarity to measure angle between plurality of numerical vectors to determine similarity;-euclidean distance to measure straight-line distance between plurality of numerical vectors; and-L2-Norm to normalize the vector magnitude and measure similarity.

6. The method (100) as claimed in claim 1, wherein generating and transmitting the verification request (103), comprising:- transmitting the verification request (401) from the authorized the client executive device to the server unit to initiate customer authentication;-generating (402) a verification link by the server unit to at least one of the field device;-accessing the link (403) by at least one of the field device and start communicating with the server unit;-generating an One-Time Password (OTP) (404) from the server unit and sending it to the customer device upon accessing the verification link by at least one of the field device; and-submitting the OTP received (405) on to the server unit along with geographical location data by the customer device for customer authentication.

7. A system (500) for customer identification and validation for onboarding a customer, comprising:• A client executive device (502), configured to:-register one or more customer details to a server unit (510) for triggering a verification request and enabling customer document uploading, wherein the customer details comprises customer’s name, contact number, address etc.; and -transmit the registered customer details to the server unit (510) for authentication and retrieval;• A server unit (510), configured to:-store the registered customer details in a memory unit (511) for identity verification,-generate and transmit a verification request to at least one field device (503) for initiating authentication, wherein the field device (503) comprises a client field executive device (504) or a customer device (505); and-receive and process an image of one or more physical objects provided by the customer or placed in and around the vicinity where the customer is located at that moment, along with geographical location data, wherein the physical objects includes a live image of one or more customer identification documents, for verifying the geographical location of the customer;• A geographical location verification module (506), executed by the server unit (510), configured to:-compare the submitted geographical location data with the address registered in an external database (512) to check for an exact match, If a match is found, convert the address into longitude and latitude coordinates for geolocation comparison;-segment the textual address into more than one location components, wherein location components includes house number, street name, locality, district, state, and postal code etc.;-iteratively remove each segment of the address one by one to identify possible geographical coordinates by focus on broader location components; wherein the location components segmented from narrowest to broadest side; and-identify more than one possible location coordinates within the identified area and perform a coordinate similarity check to determine the most probable address location;• a facial verification module (507), executed by the server unit (510), configured to capture and submit a live facial image from the customer device (505) for face verification of the customer; and• a verification module (509), executed by the server unit (510), configured to:-comparing the submitted geographical location coordinates from the customer device (505) with the derived geographical location coordinates obtained from the external database (512) with customer identification documents and the facial image with facial templates extracted from identification documents stored in the external database (512); and-combining and generating a verification report (513) by the server unit (510) incorporating results from geographical location verification, document validation, and facial recognition establishing authenticity of the customer's identity and address verification.

8. The system (500) as claimed in claim 7, comprising:• the customer device (505), configured to: capture and upload one or more customer identification documents, transmit the captured document images and geographical coordinates to the server unit (510) for processing; and• the server unit (510), communicatively connected to the external database (512) and configured to execute a document verification module (508), comprising: receive and process customer-uploaded identification documents and geographical coordinates; extract textual and image-based data from the uploaded documents using an Optical Character Recognition (OCR) model, selected from Convolutional Neural Networks, Recurrent Neural Networks or Transformer-Based (LLM) Models; identify and extract more than one key document attribute, such as name, date of birth, address, document number etc., from customer-uploaded identification documents and the geographical coordinates; compare the extracted OCR text with official documents stored in the external database (512) to confirm customer authenticity and validate the customer’s residential address; andgenerates the verification report incorporating the document verification result and address verification results.

9. The system (500) as claimed in claim 7, wherein the facial verification module (507) comprising:• the customer device (504), configured to: capture a live facial image of the customer for identity authentication, transmit the captured facial image to the server unit (510) for processing; and• the server unit (510), communicatively connected to the external database (512) and configured to execute the facial verification module, comprising: extract facial features from the received image using a Convolutional Neural Network (CNN) model, convert extracted facial features into numerical vector representations using a neural network-based facial recognition model, compare the generated numerical vectors with facial templates stored in the external database (512) by applying at least one similarity distance metric, assign a face match score based on the computed similarity metrics and evaluating the score against a predefined threshold value; and triangulate the face match score by integrating geolocation validation and document verification in the validation report by the server unit (510).

10. The system (500 as claimed in claim 7, further comprising an OTP based authentication module (514), comprising:• the client executive device (502), configured to transmit a verification request to the server unit (510) to initiate customer authentication;• the server unit (510), communicatively connected to the memory unit (511), configured to execute an OTP-based authentication module, comprising:-generate a verification link upon receiving the verification request,-transmit the generated verification link to at least one field device (503) for initiating authentication;-enable at least one field device (503) to access the verification link and start communication with the server unit (510);-generating an One-Time Password (OTP) from the server unit (510) and sending it to the customer device (505) upon accessing the verification link by at least one of the field device (503); and-submitting the OTP received on to the server unit (510) along with geographical location data by the customer device (505) for customer authentication.

Images