Virtual Yoga Instructor with Real-Time Feedback
Date of Award
12-12-2024
Degree Name
M.S. in Computer Science
Department
Department of Computer Science
Advisor/Chair
Tam Nguyen
Abstract
We introduce the virtual yoga instructor which is a gamified application designed to make yoga learning engaging and accessible. The application consists of two core components: a virtual instructor and feedback module via pose estimation. The virtual instructor, implemented in Unity with an animated 3D humanoid model, represents a human yoga instructor. The instructor demonstrates each pose and guides users through an instruction and feedback panel. Meanwhile, the feedback module utilizes MediaPipe Pose to detect and analyze body landmarks. It calculates critical angles across various body parts to assess posture accuracy. By comparing the user’s angles with predefined reference angles for each yoga pose, the system generates a score and provides feedback, guiding the user to correct their posture and match the reference more closely. Communication between the pose estimation module and the virtual instructor is achieved through socket programming, enabling the application to run on the same device or across multiple devices on the same network. User evaluations were conducted with three versions of the application to assess performance based on ease of use, informativeness, engagement, retention, and interaction. The results indicate that our proposed work provides a convenient solution for practicing yoga safely at home, offering immediate corrective guidance to improve posture.
Keywords
Computer Science
Rights Statement
Copyright © 2024, author.
Recommended Citation
Chithambaram, Nikesh, "Virtual Yoga Instructor with Real-Time Feedback" (2024). Graduate Theses and Dissertations. 7466.
https://ecommons.udayton.edu/graduate_theses/7466
