Automated Evaluation (Context based) of Handwritten Assessment

Let's See

Automated Evaluation (Context based) of Handwritten Assessment

This project introduces an AI-based system for the automated evaluation of handwritten assessments with a focus on contextual understanding. Using Moondream AI for Optical Character Recognition (OCR), the system extracts text from scanned images of both the student’s and the teacher’s handwritten responses. A Large Language Model (LLM) then acts as a contextual judge, comparing the student’s answer to the key not just for keyword similarity, but for semantic meaning and reasoning. The system generates a score along with detailed feedback, highlighting the rationale and pointing out specific errors. This approach modernizes assessment by integrating OCR technology with intelligent, context-aware evaluation.

Keywords: Context-Based Grading, OCR, Transformer, MoonDream, LLM as Judge, Vision Transformer, AI in Education, Handwritten Assessment
Tools: VS code, MoonDream, Ollama, Collab, LLM, Streamlit, NLP,Qwen3
Department: Department of Electrical Engineering
Project Poster
Blog
Project Team Members
Name Email CV
Muhammad Saad saad2021@namal.edu.pk
Muhammad Makki makki2021@namal.edu.pk