Year
2022
PLATFORM
IRJMETS
Category
Research
focus
Human Motion Analytics
Focused on translating physical motion into mathematical feedback, the research builds on concepts from pose estimation, classification, and exercise validation. I studied and applied different ML algorithms (Random Forest, Logistic Regression, ANN), building a reliable system that mirrors trainer-like feedback in digital environments.
The interface and experience were designed for intuitive understanding—even by users without technical background. I ensured that posture feedback and repetition counters were clean, responsive, and aligned with how users interact during physical activity.
We implemented MediaPipe's 33-point body landmarking system, extracted real-time vectors, and mapped them into a neural network classifier. I co-engineered the backend using Python with visualization layers for performance tracking and training flow.
This project proves how AI can enhance health and well-being. By embedding ML directly into the user’s environment, we created a tool that’s accurate, educational, and accessible across platforms. The study lays groundwork for future mobile integration and broader accessibility in digital fitness ecosystems.
