A Novel Hybrid Model for Spatio-Temporal Human Activity Recognition in Video Streams

Rashidkhan R Pathan; Hetal S Chavda

doi:10.36948/ijfmr.2025.v07i05.57511

A Novel Hybrid Model for Spatio-Temporal Human Activity Recognition in Video Streams

Author(s)	Prof. Rashidkhan R Pathan, Prof. Hetal S Chavda
Country	India
Abstract	Human Activity Recognition (HAR) from video streams has become an important research area in computer vision because of its vast applications in surveillance, health monitoring, sports analysis, and human-machine interaction. Conventional approaches based on either spatial or temporal features alone tend to fail to make accurate decisions in challenging real-world situations. In this work, we introduce a new hybrid deep learning architecture that successfully combines Convolutional Neural Networks (CNNs) for spatial feature extraction with Gated Recurrent Units (GRUs) for temporal sequence modeling. Our model captures both appearance and motion dynamics intrinsic in human actions through consecutive video frames. The suggested model is tested on three benchmark video datasets, namely UCF11, KTH, and UCF Sports Action, containing various human actions and adverse environmental factors like camera movements, occlusion, and changed angles. Experimental evaluations show that the suggested CNN-GRU model attains higher accuracy of 97.37% on the UCF11 dataset, 97.4% on KTH, and 97.5% on UCF Sports compared to some current deep-learning baselines. We also use methods like data augmentation, dropout regularization, and learning rate scheduling to increase generalization and avoid overfitting. The results demonstrate the power of our compact yet robust hybrid model in spatio-temporal video classification tasks and offer a scalable solution for real-world HAR systems. Future directions involve expanding the model to real-time recognition and investigating transformer-based optimization for better context modeling.
Keywords	Human Activity Recognition (HAR), Spatio-Temporal Features, Video Classification, Convolutional Neural Network (CNN), Gated Recurrent Unit (GRU), Deep Learning, Action Recognition, Video Streams, Temporal Modeling, Computer Vision
Field	Computer Applications
Published In	Volume 7, Issue 5, September-October 2025
Published On	2025-10-22
DOI	https://doi.org/10.36948/ijfmr.2025.v07i05.57511

View / Download PDF File

E-ISSN 2582-2160

doi

CrossRef DOI is assigned to each research paper published in our journal.

IJFMR DOI prefix is
10.36948/ijfmr

Downloads

Research Paper Format Copyright Permission Form and Undertaking Form Cover Page Vol 8 Isu 2 Cover Page Vol 8 Isu 1 Cover Page Vol 7 Isu 6

All research papers published on this website are licensed under Creative Commons Attribution-ShareAlike 4.0 International License, and all rights belong to their respective authors/researchers.

CC-BY-SA

About IJFMR Fees & Payment Current Issue Publication Archive	Submit Research Paper Track Submission Status Publication Guidelines Publication Ethics Peer Review & Plagiarism	Join as a Reviewer Editors & Reviewers Reviewer Referral Program Get Reviewer Membership Certi.	Website/Journal Policies Usage Policy Content Policies Privacy Policy

Contact Us		+91-9687-828-838	editor@ijfmr.com

International Journal For Multidisciplinary Research

A Widely Indexed Open Access Peer Reviewed Multidisciplinary Bi-monthly Scholarly International Journal

A Novel Hybrid Model for Spatio-Temporal Human Activity Recognition in Video Streams

Share this