Citation

Bari, Ahsanul and Al Farid, Fahmid and Sarker, Md. Tanjil and Mansor, Sarina and Abdul Karim, Hezerul and Bhuiyan, Md Roman and Bannah, Hasanul (2024) Lightweight Deep Learning for Human Activity Recognition in Ambient Assisted Living. In: 2024 Multimedia University Engineering Conference (MECON), 23-25 July 2024, Cyberjaya, Malaysia.

Text Lightweight Deep Learning for Human Activity Recognition in Ambient Assisted Living.pdf - Published Version Restricted to Repository staff only Download (376kB)

Abstract

This Lightweight Deep Learning (LDL) for MultiView Human Activity Recognition in Ambient Assisted Living Systems can significantly improve the conditions of daily activities for people living with the elderly, disabled, or simply needing assistance. Human Activity Recognition (HAR) is essential for these systems; currently, the task of identifying the wide array of human activities remains complicated. This study introduces an improved deep learning solution for recognizing human activity from multiple viewpoints in AAL environments. By utilizing skeleton data from various perspectives, such as the RHM-HAR-SK dataset, we demonstrate useful improvements in accuracy for activity recognition when compared with singleview methods. Our LeNet and Vision Transformer models have been carefully designed to excel with skeletal data, delivering outstanding performance while keeping computational requirements to a minimum. In this research study also highlights the benefits of multi-view human activity recognition in ambient assisted living and emphasizes the effectiveness of efficient deep learning models in enabling assistive technologies that are real-time and resource friendly. These advancements support more independence and safety by supporting the development of AAL systems that are more sensitive and intelligent

Item Type:	Conference or Workshop Item (Paper)
Uncontrolled Keywords:	Lightweight Deep Learning
Subjects:	Q Science > Q Science (General) > Q300-390 Cybernetics
Divisions:	Faculty of Engineering (FOE)
Depositing User:	Ms Nurul Iqtiani Ahmad
Date Deposited:	12 Feb 2025 00:24
Last Modified:	12 Feb 2025 00:24
URII:	http://shdl.mmu.edu.my/id/eprint/13411

Downloads

Edit (login required)