Citation

Salem, Mohammed Ahmed and Besar, Rosli and Abdulkareem, Abdulaziz Mohsen and Salleh Abas, Fazly and Aziz, Azlan Abdul and Amir Hamzah, Nur Asyiqin and Ab Aziz, Nor Azlina (2025) Optimized Motion Recognition for Work Environment Using Improved Artificial and Swarm Intelligence. In: 2025 Multimedia University Engineering Conference, MECON 2025, 21 July 2025 - 23 July 2025, Cyberjaya, Malaysia.

Text 453.pdf - Published Version Restricted to Repository staff only Download (1MB)

Abstract

Workplace productivity is influenced by motion efficiency, activity monitoring, and real-time adaptation to dynamic environments. Traditional motion detection systems frequently struggle with high false alarm rates and poor adaptability to complex work activities. This research proposes an AI-based motion recognition algorithm enhanced with Particle Swarm Optimization (PSO) to improve accuracy and computational efficiency in workplace activity recognition. The framework integrates deep learning models, including Convolutional Neural Networks-Long Short-Term Memory Networks (CNN-LSTM) and Long-term Recurrent Convolutional Networks (LRCN), with PSO for hyperparameter optimization, enabling automated tuning of learning rate, dropout rate, and batch size. Additionally, a customized Workplace Human Activity Recognition (CWHAR) dataset was developed to enhance training robustness. A total of 24 activity classes were categorized into six different groups to ensure a structured and diverse dataset for model training and evaluation. The results demonstrate that the PSO-optimized CNN+LRCN model achieved 95.83% accuracy, compared to 83.86% without PSO (14.3% improvement), while the PSO-optimized CNN+ConvLSTM model reached 90% accuracy, compared to 78.42% without PSO (14.8% improvement). These findings confirm that PSO-based hyperparameter optimization significantly enhances model accuracy and robustness, making it a promising approach for intelligent workplace activity recognition.

Item Type:	Conference or Workshop Item (Paper)
Uncontrolled Keywords:	Particle Swarm Optimization (PSO), Convolutional Neural Networks-Long Short-Term Memory Networks (CNN LSTM) and Long-term Recurrent Convolutional Networks (LRCN).
Subjects:	H Social Sciences > HD Industries. Land use. Labor > HD28-70 Management. Industrial Management > HD30.2 Electronic data processing. Information technology. Including artificial intelligence and knowledge management
Divisions:	Faculty of Engineering and Technology (FET)
Depositing User:	Ms Suzilawati Abu Samah
Date Deposited:	19 Mar 2026 02:33
Last Modified:	19 Mar 2026 02:33
URII:	http://shdl.mmu.edu.my/id/eprint/15615

Downloads

Edit (login required)