Description:AN INTELLIGENT VENDOR IDENTIFICATION & VERIFICATION SYSTEM & METHOD THEREOF
TECHNICAL FIELD
[01] The present invention in general pertains to the field of security systems and more specifically to systems and methods designed for automated vendor identification and verification. This innovation aims to enhance security protocols, streamline operations, and improve overall efficiency across various commercial settings, including retail and logistics, healthcare facilities, corporate offices, government facilities, and event venues.

BACKGROUND OF THE INVENTION
[02] The following description of the related art is intended to provide background information pertaining to the field of the disclosure. This section may include certain aspects of the art that may be related to various features of the present disclosure. However, it should be appreciated that this section is used only to enhance the understanding of the reader with respect to the present disclosure, and not as admissions of the prior art.
[03] In today's complex commercial landscape, accurate and efficient vendor identification & verification is a critical task that underpins various business processes, including financial management, procurement, and compliance. Traditional methods of vendor verification, often involving manual data entry and comparison, are time-consuming, error-prone, and susceptible to human error. To address these limitations, a robust and automated vendor detection system is required.
[04] One of the primary challenges in vendor identification lies in the variability and inconsistency of vendor information across different documents. Vendor addresses, for instance, may be formatted differently or contain minor discrepancies, such as variations in abbreviations, punctuation, or the inclusion of additional details. These inconsistencies can hinder the accurate matching of vendor information with existing records.
[05] Additionally, vendor logos can appear in various formats, sizes, and orientations, making it difficult for traditional image recognition techniques to reliably identify them. Logos may be presented in different colours, with varying degrees of clarity, or even partially obscured by other elements on the document. This variability poses a significant challenge for automated logo recognition systems.
[06] Furthermore, the quality of scanned documents can significantly impact the accuracy of vendor detection. Low-resolution or noisy images can hinder the accurate extraction of text and image data, leading to errors in the identification process. Poor image quality can result from various factors, including the original document's condition, the scanning process, and the subsequent image processing techniques.
[07] In addition, both buyer and seller information is typically present on an invoice, and this information may also exist in the vendor master, therefore there is a risk of incorrectly identifying the relevant vendor.
[08] To overcome these challenges, there is a need for an intelligent vendor detection system that can accurately identify and verify vendors based on the information provided in documents such as invoices, bank statements, and cheques. This system should be able to handle variations in data format, quality, and presentation, ensuring accurate and reliable vendor identification.
[09] One key aspect of the system is the ability to accurately extract vendor information from documents, including addresses, TIN numbers, GST numbers, logos, and website details. This requires advanced text recognition and image processing techniques that can handle various font styles, sizes, and layouts. Additionally, the system must be able to identify and correct common errors, such as typos, misspellings, and inconsistencies in formatting.
[10] Another critical component of the system is the ability to match extracted vendor information with existing vendor records. This involves comparing vendor names, addresses, TIN numbers, GST numbers, and logos to identify potential matches. To improve matching accuracy, the system may employ techniques such as fuzzy string matching, phonetic matching, and semantic similarity analysis.
[11] In general, vendor identification, a seemingly straightforward task, can pose significant challenges for various industries. These challenges can lead to operational inefficiencies, financial losses, and reputational damage. Here are some of the key industry-facing problems:
Financial Implications: Incorrect Payments: Misidentified vendors can lead to incorrect payments, resulting in overpayments or underpayments, Fraudulent Activities: By prioritizing accurate vendor identification, organizations can take proactive steps to prevent fraudulent activities, such as vendor fraud and identity theft, Delayed Payments and Penalties: Delayed payments due to vendor identification errors can result in late fees and penalties.
Operational Inefficiencies: Slow Processing Times: Manual verification and data entry processes can slow down invoice processing and payment cycles, Increased Administrative Burden: Incorrect vendor information can lead to increased administrative overhead, such as manual data correction and reconciliation, Supply Chain Disruptions: Inaccurate vendor information can disrupt supply chain operations, leading to stock outs, delays, and customer dissatisfaction.
Compliance and Regulatory Risks: Non-Compliance with Tax Regulations: Incorrect vendor information, especially regarding tax identification numbers, can lead to non-compliance with tax regulations and potential penalties, Anti-Money Laundering (AML) and Know Your Customer (KYC) Regulations: Failure to properly identify vendors can expose organizations to AML and KYC risks, Ethical Sourcing and Sustainability Regulations: Incorrect vendor information can hinder efforts to comply with ethical sourcing and sustainability regulations.
Data Quality and Integrity: Inaccurate Data: Incorrect vendor information can lead to inaccurate data in financial systems and other databases, compromising data quality and integrity, Data Security Risks: Inaccurate vendor information can increase the risk of data breaches and security vulnerabilities.
Decision Making and Strategic Planning: Poor Vendor Performance: Incorrect vendor identification can lead to the selection of underperforming vendors, impacting the overall performance of the organization, Suboptimal Supplier Relationships: Misidentified vendors can hinder the development of strong and collaborative relationships with suppliers.
Specific Industry Impacts: Healthcare: Incorrect vendor identification can lead to medical errors, billing errors, and compliance issues, Retail: Misidentified vendors can result in product quality issues, delayed shipments, and customer dissatisfaction, Manufacturing: Incorrect vendor information can disrupt production schedules, increase costs, and impact product quality, Financial Services: Misidentified vendors can expose financial institutions to fraud, money laundering, and regulatory risks, Government: Incorrect vendor identification can lead to inefficient procurement processes, wasted public funds, and fraud.
[12] To address the challenges inherent in vendor identification, organizations must implement robust processes that combine technological innovation with rigorous data management. These processes should include:
• Data Quality and Standardization: Ensuring the accuracy and consistency of vendor data is paramount. This involves establishing clear data standards, implementing data validation rules, and regularly reviewing and updating vendor information.
• Advanced Data Capture Technologies: Leveraging artificial intelligence and machine learning can significantly automate data extraction and validation processes. By utilizing these technologies, organizations can efficiently extract accurate vendor information from various documents, such as invoices, contracts, and purchase orders.
• Real-time Vendor Verification: Implementing real-time verification processes, such as biometric authentication or digital signature verification, can help identify and prevent fraudulent activities. This ensures that only authorized vendors can access sensitive information and resources.
• Robust Data Security Measures: Protecting sensitive vendor information is essential. Organizations must implement strong security measures, including encryption, access controls, and regular security audits, to safeguard against unauthorized access and data breaches.
• Continuous Monitoring and Improvement: Regular monitoring and improvement of the vendor identification process are crucial to maintain data accuracy and identify emerging trends. By continuously analyzing vendor data and feedback, organizations can refine their processes and improve their overall efficiency.
[13] Hence, there is need of an intelligent vendor detection system that can significantly improve the accuracy and efficiency of vendor identification processes, reducing errors, saving time, and enhancing overall business operations.
[14] In addition, there is also a need to incorporate machine learning techniques in vendor detection system that can significantly enhance the accuracy and efficiency of vendor identification. By training machine learning models on historical vendor data, the system can learn to recognize patterns, identify anomalies, and make more accurate predictions. This can help to automate routine tasks, reduce manual effort, and improve the overall quality of vendor information.
[15] Moreover to the technical aspects, it is essential to consider the legal and regulatory implications of vendor identification. Organizations must ensure compliance with data privacy regulations, such as GDPR and CCPA, to protect sensitive vendor information. By implementing robust data protection measures, organizations can minimize the risk of legal liabilities and reputational damage.

OBJECTS OF THE INVENTION

[16] Some of the objects of the present disclosure, which at least one embodiment herein satisfies are listed herein below.
[17] It is an object of the present invention to provide a system and method for automated vendor identification and verification that utilizes advanced data capture techniques, machine learning, and data analytics to accurately identify and verify vendors based on information extracted from various documents, such as invoices, contracts, and purchase orders.
[18] It is an object of the present invention to provide a system and method for cleaning and standardizing vendor data to improve data quality and consistency, including techniques for handling variations in address formats, names, and other relevant information.
[19] It is an object of the present invention to provide a system and method for automated data extraction from vendor documents that employs advanced optical character recognition (OCR) and natural language processing (NLP) techniques to accurately extract vendor information, such as names, addresses, Tax IDs, and contact details.
[20] It is an object of the present invention to provide a system and method for vendor matching that can accurately identify and match vendors based on various attributes, including name similarity, address similarity, and other relevant factors.
[21] It is an object of the present invention to provide a system and method for protecting sensitive vendor information, including personal data and financial information.
[22] It is an object of the present invention to provide a system and method for integrating and synchronizing vendor data across multiple systems and databases.
[23] It is an object of the present disclosure to provide a system and a method that facilitates novel image processing technique designed to enhance the accuracy and efficiency of vendor detail extraction from multiple scanned documents.
[24] It is an object of the present disclosure to provide a system and a method incorporating machine learning model specifically trained to identify and extract vendor details from diverse document formats.
[25] It is an object of the present disclosure to provide a system and a method incorporating machine learning model specifically trained on a large dataset of labelled documents to achieve high accuracy and adaptability.
[26] It is an object of the present disclosure to provide a system and a method that could focus on a user-friendly interface allowing users to easily interact with the automated vendor detail extraction system.
[27] It is an object of the present disclosure to provide a system and a method for integrating the automated vendor detail extraction system with enterprise resource planning (ERP) systems that would streamline the workflow and improve data accuracy.

SUMMARY OF THE INVENTION

[28] In accordance with an embodiment, the present disclosure provides an intelligent vendor identification method , wherein said method comprising the steps of acquiring, an image of the negotiable instruments via one or more processors of a computing device, and pre-processing the same by utilizing an imaging library to correct skew and remove noise, thereby enhancing image quality for further analysis and receiving a plurality of samples from the pre-processed image for subsequent analysis; analysing, each of the plurality of samples via one or more processors, to identify at least one entity associated with the negotiable instrument by employing a learning model including trained dataset; validating, the identified entity, via one or more processors, against predefined patterns to confirm its accuracy; extracting data related to at least one identified entity from the negotiable instruments in the event of successful validation
via one or more processors; and in the event of failed entity identification, extracting a vendor address from the negotiable instruments, and implementing fuzzy matching techniques to compare the extracted vendor address against a vendor master list for the purpose of identifying the corresponding vendor; and updating the vendor information, including the vendor name and address, in a secure, tamper-proof database, ensuring data privacy and integrity through encryption and access controls, and storing the updated information in a structured format for efficient retrieval and analysis.
[29] In accordance with an aspect, an image processing library is employed to correct skew and eliminate noise from the image of the entity, thereby improving the overall quality of the images.
[30] In accordance with an aspect, an OCR engine is utilized to identify at least one of the entity after correcting skew and removing noise from the image of the entity.
[31] In accordance with an aspect, the trained dataset is utilized to detect and locate at least one of the entity of the negotiable instruments via employing bounding box mechanism.
[32] In accordance with an aspect, data augmentation technique is employed to extract the data of at least one of the entity.
[33] In accordance with an aspect, the data augmentation technique include, but are not limited to, rotation, flipping, cropping, and intensity adjustments, to generate a diverse set of training images.
[34] In accordance with an aspect, the trained dataset comprises the plurality of samples, where each of the plurality of samples are categorized into at least one of a training subset, a validation subset, and a test subset.
[35] In accordance with an aspect, a performance metrics is applied on the trained dataset to track an impact of the training subset on the trained dataset.
[36] In accordance with an embodiment, the present disclosure provides an intelligent vendor identification system ,wherein said system comprising one or more processors associated with a computing device; and a memory operatively coupled to the one or more processors, wherein the memory comprises processor-executable instructions, which on execution, cause the one or more processors to: acquire, an image of the negotiable instrument, and pre-processing the same by utilizing an imaging library to correct skew and remove noise, thereby enhancing image quality for further analysis and receiving a plurality of samples from the pre-processed image for subsequent analysis; analyse, each of the plurality of samples via one or more processors, to identify at least one entity associated with the negotiable instruments by employing a learning model including trained dataset; validate, the identified entity, via one or more processors, against predefined patterns to confirm its accuracy; extract, the data related to at least one
identified entity from the negotiable instrument in the event of successful validation via one or more processors, and in the event of failed entity identification, extracting a vendor address from the negotiable instruments, and implementing fuzzy matching techniques to compare the extracted vendor address against a vendor master list for the purpose of identifying the corresponding vendor; and updating the vendor information, including the vendor name and address, in a secure, tamper-proof database via one or more processors, ensuring data privacy and integrity through encryption and access controls, and storing the updated information in a structured format for efficient retrieval and analysis.
[37] In accordance with an aspect, an image processing library is employed to correct skew and eliminate noise from the image of the entity, thereby improving the overall quality of the images.
[38] In accordance with an aspect, an OCR engine is utilized to identify at least one of the entity after correcting skew and removing noise from the image of the entity.
[39] In accordance with an aspect, the trained dataset is utilized to detect and locate at least one of the entity of the negotiable instrument via employing bounding box mechanism.
[40] In accordance with an aspect, data augmentation technique is employed to extract the data of at least one of the entity.
[41] In accordance with an aspect, the data augmentation technique include, but are not limited to, rotation, flipping, cropping, and intensity adjustments, to generate a diverse set of training images.
[42] In accordance with an aspect, the trained dataset comprises the plurality of samples, where each of the plurality of samples are categorized into at least one of a training subset, a validation subset, and a test subset.
[43] In accordance with an aspect, a performance metrics is applied on the trained dataset to track an impact of the training subset on the trained dataset.
[44] In accordance with an embodiment, the present disclosure provides a non-transitory computer-readable medium incorporated in an intelligent vendor identification system, wherein said medium comprising processor executable instructions that cause a processor to: acquire, an image of the negotiable instruments, and pre-processing the same by utilizing an imaging library to correct skew and remove noise, thereby enhancing image quality for further analysis and receiving a plurality of samples from the pre-processed image for subsequent analysis; analyse, each of the plurality of samples via one or more processors, to identify at least one entity associated with the negotiable instrument by employing a learning model including trained dataset; validate, the identified entity, via one or more processors, against predefined patterns to confirm its accuracy; extract data related to at least one identified entity from the negotiable
instruments via one or more processors in the event of successful validation, and in the event of failed entity identification, extracting a vendor address from the negotiable instruments, and implementing fuzzy matching techniques to compare the extracted vendor address against a vendor master list for the purpose of identifying the corresponding vendor; and updating the vendor information, including the vendor name and address, in a secure, tamper-proof database via one or more processors, ensuring data privacy and integrity through encryption and access controls, and storing the updated information in a structured format for efficient retrieval and analysis.

BRIEF DESCRIPTION OF THE DRAWINGS
[45] The accompanying drawings are included to provide a further understanding of the present invention, and are incorporated in and constitute a part of this specification. The drawings illustrate exemplary embodiments of the present invention and, together with the description, serve to explain the principles of the present invention.
[46] FIG.1 illustrates an exemplary block diagram representation of a network architecture implementing vendor identification system according to embodiments of the present disclosure.
[47] FIG.2 illustrates a system designed for intelligent vendor identification in accordance with an embodiment of the present disclosure.
[48] FIG.3 illustrates a flowchart depicting a process for vendor identification from documents in accordance with an embodiment of the present disclosure.
[49] FIG.4 illustrates a process for extracting and validating information from invoices in accordance with an embodiment of the present disclosure.
[50] FIG.5 illustrates an exemplary user interface of the invoice to be used for data extraction, in accordance with an embodiment of the present disclosure.
[51] FIG. 5A illustrates an exemplary image of the logo field of the invoice to be extracted, in accordance with an embodiment of the present disclosure.
[52] FIG. 5B illustrates an exemplary image of the GST field of the invoice that is to be extracted, in accordance with an embodiment of the present disclosure.
[53] FIG. 6 illustrates, a series of steps involved in a process for training a trained dataset, in accordance with an embodiment of the present disclosure.
[54] FIG. 7 is a flow diagram depicting a proposed method for extracting data from the negotiable, in accordance with an embodiment of the present disclosure.

DETAILED DESCRIPTION OF INVENTION
[55] In the following description, for the purposes of explanation, various specific details are
set forth in order to provide a thorough understanding of embodiments of the present disclosure. It will be apparent, however, that embodiments of the present disclosure may be practiced without these specific details. Several features described hereafter can each be used independently of one another or with any combination of other features. An individual feature may not address all of the problems discussed above or might address only some of the problems discussed above. Some of the problems discussed above might not be fully addressed by any of the features described herein.
[56] The ensuing description provides exemplary embodiments only and is not intended to limit the scope, applicability, or configuration of the disclosure. Rather, the ensuing description of the exemplary embodiments will provide those skilled in the art with an enabling description for implementing an exemplary embodiment. It should be understood that various changes may be made in the function and arrangement of elements without departing from the spirit and scope of the disclosure as set forth.
[57] Reference throughout this specification to “one embodiment” or “an embodiment” or “an instance” or “one instance” means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the present disclosure. Thus, the appearances of the phrases “in one embodiment” or “in an embodiment” in various places throughout this specification are not necessarily all referring to the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiment.
[58] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. As used herein, the singular forms “a”, “an”, and “the” are intended to include the plural forms as well, unless the context indicates otherwise. It will be further understood that the terms “comprises” and/or “comprising,” when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof. As used herein, the term “and/or” includes any and all combinations of one or more of the associated listed items.
[59] Aspects of present invention relates to a system and a method for intelligent vendor detection. This system leverages advanced image processing techniques and machine learning algorithms to automatically identify and extract vendor information from various types of documents, including invoices, purchase orders, and bank statements. The system addresses the challenges associated with manual data extraction, such as variations in document formats, image quality, and font styles.
[60] Various embodiments of the present disclosure will be explained in detail with reference to FIGs. 1-8.
[61] FIG.1 illustrates an exemplary block diagram representation of a network architecture 100 implementing vendor identification system 110 (also referred to as a system 110, herewith), according to embodiments of the present disclosure.
[62] The network architecture (100) depicted in FIG.1 comprises multiple components such as a system (110), a computing device (108), a centralized server (118), a decentralized database (120), a communication network (106), a set of electronic devices (104), and a set of users (102).
[63] The multiple components involved in a network architecture 100 are defined below –
• System (110): This represents the core computing unit, likely a server or a powerful workstation. It includes a processor (112), memory (116), and I/O interface (114) for communication.
• Computing Device (108): This represents a general-purpose computing device, such as a laptop or desktop computer. It connects to the network for accessing resources and services.
• Centralized Server (118): This server stores and processes data centrally, acting as a single point of control for the network.
• Decentralized Database (120): This component distributes data across multiple nodes, providing redundancy and improved performance.
• Communication Network (106): This network enables communication between devices and the server, allowing data transfer and resource sharing. It can be wired or wireless.
• Electronic Devices (104): This category includes various electronic devices that can connect to the network, such as smartphones, tablets, IoT devices, etc.
• Users (102): These are the end-users who interact with the network through electronic devices.
[64] The functionality of multiple components involved in a network architecture 100 are defined below -
• The system (110) communicates with the centralized server (118) and the decentralized database (120) via a communication network (106). The centralized server (118) can be configured as a standalone server, a remote server, a cloud-based server, or a distributed server system. It may also include hardware supporting cloud services, home servers, virtualized servers, or multiple processors executing server-side functions.
• The communication network (106) mentioned can either be wired or wireless. The wireless communication network may employ various technologies, such as carrier networks (e.g., GSM (Global System for Mobile Communications), UMTS (Universal Mobile Telecommunications System), and LTE (Long-Term Evolution), NR (New Radio)), the internet, intranets, LANs, WANs, or mobile networks.
• The system (110) can be implemented as a standalone device, such as the centralized server (118), or integrated into computing devices (108) or electronic devices (104). These electronic devices (104) may be used by users (102), who could be individuals or entities like insurance companies, logistics firms, government agencies, auditors, or regulatory bodies.
• The system (110) may include a processor (112), an I/O interface (114), and a memory
(116). The I/O interface (114) receives user input from electronic devices (104). The processor (112), often integrated with the computing device (108), executes instructions stored in the memory (116) to extract vendor details from negotiable instruments. By combining these components and leveraging the communication network, the system
(110) enables efficient and accurate data extraction, processing, and analysis.
• The break down shows how each component of the network architecture contributes to the process of vendor identification:
o Data Acquisition:
? Electronic Devices (104):
• Users input vendor information (e.g., name, address, contact details) into these devices.
• This data could be extracted from invoices, purchase orders, or other relevant documents.
o Data Transmission:
• The collected data is transmitted to the centralized server (118) via the communication network (106).
o Data Processing:
• Centralized Server (118):
o Receives and stores the incoming vendor data.
o Pre-processes the data to clean and normalize it.
o Extracts relevant features from the vendor information, such as name, address, and product/service categories.
o System (110):
? Utilizes the computing device (108), the processor (112), and memory
(116) to perform data analysis and processing tasks.
? Employs machine learning algorithms to identify patterns and anomalies in the vendor data.
? Compares the extracted features with a database of known vendors.
? Uses techniques like fuzzy matching to handle variations in vendor information.
o Vendor Identification:
? Matching Algorithms: The system (110) employs various algorithms to match incoming vendor data with existing records:
• Exact matching: Directly compares vendor names, addresses, and other identifiers.
• Fuzzy matching: Handles variations in spelling, abbreviations, and formatting.
• Machine learning: Uses trained models to identify patterns and classify vendors.
o Data Validation:
? Validates the identified vendor information against predefined criteria.
? Checks for inconsistencies, errors, or potential fraud

o Data Storage and Analysis:
? Centralized Database (118) and Decentralized Database (120): Stores the verified vendor information, including:
• Vendor name and address
• Contact information
• Product or service offerings
• Performance history
• Financial information
? Analyses the stored data to identify trends, risks, and opportunities.
? Generates reports and insights for decision-making.

o User Interaction:
? Electronic Devices (104):
• Users can access the system through their devices to view vendor information, search for specific vendors, and generate reports.
• The system can provide alerts or notifications for potential issues or irregularities.
[65] In between a key aspect of the process involves encryption to protect sensitive information. Here's a breakdown of the encryption sections:
• Encryption of the Identified Entity:
• Purpose: To safeguard the identity of the vendor.
• Method: A secure encryption algorithm is employed to encrypt the identified entity (e.g., vendor name) before further processing.
• Encryption of Extracted Data:
• Purpose: To protect sensitive vendor information like name and address.
• Method: A secure encryption algorithm is used to encrypt the extracted data, ensuring its confidentiality.
• Key Points about Encryption:
• Secure Encryption Algorithm: The choice of encryption algorithm is crucial. Strong algorithms like AES (Advanced Encryption Standard) are often used for their robustness.
• Encryption Key: A secure encryption key is used to encrypt and decrypt the data. This key must be protected to maintain the security of the information.
• Decryption Key: A decryption key is used to decrypt the encrypted data for validation and analysis purposes. This key must be securely stored and used only by authorized personnel.
• Importance of Encryption:
• Data Privacy: Encryption helps protect sensitive information from unauthorized access.
• Data Integrity: Encryption ensures that the data remains unaltered during transmission and storage.
Compliance: In many industries, encryption is required by regulations to protect customer data.
[66] By implementing robust encryption techniques, the system can effectively safeguard the privacy and security of the vendor information, ensuring compliance with data protection regulations.
[67] FIG.2 illustrates a system 110 designed for intelligent vendor identification in accordance with an embodiment of the present disclosure. It consists of various components working together to analyse images of negotiable instruments, extract relevant information, and identify associated vendors.
Key Components and Their Functions
• System (110):
a. The overarching system that encompasses all other components.

b. Coordinates the overall operation and data flow.

I. Computing Device (108):
1. A general-purpose computing device, such as a laptop or server.

2. Provides the computational resources for processing and analysis.

II. Processing Engine (112):
1. The core processing unit that executes instructions and performs calculations.
2. Handles tasks like image processing, data analysis, and decision- making.
III. Memory (114):
1. Stores data, instructions, and intermediate results.
2. Essential for the system's operation.

IV. Interfaces (116):
1. Enable communication with external devices and networks.

2. Handles input/output operations, such as receiving images and displaying results.
V. Processing Engine(s) (112): A specialized unit or software module dedicated to specific processing tasks. May include components like:
1. Processing Unit (112-P): Handles basic data processing and manipulation. 2.Analysing Unit (112-A): Performs analysis tasks, such as pattern recognition
and data mining.
3. Extracting Unit (112-E): Extracts relevant information from the images, such as vendor names, addresses, and other details.
4. Validating Unit (112-V): Verifies the accuracy and consistency of the extracted data.
5. Conditional Logic Unit (112-C)- Implementing a decision-making process to trigger optical character recognition module when the deep learning model's employed by the Validating unit determines that the prediction confidence for vendor logo identification falls below a predefined threshold; The processing Unit (112-P) further recognizes text from vendor address images and store the extracted vendor address information in the data storage module.
6. Other Units (112-O): May include additional specialized units for specific tasks (e.g., security, encryption)
VI. Database (122): Stores vendor information, system configuration, and historical data. Enables efficient data retrieval and analysis.
[68] FIG.3 illustrates a flowchart depicting a process for vendor identification from documents in accordance with an embodiment of the present disclosure. Here's a breakdown of the steps involved:
I. Step 301: Receive Image and Pre-processing: The process begins with receiving an image for processing. This is followed by pre-processing steps, which could include noise reduction, image enhancement, or format conversion to prepare the image for analysis.
II. Step 302: Vendor Detection - Three Parallel Approaches: The core of the process involves three parallel methods for detecting the vendor:
? Logo-based Vendor Identification: This method attempts to identify the vendor based on its logo present in the image. It sends the image to a trained model for vendor prediction. If a vendor is found based on the logo, a weightage is assigned to this detection. If no vendor is found, this branch of the process ends.
? Full Text Search-based Vendor Identification: This method uses Optical Character Recognition (OCR) to extract text from the image. It then performs a full-text search against a vendor database to find potential matches. The top 30 vendor matches are retrieved. If matches are found, they are returned as detected vendors.
? Vendor Identification with Address Block Detection: This method focuses on identifying the vendor based on address information extracted from the image. It connects with a vendor master database and matches extracted address information against the database using regular expressions (regex). If matches are found, they are returned as detected vendors.
III. Step 303: Check for Consistent Vendor Detection: After the three parallel methods have completed, the process checks if a vendor was detected in all three cases. If a vendor was detected in all three cases, the process assigns the detected vendor with the top weightage and proceeds. If a vendor was not detected in all three cases, the process proceeds to the next step.
IV. Step 304: Apply Logic to find Vendor with Highest Weightage: If the vendor wasn't detected by all three methods, this step applies logic to determine the vendor with the highest weightage among the results from the different methods. This logic likely considers the confidence levels or weightings assigned by each method.
V. Step 305: Final Vendor Check: After applying the weightage logic, the process checks if a vendor was found at all. If a vendor was found, the identified vendor is returned. If no vendor was found, a message is returned indicating that no vendor was detected in the image.
VI. Step 306: Exit: The process concludes.
[69] In essence, this process uses a multi-pronged approach to increase the accuracy of vendor identification by combining logo recognition, text search, and address matching. It also incorporates a weighting system to handle cases where the different methods produce conflicting or incomplete results.
[70] This approach aims to improve the accuracy and efficiency of vendor identification by combining both logo-based and address-based methods.
[71] FIG.4 illustrates a process for extracting and validating information from invoices in accordance with an embodiment of the present disclosure. Here's a breakdown of the steps involved:
• Image Processing of Invoice (402):
I. The process begins with an image of an invoice.

II. The image is pre-processed to enhance its quality and prepare it for further analysis. This may involve tasks like noise reduction, contrast adjustment, and skew correction.
• Image Processing of Logo (404):
III. The image of the invoice is further processed to identify and extract the logo. This step is crucial for vendor identification.
• YOLOv7 Model Application (406):
IV. The pre-processed image, including the extracted logo, is fed into a YOLOv7 model, a state-of-the-art object detection model.
V. The model aids in the accurate localization of text blocks and other key fields within the image, ultimately facilitating efficient data extraction and vendor identification.
• Check for GST Field (408):
VI. The system checks for the presence of a GST field within the detected entities.
• OCR Text Extraction (410):
VII. Optical Character Recognition (OCR) techniques are employed to extract text from the detected entities, including the GST number.
• Field Validation and Post-Processing (412):
VIII. The extracted text, including the GST number, is validated to ensure accuracy and completeness.
IX. Post-processing steps, such as cleaning and formatting the extracted data, may be applied.
• Validation Data of All Invoice Entities (414):
X. The extracted data is compared against a predefined set of validation rules or a database of known invoice formats.
XI. This step ensures that the extracted data is consistent with the expected invoice structure.
• Fall-back Logic (416):
XII. If the automatic extraction and validation processes fail, a fall-back mechanism is triggered.
XIII. This might involve manual intervention or alternative extraction techniques to recover the missing information.
[72] Overall, the diagram depicts a robust and automated process for extracting and validating information from invoices, with a focus on logo mentioned in invoice and if the accurate vendor detail is not determined than accurate GST number extraction is required for accurate validation of the vendor details.
[73] FIG.5 illustrates an exemplary user interface 500 of the invoice to be used for data extraction, in accordance with an embodiment of the present disclosure. With respect to FIG.5, fields that are commonly extracted from the invoice are, for example:
• Goods and Services Tax Identification Number (GSTIN): a unique 15 digit identification number assigned to every taxpayer (primarily dealer or supplier or any business entity) registered under the GST regime.
• Vendor Address: A vendor address is the physical location of a supplier, including the street, city, state/province, country, and postal code.
? Invoice Date: Date on which the invoice is issued.
? Invoice Time: Time on which the invoice is issued.
? Amount: A payable amount mentioned on an invoice.
? Corporate Identification Number (CIN): A 21-digit alphanumeric code that uniquely identifies a company incorporated in India.
[74] FIG. 5A illustrates an exemplary image of the logo field 502 of the invoice to be extracted, in accordance with an embodiment of the present disclosure. With respect to FIG. 5A, is shown the logo field that is enclosed in boundaries.
[75] FIG. 5B illustrates an exemplary image of the GST field 504 of the invoice that is to be extracted, in accordance with an embodiment of the present disclosure.
[76] When the GST data is embedded within structured elements, the pre-existing imaging library may be used to extract the GST field. This ensures that only relevant text data is available for the OCR processing. As may be appreciated, extraction of the tables may be made easier by using borders and identifying cells of the table, based on lines.
[77] When documents are scanned into computers as images, they are stored as collections of pixels with varying colours. This transforms the task into an image processing challenge that requires efficient data extraction. With advancements in deep learning, traditional methods have been enhanced by powerful algorithms like Convolutional Neural Networks (CNNs). Instead of manually specifying parameters, CNNs are designed to learn them automatically from large datasets. By leveraging sophisticated techniques, CNNs may effectively extract data from structured elements in documents, even when the image is scanned poorly with distortion and noise."
[78] FIG. 6 illustrates, seems to outline a process 600 for training a YOLO model to extract data from invoices in accordance with an embodiment of the present disclosure. Here's a breakdown of the steps:
• Data Capture (Step 602): Multiple invoice images are collected and used as the data set for training the model.
• Data Annotation (Step 604): These invoice images are annotated with labels and bounding boxes. Annotations can be for various invoice parameters like GST No., vendor name, address, amount, date, and time. Bounding boxes are likely used to highlight the specific location of this information within the invoice image.
• Data Splitting (Step 606): The data set is divided into three parts: a training set, a validation set, and a test set. The training set is used to train the model, the validation set is used to monitor the model's performance during training, and the test set is used to evaluate the model's accuracy after training.
• Model Setup and Training (Step 608 & 610): A YOLO model is set up, and the training set is fed into the model for training. During training, the model learns to recognize the patterns and relationships between the image data and the corresponding annotations.
• Model Validation (Step 612): The validation set is used to assess the model's performance. This helps identify any areas where the model may be struggling and allows for adjustments during training.
• Model Testing (Step 614): Once trained, the model's performance is evaluated using the test set. This provides an independent measure of how well the model can generalize to unseen invoice images.
• Model Fine-Tuning (Step 616): Based on the test results, the model may be further adjusted or fine-tuned to improve its accuracy in extracting data from invoices. This may involve going back and modifying the training data or the model architecture.
[79] Overall, this process creates a YOLO model that can be used to automatically extract data from invoice images.
[80] FIG. 7 is a flow diagram depicting a proposed method 700 for extracting data from the negotiable, in accordance with an embodiment of the present disclosure.
• At step 702 - acquiring, an image of the negotiable instruments via one or more processors of a computing device, and pre-processing the same by utilizing an imaging library to correct skew and remove noise, thereby enhancing image quality for further analysis and receiving a plurality of samples from the pre-processed image for subsequent analysis;
• At step 704- analysing, each of the plurality of samples from the pre-processed image via one or more processors, to identify at least one entity associated with the negotiable instrument by employing a learning model, including a trained dataset, and encrypting the identified entity using a secure encryption algorithm;
• At step 706- validating, the encrypted identified entity, via one or more processors, against predefined patterns to confirm its accuracy, using a secure decryption key to decrypt the entity before validation;
• At step 708- extracting data related to at least one identified entity from the negotiable instruments in the event of successful validation via one or more processors and encrypting the extracted data, including the vendor name and address, using a secure encryption algorithm and in the event of failed entity identification, extracting a vendor address from the negotiable instruments, and implementing fuzzy matching techniques to compare the extracted vendor address against a vendor master list for the purpose of identifying the corresponding vendor; and
• At step 710- updating the vendor information, including the vendor name and address, in a secure database, ensuring data privacy and integrity through encryption and access controls.
[81] Those skilled in the art would appreciate that embodiments of the present disclosure provides a negotiable instruments extraction system that employs a comprehensive workflow to ensure accurate and efficient text extraction. First, an image pre-processing module is used to correct skew, remove noise, and enhance contrast so as to optimize the negotiable instruments appearance for machine learning and OCR. This step improves field recognition and text extraction clarity. Further, a YOLO v7 model that is trained on a diverse dataset of images of the negotiable instruments with labelled fields, is generalized across various layouts. This is done to enable accurate detection of the entity without relying on template-based approaches. In addition, a module is used to isolate and remove irrelevant elements around fields like GST No. and vendor address, thereby retaining only relevant text for the OCR.
[82] This reduces visual artifacts and enhances OCR accuracy. Once fields are localized by the YOLOv7 model, the OCR may be applied to each specific region, while focusing on targeted areas to ensure precise data extraction. Field-specific validation checks may then verify an accuracy of extracted information, such as GST codes, Vendor address, and date formats to correct errors and ensure compliance with required standards. Together, these steps create a robust and a reliable system for extracting data from multiple negotiable instruments.
[83] In addition, the Automated Logo Detection System, powered by YOLOv7, brings several technical and commercial advantages, making it a valuable tool across multiple industries.
Technical Advantages:

• High Speed and Real-Time Detection: YOLOv7 is renowned for its fast- processing capabilities, enabling real-time logo detection in both images and videos. This is essential in dynamic environments like social media or live- streamed content, where logos need to be tracked quickly and efficiently.
• Robust Detection in Complex Scenarios: The system handles variations in logo presentation, such as different orientations, scales, partial occlusions, and even noisy backgrounds. YOLOv7’s multi-scale prediction ensures accurate detection even when logos are rotated, resized, or partially cut, making it robust in handling real-world challenges.
• Improved Accuracy: YOLOv7 features anchor-free detection and better feature extraction, resulting in higher accuracy. This helps reduce false positives and negatives, ensuring that only the relevant logos are detected and cataloged, improving overall system reliability.
• Scalability: The system can process large volumes of data without performance degradation. This scalability ensures that businesses can handle massive datasets, whether they are scanning thousands of product images or monitoring millions of social media posts. YOLOv7’s optimized architecture allows for efficient scaling while maintaining performance.
• Lightweight and Efficient: YOLOv7 is optimized for efficiency, meaning it can be deployed on devices with limited computational resources (such as mobile or edge devices) or in the cloud for large-scale operations. This
flexibility allows for seamless integration into a wide range of platforms and
devices.
• Versatility across Different Media Types: The system works across multiple types of media—images, videos, and even streaming content- making it adaptable for various industries, including e-commerce, advertising, surveillance, and branding. It also supports multiple formats of input, enhancing its utility.
Commercial Advantages:
• Cost Savings on Manual Labor: By automating the process of logo detection and cataloging, businesses can save significantly on the cost of manual labor. The system reduces the need for human workers to manually inspect images and videos, improving efficiency and cutting down operational expenses.
• Brand Protection and Compliance: The system helps companies protect their intellectual property by monitoring the use of their logos across different platforms. Businesses can automatically track logo usage, identify unauthorized use, and ensure compliance with brand guidelines, helping avoid potential legal issues and counterfeit concerns.
• Enhanced Marketing Insights: The automated logo detection system can be integrated into marketing analytics to track the visibility of logos in campaigns, advertisements, and media content. This provides businesses with actionable insights into how often their logos are appearing, their prominence, and overall brand exposure, allowing them to fine-tune marketing strategies.
• Scalable for Growing Businesses: As companies expand, their need for monitoring and cataloging logo appearances grows. The scalability of the system means it can easily handle this increased workload, whether it’s tracking logos across global social media platforms or processing logos from thousands of product images on e-commerce websites.
• Real-Time Monitoring in Dynamic Environments: In environments like social media, where content is rapidly changing, real-time logo detection provides immediate insights into brand mentions and logo use. This allows businesses to react quickly to trends, manage reputation, and track viral content, giving them a competitive edge in monitoring brand perception.
• Versatility across Industries: The system is applicable to a wide range of sectors. Advertising agencies can use it to evaluate brand visibility, legal teams can use it for brand protection, e-commerce companies can use it to ensure proper representation of brands on their platforms, and surveillance systems can use it to track logos for security purposes. This versatility opens multiple revenue streams for businesses deploying this solution.
• Competitive Advantage: Companies using the system can stay ahead of the competition by leveraging automation for brand monitoring and marketing. The faster processing times and real-time detection allow for quicker decision-making, enabling companies to respond to market trends and consumer behavior faster than competitors.
[84] It should be apparent to those skilled in the art that many more modifications besides those already described are possible without departing from the inventive concepts herein. The inventive subject matter, therefore, is not to be restricted except in the spirit of the appended claims. Moreover, in interpreting both the specification and the claims, all terms should be interpreted in the broadest possible manner consistent with the context. In particular, the terms “comprises” and “comprising” should be interpreted as referring to elements, components, or steps in a non-exclusive manner, indicating that the referenced elements, components, or steps may be present, or utilized, or combined with other elements, components, or steps that are not expressly referenced. Where the specification claims refers to at least one of something selected from the group consisting of A, B, C ….and N, the text should be interpreted as requiring only one element from the group, not A plus N, or B plus N, etc.
[85] The foregoing description of the specific embodiments will so fully reveal the general nature of the embodiments herein that others can, by applying current knowledge, readily modify and/or adapt for various applications such specific embodiments without departing from the generic concept, and, therefore, such adaptations and modifications should and are intended to be comprehended within the meaning and range of equivalents of the disclosed embodiments. It is to be understood that the phraseology or terminology employed herein is for the purpose of description and not of limitation. Therefore, while the embodiments herein have been described in terms of preferred embodiments, those skilled in the art will recognize that the embodiments herein can be practiced with modification within the spirit and scope of the appended claims.
[86] While the foregoing describes various embodiments of the invention, other and further embodiments of the invention may be devised without departing from the basic scope thereof. The scope of the invention is determined by the claims that follow. The invention is not limited to the described embodiments, versions or examples, which are included to enable a person having ordinary skill in the art to make and use the invention when combined with information and knowledge available to the person having ordinary skill in the art.
TECHNICAL ADVANTAGES
[87] The present invention provides a system and method for automated vendor identification and verification that utilizes advanced data capture techniques, machine learning, and data analytics to accurately identify and verify vendors based on information extracted from various documents, such as invoices, contracts, and purchase orders.
[88] The present invention provides a system and method for cleaning and standardizing vendor data to improve data quality and consistency, including techniques for handling variations in address formats, names, and other relevant information.
[89] The present invention provides a system and method for automated data extraction from vendor documents that employs advanced optical character recognition (OCR) and natural language processing (NLP) techniques to accurately extract vendor information, such as names, addresses, Tax IDs, and contact details.
[90] The present invention provides a system and method for vendor matching that can accurately identify and match vendors based on various attributes, including name similarity, address similarity, and other relevant factors.
[91] The present invention provides a system and method for protecting sensitive vendor information, including personal data and financial information.
[92] The present invention provides a system and method for integrating and synchronizing vendor data across multiple systems and databases.
[93] The present invention provides a system and a method that facilitates novel image processing technique designed to enhance the accuracy and efficiency of vendor detail extraction from multiple scanned documents.
[94] The present invention provides a system and a method incorporating machine learning model specifically trained to identify and extract vendor details from diverse document formats.
[95] The present invention provides a system and a method incorporating machine learning model specifically trained on a large dataset of labelled documents to achieve high accuracy and adaptability.
[96] The present invention provides a system and a method that could focus on a user- friendly interface allowing users to easily interact with the automated vendor detail extraction system.
[100] The present invention provides a system and a method for integrating the automated vendor detail extraction system with enterprise resource planning (ERP) systems that would streamline the workflow and improve data accuracy.

, Claims:We claim
1. An intelligent vendor identification method (700), wherein said method (700) comprising the steps of
acquiring, an image of the negotiable instruments via one or more processors (112) of a computing device (108), and
pre-processing the same by utilizing an imaging library to correct skew and remove noise, thereby enhancing image quality for further analysis and receiving a plurality of samples from the pre-processed image for subsequent analysis;
analysing, each of the plurality of samples from the pre-processed image via one or more processors (112), to identify at least one entity associated with the negotiable instrument by employing a learning model, including a trained dataset, and encrypting the identified entity using a secure encryption algorithm;
validating, the encrypted identified entity, via one or more processors (112), against predefined patterns to confirm its accuracy, using a secure decryption key to decrypt the entity before validation;
extracting data related to at least one identified entity from the negotiable instruments in the event of successful validation via one or more processors (112) and encrypting the extracted data, including the vendor name and address, using a secure encryption algorithm and in the event of failed entity identification, extracting a vendor address from the negotiable instruments, and implementing fuzzy matching techniques to compare the extracted vendor address against a vendor master list for the purpose of identifying the corresponding vendor; and
updating the vendor information, including the vendor name and address, in a secure database, ensuring data privacy and integrity through encryption and access controls.
2. The method (700) as claimed in claim 1, wherein an image processing library is employed to correct skew and eliminate noise from the image of the entity, thereby improving the overall quality of the images.
3. The method (700) as claimed in claim 2, wherein an OCR engine is utilized to identify at least one of the entity after correcting skew and removing noise from the image of the entity.
4. The method (700) as claimed in claim 1, wherein the trained dataset is utilized to detect and locate at least one of the entity of the negotiable instruments via employing bounding box mechanism.
5. The method (700) as claimed in claim 1, wherein data augmentation technique is employed to extract the data of at least one of the entity.
6. The method (700) as claimed in claim 5, wherein the data augmentation technique includes, but are not limited to, rotation, flipping, cropping, and intensity adjustments, to generate a diverse set of training images.
7. The method (700) as claimed in claim 1, wherein the trained dataset comprises the plurality of samples, where each of the plurality of samples are categorized into at least one of a training subset, a validation subset, and a test subset.
8. The method (700) as claimed in claim 1, wherein a performance metrics is applied on the trained dataset to track an impact of the training subset on the trained dataset.
9. An intelligent vendor identification system (110), wherein said system (110) comprising one or more processors (112) associated with a computing device; and
a memory (116) operatively coupled to the one or more processors (112), wherein the memory (116) comprises processor-executable instructions, which on execution, cause the one or more processors (112) to:
acquire, an image of the negotiable instruments via one or more processors (112) of a computing device (108), and pre-processing the same by utilizing an imaging library to correct skew and remove noise, thereby enhancing image quality for further analysis and receiving a plurality of samples from the pre-processed image for subsequent analysis;
analyse, each of the plurality of samples from the pre-processed image via one or more processors (112), to identify at least one entity associated with the negotiable instrument by employing a learning model, including a trained dataset, and encrypting the identified entity using a secure encryption algorithm;
validate, the encrypted identified entity, via one or more processors, against predefined patterns to confirm its accuracy, using a secure decryption key to decrypt the entity before validation; and extract, the data related to at least one identified entity from the negotiable instruments in the event of successful validation via one or more processors (112) and
encrypting the extracted data, including the vendor name and address, using a secure encryption algorithm and in the event of failed entity identification, extracting a vendor address from the negotiable instruments, and implementing fuzzy matching techniques to compare the extracted vendor address against a vendor master list for the purpose of identifying the corresponding vendor;
update, the vendor information, including the vendor name and address, in a secure database, ensuring data privacy and integrity through encryption and access controls.
10. The system (110) as claimed in claim 11, wherein an image processing library is employed to correct skew and eliminate noise from the image of the entity, thereby improving the overall quality of the images.
11. The system (110) as claimed in claim 12, wherein an OCR engine is utilized to identify at least one of the entity after correcting skew and removing noise from the image of the entity.
12. The system (110) as claimed in claim 11, wherein the trained dataset is utilized to detect and locate at least one of the entity of the negotiable instrument via employing bounding box mechanism.
13. The system (110) as claimed in claim 11, wherein data augmentation technique is employed to extract the data of at least one of the entity.
14. The system (110) as claimed in claim 13, wherein the data augmentation technique includes but are not limited to, rotation, flipping, cropping, and intensity adjustments, to generate a diverse set of training images.
15. The system (110) as claimed in claim 11, wherein the trained dataset comprises the plurality of samples, where each of the plurality of samples are categorized into at least one of a training subset, a validation subset, and a test subset.
16. The system (110) as claimed in claim 11, wherein a performance metrics is applied on the trained dataset to track an impact of the training subset on the trained dataset.
17. A non-transitory computer-readable medium incorporated in an intelligent vendor identification system (110), wherein said medium comprising processor executable instructions that cause a processor (112) to:
acquire, an image of the negotiable instruments via one or more processors (112) of a computing device (108), and pre-processing the same by utilizing an imaging library to correct skew and remove noise, thereby enhancing image quality for further analysis and receiving a plurality of samples from the pre-processed image for subsequent analysis;
analyse, each of the plurality of samples from the pre-processed image via one or more processors (112), to identify at least one entity associated with the negotiable instrument by employing a learning model, including a trained dataset, and encrypting the identified entity using a secure encryption algorithm;
validate, the encrypted identified entity, via one or more processors (112), against predefined patterns to confirm its accuracy, using a secure decryption key to decrypt the entity before validation; and
extract, the data related to at least one identified entity from the negotiable instruments in the event of successful validation via one or more processors (112) and encrypting the extracted data, including the vendor name and address, using a secure encryption algorithm and in the event of failed entity identification, extracting a vendor address from the negotiable instruments, and implementing fuzzy matching techniques to compare the extracted vendor address against a vendor master list for the purpose of identifying the corresponding vendor;
update, the vendor information, including the vendor name and address, in a secure database, ensuring data privacy and integrity through encryption and access controls.