View Proposal


Proposer
Aigerim Khairullina
Title
AI-Powered Student Engagement Monitoring Platform
Goal
To develop a localized, privacy-preserving software platform that monitors and classifies student engagement levels in real-time during a simulated learning environment.
Description
Using a standard laptop webcam, the system captures live video feeds via OpenCV, tracks multiple participants with unique IDs using the YOLO framework, and extracts detailed facial coordinates (head pose, eye state, and gaze direction) via the MediaPipe Face Mesh library. The core of the project involves building an autonomous pipeline where the student selects and implements an appropriate Machine Learning (e.g., Random Forest, XGBoost) or Deep Learning (e.g., LSTM Neural Networks) classification model using Scikit-learn or PyTorch/TensorFlow to accurately predict focus metrics and render them on an interactive dashboard.
Resources
Background
Url
Difficulty Level
High
Ethical Approval
Full
Number Of Students
1
Supervisor
Aigerim Khairullina
Keywords
machine learning, mediapipe, yolo, engagement detection, real-time analytics
Degrees
Bachelor of Science in Computational Sciences and Software Engineering