Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{173776,
        author = {Shaik Hafijulla Irshad and S.Suneetha},
        title = {Detecting Video Anomalies through Deep Learning Using Optimized Attention-Enhanced Auto encoders},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {11},
        number = {10},
        pages = {1671-1681},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=173776},
        abstract = {Automated monitoring systems, security, and surveillance all depend on video anomaly detection. Conventional methods frequently have limited applicability in complicated situations and significant false alarm rates. Here, we provide a novel deep learning framework that enhances anomaly detection performance in video sequences by utilizing Optimised Attention-Enhanced Auto encoders. By incorporating an attention mechanism into an auto encoder architecture, our approach improves the learning of spatial-temporal representations by allowing the model to suppress irrelevant information and concentrate on important regions. To increase accuracy and efficiency, we also use an optimization technique to fine-tune network setups and hyper parameters. Experimental tests on benchmark datasets, such as ShanghaiTech and UCSD Ped2, show that our strategy drastically lowers false positives while achieving better performance than state-of-the-art techniques. Optimised Attention-Enhanced Auto encoders are a revolutionary deep learning-based framework that we present in this work to detect video anomalies. The attention mechanism in our model is integrated into auto encoder architecture to effectively learn normal patterns in video sequences and improve the representation of spatial-temporal information. By concentrating on the most pertinent areas, the attention mechanism helps the model identify even the smallest abnormalities. For increased accuracy and efficiency, we also present an optimization technique to fine-tune the network design and hyper parameters. In terms of precision, recall, and F1-score, experiments on benchmark datasets like UCSD Ped2 and ShanghaiTech show that our suggested strategy performs better than current state-of-the-art techniques.},
        keywords = {Deep Learning Encoders, Strategies for Optimizing, Attention Mechanisms, Spatiotemporal, Monitoring Systems for Feature Learning, Neural Networks.},
        month = {March},
        }