Tools
Tools
Sen, Shivam (2025) AuraFlow: A Hybrid Artificial Intelligence System for Yoga Pose Correction Using Joint-Level Analysis and Generative Feedback. Masters thesis, Dublin, National College of Ireland.
Preview |
PDF (Master of Science)
Download (2MB) | Preview |
Preview |
PDF (Configuration Manual)
Download (2MB) | Preview |
Abstract
Digital wellness technology has emerged to offer recent research opportunities, especially in the challenging health habits such as yoga where granular real-time biomechanical feedback has been historically missing in the easily accessible systems. Although currently available AI fitness apps may often predict the pose with high accuracy, these apps lack the interpretable joint-level corrective feedback that would allow ensuring user safety and performance effectiveness on a fundamental level. The following thesis is an amendment to this limitation because it suggests formulates and tests a hybridized AI system that acts as a smart yoga coach. The most important scientific contribution is the Joint Correctness Index (JCI), a new explainable algorithm designed for providing a real-time per-joint alignment score. A system called AuraFlow is to implement and validate this framework. This was done on a methodology basis by comparing five machine learning classifiers on the Yoga-82 dataset. The Random Forest model was identified as the best out of the models used with the accuracy of 92.13% in 43 poses. Then JCI algorithm is used on the categorized poses. Also a Large Language Model (LLM) is incorporated to convert these quantitative JCI scores into particular and eloquent coaching responses. Another interesting discovery which follows this study is that a simple rule-based algorithm (JCI) can be extend to a conventional machine learning model so that it delivers accurate and real time performance (roughly 9-12 frames per second) on a standard consumer hardware. This study gives a viable framework to formulate more viable and consumer-friendly AI-driven coaching solutions.
| Item Type: | Thesis (Masters) |
|---|---|
| Supervisors: | Name Email Tomer, Vikas UNSPECIFIED |
| Subjects: | G Geography. Anthropology. Recreation > GV Recreation Leisure Q Science > QH Natural history > QH301 Biology > Methods of research. Technique. Experimental biology > Data processing. Bioinformatics > Artificial intelligence Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Artificial intelligence R Medicine > RA Public aspects of medicine > RA0421 Public health. Hygiene. Preventive Medicine > Personal Health and Hygiene |
| Divisions: | School of Computing > Master of Science in Data Analytics |
| Depositing User: | Ciara O'Brien |
| Date Deposited: | 03 Jul 2026 10:50 |
| Last Modified: | 03 Jul 2026 10:50 |
| URI: | https://norma.ncirl.ie/id/eprint/9460 |
Actions (login required)
 |
View Item |