A Real-time Machine Learning Framework for Smart Home-based Yoga Teaching System - Conference paper

Aug 4, 2026

Home-based yoga practice has grown significantly after Covid-19, yet exercising without a trainer often leads to incorrect pose execution that can cause muscle damage. To address real-time computational constraints in smart-home instruction, this research presents a real-time machine learning framework combining pose estimation, pose classification, and immediate feedback. Five yoga poses are supported, with BlazePose generating 3D landmark points for downstream classification. Multiple classical and neural models are evaluated on accuracy, latency, and model size, and the XGBoost classifier achieves the best overall performance. The framework enables real-time pose correction and can be integrated into mobile applications for unsupervised practice.

LanguageEnglish

CategoryTechnology

FormatPDF

Pages10

Reading Time25 min

AccessFree to download

Abstract

Introduction

Motivation and challenges

Pose estimation background (BlazePose)

Research aim and contribution

Technological University Dublin ARROW@TU Dublin | Conference papers | School of Computer Science | | --- | --- | | 2023
A Real-time Machine Learning Framework for Smart Home-based Yoga Teaching System
Jothika Sunney
National College of Ireland, Ireland
Musfira Jilani
National College of Ireland, Ireland
Pramod Pathak
Technological University Dublin, Ireland, [pramod.pathak@tudublin.ie](pramod.pathak@tudublin.ie)[ ](pramod.pathak@tudublin.ie)See next page for additional authors
Follow this and additional works at: [https://arrow.tudublin.ie/scschcomcon](https://arrow.tudublin.ie/scschcomcon)
Part of the Computer Engineering Commons | | Recommended Citation Sunney, Jothika; Jilani, Musfira; Pathak, Pramod; and Stynes, Paul, "A Real-time Machine Learning Framework for Smart Home-based Yoga Teaching System" (2023) . Conference papers. 402. [https://arrow.tudublin.ie/scschcomcon/402](https://arrow.tudublin.ie/scschcomcon/402) This Conference Paper is brought to you for free and open access by the School of Computer Science at ARROW@TU Dublin. It has been accepted for inclusion in Conference papers by an authorized administrator of ARROW@TU Dublin. For more information, please contact [arrow.admin@tudublin.ie](arrow.admin@tudublin.ie), [aisling.coyne@tudublin.ie](aisling.coyne@tudublin.ie), [gerard.connolly@tudublin.ie](gerard.connolly@tudublin.ie), [vera.ki](vera.ki)[lshaw@tudublin.ie](lshaw@tudublin.ie). This work is licensed under a Creative Commons Attribution-Share Alike 4.0 International License. Funder: This research received no external funding Authors Jothika Sunney, Musfira Jilani, Pramod Pathak, and Paul Stynes This conference paper is available at ARROW@TU Dublin: [https://arrow.tudublin.ie/scschcomcon/402](https://arrow.tudublin.ie/scschcomcon/402) 2023 7th International Conference on Machine Vision and Information Technology (CMVIT) | 978-1-6654-6485-7/23/$31.00 ©2023 IEEE | DOI: 10. 1 109/CMVIT57620.2023.00029 2023 7th International Conference on Machine Vision and Information Technology (CMVIT) A Real-time Machine Learning Framework for Smart Home-based Yoga Teaching System 1st Jothika Sunney School of Computing National College of Ireland Dublin, Ireland [jothikasunney@gmail.com](jothikasunney@gmail.com) 2nd Musﬁra Jilani School of Computing National College of Ireland Dublin, Ireland musﬁ[ra.jilani@ncirl.ie](ra.jilani@ncirl.ie) 3rd Pramod Pathak Faculty of Digital&Data Technological University Dublin Dublin, Ireland [pramod.pathak@TUDublin.ie](pramod.pathak@TUDublin.ie) 4th Paul Stynes School of Computing National College of Ireland Dublin, Ireland [paul.stynes@ncirl.ie](paul.stynes@ncirl.ie) Abstract—Practicing yoga poses in a home-based environment has increased due to Covid19. Yoga poses without a trainer can be challenging, and incorrect yoga poses can cause muscle damage. Smart home-based yoga teaching systems may aid in performing accurate yoga poses. However, the challenge with such systems is the computational time required to detect yoga poses. This research proposes a real-time machine learning framework for teaching accurate yoga poses. It combines a pose estimation model, a pose classiﬁcation model, and a real-time feedback mechanism. The dataset consists of ﬁve popular yoga poses namely the downdog pose, the tree pose, the goddess pose, the plank pose, and the warrior pose. The BlazePose model was used for yoga pose estimation which transforms the image data into 3D landmark points. The output of the pose estimation model was then passed to the pose classiﬁcation model for yoga pose detection. Four machine learning classiﬁers namely, Random Forest, Support Vector Machine, XGBoost, Decision Tree, and two neural network classiﬁers LSTM and CNN were evaluated based on accuracy, latency and size. Results demonstrate that XGBoost outperforms other models with an accuracy of 95.14 percentage, latency of 8 ms, and size of 513 KB. The output of

FAQs

What problem does the proposed smart home yoga teaching system address?

It targets the difficulty and risk of practicing incorrect yoga poses at home, where lack of a trainer can lead to muscle damage. It also addresses the computational time required for pose detection in real-time applications.

What performance results does the best model achieve?

XGBoost outperforms the other classifiers, reaching 95.14% accuracy with 8 ms latency and a model size of 513 KB. Its predicted output is used to provide real-time feedback to correct yoga poses.

Aug 4, 2026

Upload by Olivia Brown