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{200782,
        author = {Divyam Jangada and Yashraj Nikam and Aditya Lagad and Palak Chawarkar},
        title = {Multimodal AI for Crowd Risk Detection and Prediction},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {12},
        pages = {5826-5832},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=200782},
        abstract = {The management of mass gatherings in rapidly expanding urban environments presents a critical public safety challenge. Traditional crowd monitoring systems are largely reactive and often fail to capture the complex visual, behavioral, and acoustic precursors of crowd-related incidents. This paper proposes an integrated multimodal deep learning framework for proactive crowd risk detection and prediction. The system employs YOLOv8 for real-time person detection and CResNet for enhanced feature representation and crowd density characteri- zation, ensuring robustness in dense and occluded environments. The proposed framework integrates visual density estimation, skeletal pose-based gesture analysis, and acoustic panic detection. By replacing conventional rule-based mathematical constraints with deep feature extraction, the system utilizes an anchor- free detection mechanism and a Long Short-Term Memory (LSTM) network for temporal risk forecasting. Experimental results demonstrate a peak accuracy of 95.4.},
        keywords = {Crowd Management, YOLOv8, Behavioral Analysis, Multimodal Fusion, Predictive Analytics, Smart Cities, Deep Learning.},
        month = {May},
        }