Digitized Paper Evaluation for Universities using Machine Learning Technique

Automating the assessment of handwritten university papers is a crucial advancement in today's digital era. This study focuses on leveraging Handwriting Text Recognition (HTR) and Machine Learning (ML) techniques to streamline the evaluation process. By extracting key features and automating the grading process, this research aims to enhance efficiency and accuracy. The project emphasizes two primary objectives: feature extraction and automated calculation to replace manual evaluation methods. By employing a 70:30 split and training on 1000 images, the system undergoes rigorous filtering, preprocessing, region of interest (ROI) identification, segmentation, and validation based on ground truth data. Despite challenges like strike-throughs, overlapping letters, and inconsistent ink, this initiative ultimately aims to overcome the limitations of traditional paper evaluation methods.