Person Re-Identification for Public Safety in Indian Railway Using Deep Learning

Abstract

Person re-identification (Re-ID) is a critical task in intelligent surveillance systems, aiming to match and track individuals across multiple non-overlapping camera views. In the context of Indian Railways, which serves millions of passengers daily, ensuring public safety is a paramount challenge due to the vast, crowded, and dynamic nature of railway stations. This paper proposes a deep learning-based person re-identification system tailored for Indian railway environments to enhance public safety and security monitoring. The proposed framework leverages convolutional neural networks (CNNs) with attention mechanisms to extract robust discriminative features from CCTV footage, enabling accurate identification of persons across different camera viewpoints, occlusions, and varying illumination conditions. The system integrates a multi-branch network architecture combining global and local feature representations to address the challenges of appearance changes and viewpoint variations typical in crowded railway stations. Experimental results demonstrate that the proposed model achieves a Rank-1 accuracy of 91.3% and mAP of 78.6% on a custom Indian railway surveillance dataset.