Scholarly open access journals, Peer-reviewed, and Refereed Journals, Impact factor 8.14 (Calculate by google scholar and Semantic Scholar | AI-Powered Research Tool) , Multidisciplinary, Monthly, Indexing in all major database & Metadata, Citation Generator, Digital Object Identifier(DOI)
Human Action Recognition (HAR) has recently become an essential technology in contemporary computer vision, supporting important applications from fully automated video surveillance and anomaly detection to sophisticated human-computer interaction and senior care monitoring. Nonetheless, precise recognition remains a challenging task due to the intricate relationship between spatial visual data and temporal motion patterns, as well as differences in camera viewpoints, illumination, and presence of background clutter. Conventional methods tend to fail in capturing long-term temporal dependencies or incur unaffordable computational complexity in 3D Convolutional Neural Networks(3D-CNNs).
Human Action Recognition (HAR) has recently become an essential technology in contemporary computer vision, supporting crucial applications from fully automated video surveillance and anomaly detection to sophisticated human-computer interaction and senior care monitoring. Nonetheless, precise action recognition has been a challenging task due to the intricate coupling of spatial visual data and temporal motion patterns, as well as variations in camera viewpoints, illumination, and background complexity. Conventional methods have difficulty in capturing long-term temporal dependencies or incur unaffordable computational complexity in 3D Convolutional Neural Networks (3D-CNNs).
Experimental evaluations conducted on the UCF-101 benchmark dataset demonstrate the efficacy of our approach. The proposed model achieves competitive classification accuracy while maintaining computational efficiency superior to fully 3D architectures. These results suggest that our attention-augmented CNN-LSTM framework offers a scalable and effective solution suitable for real-time action recognition in resource-constrained environments.
Keywords:
Human Action Recognition (HAR), Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM), Spatiotemporal Feature Extraction, Attention Mechanism, Computer Vision, Deep Learning, Video Classification, Recurrent Neural Networks (RNN).
Cite Article:
"Smart Human Computer Interaction: Gesture and Action Recognition using Sequential Deep Learning", International Journal for Research Trends and Innovation (www.ijrti.org), ISSN:2456-3315, Vol.11, Issue 2, page no.a707-a714, February-2026, Available :http://www.ijrti.org/papers/IJRTI2602094.pdf
Downloads:
00081
ISSN:
2456-3315 | IMPACT FACTOR: 8.14 Calculated By Google Scholar| ESTD YEAR: 2016
An International Scholarly Open Access Journal, Peer-Reviewed, Refereed Journal Impact Factor 8.14 Calculate by Google Scholar and Semantic Scholar | AI-Powered Research Tool, Multidisciplinary, Monthly, Multilanguage Journal Indexing in All Major Database & Metadata, Citation Generator