AUTOMATED DOCUMENT VERIFICATION SYSTEM

Authors:

Tanishka Powar, Divya Patil, Shreeya Kashid, Triveni Madhale, Mrs. T. V. Deokar

D Y Patil College of engineering and Technology

Department of Computer Science and Engineering (Data Science)

Abstract - Currently, the manual process for the verification of educational certificates, mark sheets, caste/EWS/PWD certificates, and other documents used for recruitment processes is tedious, error-prone, and inadequate when it comes to processing linguistic diversity. For solving these problems, we propose an Intelligent Document Processing (IDP) solution based on Machine Learning, Deep Learning, and NLP that makes use of recent improvements made in optical character recognition (OCR). The suggested solution works according to a two-phased discrimination–generation approach where an optimized fine-tuned machine learning model helps extract data from documents, along with informing applicants about their selection. This is assured to work with high accuracy, as part of its second phase, where the process is carried out to ensure accuracy screening and validation of all documents. With a target accuracy level of Three Sigma (99.73%), the whole process will take place within 3 seconds.

Key Words: Document Verification, OCR, NLP, Machine Learning, Intelligent Document Processing, Data Validation