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{188195,
        author = {Yashaswini N B and Hanumanthappa S and Ullas M P and Gunashree H K and Vismaya K R and Rajashekar K J},
        title = {Abnormal Behaviour Detection in Massive Crowd},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {12},
        number = {7},
        pages = {1152-1158},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=188195},
        abstract = {The requirement for intelligent surveillance systems that can identify anomalous activity in intricate, high-density environments is highlighted by research on crowd anomaly detection. According to previous research, detection reliability is greatly increased when spatial characteristics, motion analysis, and temporal modelling are combined using techniques including optical flow, CNN-based feature extraction, trajectory analysis, and transformer frameworks. These studies demonstrate the superiority of deep learning over conventional feature-engineered methods, particularly hybrid spatiotemporal architectures. In order to identify anomalous activity in large crowds, this work proposes a hybrid approach that combines DenseNet-201 for spatial representation, optical flow for motion interpretation, and Bi-Directional LSTM for temporal behavior modeling. The model performs best with a batch size of 32 and a learning rate of 0.0003, achieving 55% accuracy and an AUC of 0.7187 when evaluated using the UCF-Crime dataset. The outcomes show how successful the technology is in monitoring crowds in the actual world.},
        keywords = {AUC, autoencoder, crowd, CNN, Dense Net, nonparametric test, transformer, VGGNet, and video  anomaly detection},
        month = {December},
        }