Published Paper Details:

Identity verification plays a pivotal role in the contemporary digital landscape, especially within financial and online services. The adoption of Know Your Customer (KYC) processes has become imperative to ensure secure and reliable interactions. This survey paper explores the integration of Optical Character Recognition (OCR) techniques, with a specific focus on the Convolutional Text Proposal Network (CTPN), as a machine learning approach for enhancing KYC verification. Traditional KYC methods often suffer from inefficiencies in terms of speed, accuracy, and scalability. The manual verification processes inherent in these methods not only consume valuable time but are also susceptible to errors. The advent of OCR technology has revolutionized the KYC landscape by automating the extraction and interpretation of textual information from diverse documents such as passports and driver's licenses. The survey begins by highlighting the limitations of conventional KYC methods, underscoring the need for advanced automated solutions. It provides an extensive overview of OCR, elucidating its significance in extracting text from a variety of documents while addressing challenges related to document formats, languages, and image qualities. Machine learning emerges as a critical enabler for OCR, allowing for continuous improvement in accuracy and adaptability to evolving requirements. A substantial portion of the paper is dedicated to introducing and exploring the capabilities of the Convolutional Text Proposal Network (CTPN) as an innovative machine learning approach tailored for OCR-based KYC verification. CTPN's ability to accurately localize and recognize text within images positions it as a promising solution for improving the precision and efficiency of KYC processes. The survey concludes with an evaluation of the current state of OCR-based KYC verification, identifying areas of success and potential challenges. It serves as a roadmap for future research and development, emphasizing the importance of ongoing advancements in machine learning and OCR technologies to meet the evolving demands of identity verification in the digital age.