Volume 1 - Issue 4, November - December 2025

This paper presents an AI-powered learning platform that integrates Optical Character Recognition (OCR) with deep learning and symbolic computation to interpret and solve handwritten mathematical equations. The system captures handwritten input via camera or file upload, preprocesses it, recognizes symbols using a hybrid CNN-LSTM architecture, parses the recognized symbols into structured mathematical expressions, and computes step-wise solutions using symbolic solvers. Experimental results demonstrate 96.5% symbol-level accuracy, 90.2% expression-level recognition accuracy, and 96.8% solution correctness across diverse mathematical problems. The system significantly improves learning outcomes, with users showing an average 23.7 percentage point improvement in mathematical understanding. Comprehensive evaluation on multiple datasets and user studies validate the system’s effectiveness in educational contexts.