NLP-Powered System for Academic Paper Evaluation and Performance Scoring

This paper presents an intelligent framework for automated evaluation and scoring of student answer sheets using a combination of image processing, optical character recognition (OCR), and machine learning techniques. The system is designed to reduce manual effort in assessing both objective and subjective examination papers. Initially, users can upload scanned images of answer sheets through a web-based interface. The system applies image processing techniques to enhance the quality of the uploaded documents and extract relevant textual content using OCR. For multiple-choice questions, the system identifies marked responses and compares them with the correct answers to compute scores accurately. In the case of subjective answers, the extracted text is compared with model answers using a K-Nearest Neighbors (KNN) algorithm to estimate similarity and assign marks accordingly. The platform includes modules for user registration, login, MCQ evaluation, and subjective answer assessment, ensuring smooth interaction. The evaluation results are presented clearly, highlighting correct and incorrect responses visually. Experimental outcomes demonstrate that the system provides consistent and reliable scoring. The approach minimizes human bias and speeds up the evaluation process. The integration of artificial intelligence enhances the accuracy of subjective answer assessment. The framework is scalable and can be adapted to different types of examination systems. Overall, the proposed system offers an efficient solution for automated academic evaluation. Keywords— OCR, Image Processing, Automated Evaluation, KNN Algorithm, Answer Sheet Scoring, Educational Technology, Machine Learning, Text Extraction, Smart Assessment System