Objectives
Automate Student Engagement Detection: Use computer vision to detect student faces in classroom videos automatically. Classify Engagement Levels: Classify each detected face as engaged or not engaged using emotion recognition. Calculate Total Engagement Score (TES): Provide a quantitative score representing overall class engagement during a session. Support Teacher Evaluation: Assist in evaluating teacher effectiveness based on real-time classroom engagement analytics. Develop a User-Friendly Web Interface: Build an interactive and accessible platform for uploading videos and viewing results.
Socio-Economic Benefit
Improved Quality of Education: By analyzing student engagement, teachers can adapt their teaching methods, leading to more effective learning outcomes. Teacher Accountability & Growth: Helps identify teaching gaps and encourages professional development through constructive feedback. Data-Driven Decision Making: Schools and institutions can make informed decisions regarding teaching strategies and resource allocation. Cost-Effective Evaluation: Reduces the need for manual classroom monitoring or external evaluators, saving time and resources. Bridging the Education Gap: Supports under-resourced schools by offering an affordable and scalable evaluation system. Promotes Digital Transformation in Education: Encourages the adoption of AI and smart technologies in educational environments, aligning with global trends.
Methodologies
Requirement Analysis: Gathered system requirements from academic supervisors and defined project goals. Dataset Preparation: Collected and preprocessed custom datasets for two classes: engaged and not_engaged student faces. Model Development: Detection: Used YOLOv11 to detect student faces in video frames. Classification: Applied a deep learning classifier to categorize each face as engaged or not_engaged. Integration: Combined the detection and classification models to process videos and compute the Total Engagement Score (TES). Web Development: Frontend: Built using React and Tailwind CSS for user interaction. Backend: Developed with FastAPI to serve the ML models and handle requests. Deployment: Hosted the backend on Render and frontend on Netlify for real-time video analysis via the web. Testing & Evaluation: Performed unit testing, model validation (accuracy, confusion matrix), and gathered feedback from users.
Outcome
Functional AI-Based Web Application: Successfully developed and deployed a web system that analyzes classroom videos to evaluate student engagement. Accurate Face Detection and Classification: Achieved high accuracy using YOLOv11 for face detection and deep learning for emotion-based engagement classification. Total Engagement Score (TES): Implemented TES to provide a clear, quantitative metric of classroom engagement (Excellent, Moderate, or Low). Positive User Feedback: Users (e.g., teachers and evaluators) found the system helpful for improving teaching strategies and student participation.
Project Team Members
| Registration# | Name |
|---|---|
| CU-2011-2021 | Raza Hussain |
| CU-1967-2021 | Zain Ul Abideen |
PROJECT GALLERY
Smart Teacher Evaluation System
22 Jun, 2025 09:28 pm
