Hi, I'm Manik Hossain.
I build quality into every release.

Physiotherapy Exercise Classification by Exploring Knowledge Distillation with Attention Modules

Presented at NCIM-2023, Dhaka University of Engineering & Technology · Available on IEEE Xplore

Abstract

Automated AI-based physiotherapy assistive systems are gaining popularity owing to their practical necessity, particularly for elderly people. The application of such an assistive system is highly reliant on the embedded AI model that needs to correctly classify exercise videos while being lightweight in order to be used for real-time purposes. In this paper, we take this issue into account and present a deep learning framework for physiotherapy video classification that preserves competitively high accuracy while being lightweight as well. For this, we explore knowledge distillation techniques for model compression purposes. We use both online and self-distillation techniques, and use Convolutional Long Short-Term Memory (ConvLSTM) and Long-term Recurrent Convolutional Network (LRCN) as the base networks. Moreover, we utilize different attention modules and explore different architectural combinations to further improve classification performance. We conduct detailed experimentation by exploring different combinations of distillation techniques and attention modules, along with tweaking their hyperparameters. The best results found from our experiments are reported, and those combinations are recommended for further use in relevant AI assistive systems.