Publication

Abstract : The process of customer verification has become a mandatory procedure to follow for a large number of sales and financial organizations. Traditionally, it has required collecting the required data from the uploaded images of identity cards and then manually entering them into a system which makes it a resource-intensive and error-prone task. Automation of this entire process is extremely beneficial for these organizations. The paper proposes an end-to-end system for preprocessing, text detection, and text recognition of identity cards by identifying the suitable methods and model architectures for each of the components in the system. The end-to-end system presented in the paper gives an accuracy of 93%. This paper is useful for researchers who want to work on improving the automation of text extraction from images, optical character recognition, and other similar use-cases © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.