FEATURE EXTRACTION AND CLASSIFICATION OF CERTIFICATES USING OCR

Abstract

The paper aims to create a feature extraction and  classification system for certificates based on Optical  Character Recognition (OCR) technology. The system seeks to  automate the process of certificate classification and activity point assignment by extracting pertinent textual information such as student names, course titles, issuing organizations, and dates from scanned certificate images. Using sophisticated OCR algorithms, such as EasyOCR and OpenCV, the system processes images beforehand to improve the accuracy of text recognition. Then the extracted text is processed with natural language processing (NLP) for categorizing into pre-specified types like course completion, workshop attendance, and  honors. This mechanized process significantly lessens human 
intervention and error involved in certificate validation processes, making it a scalable solution for academic institutions and organizations like KTU, MG University etc.