Context-Aware Detection Networks for Enhanced Women Safety and Threat Preventionn

Context-Aware Detection Networks for Enhanced Women Safety and Threat Prevention

Authors- Associate Professor Sarvagya Jain, Vinit Dubey, Nishesh Patil

Abstract-In order to increase the frequency of security threats to women in urban areas, advanced analysis is needed for real-time monitoring and intervening. In the current observation system, specific threats can often be detected, and context information can be analyzed in real time. In this paper, we propose a robust detection algorithm specifically designed for women’s safety that significantly improves detection accuracy and response time by using a combination of scene-based context and deep learning techniques. Our approach includes: (1) highly accurate gender-classified person detection for real-time analysis of gender distribution, (2) recognition of anomalous scenarios such as a woman alone at night or surrounded by a group of women, and (3) gesture-based SOS detection. The integration of contextual data with convolutional neural networks (CNNs) facilitates early alerts and trend analysis to support proactive interventions. By identifying hotspots and providing ongoing updates, our model enhances both public safety and resource allocation for law enforcement. A prototype of the complete system is available to the community.

DOI: /10.61463/ijset.vol.12.issue6.919