PERSON RE IDENTIFICATION FOR PUBLIC SAFETY IN INDIAN RAILWAYS USING DEEP LEARNING
DOI:
https://doi.org/10.5281/zenodo.19510391Keywords:
Person Re-Identification, Deep Learning, Indian Railways, Computer Vision, CNN, Siamese Networks, Surveillance Systems, Public Safety, Feature Extraction, Metric LearningAbstract
The proposed system titled “Person Re-Identification for Public Safety in Indian Railways Using Deep Learning” focuses on enhancing surveillance and security across railway stations and trains by leveraging advanced Deep Learning and Computer Vision techniques. With millions of passengers traveling daily, Indian Railways faces significant challenges in monitoring suspicious activities, tracking missing persons, and identifying individuals across multiple camera views. Traditional surveillance systems rely heavily on manual monitoring, which is inefficient, error-prone, and unable to handle large-scale real-time data. To address these limitations, this system introduces an automated person re-identification (Re-ID) framework capable of recognizing and tracking individuals across different cameras and time frames. The proposed methodology utilizes deep learning models such as Convolutional Neural Networks (CNNs) and Siamese Networks to extract unique feature embeddings from human images, including attributes like clothing, body shape, and gait patterns. These features are then compared across multiple video streams to match and re-identify individuals. The system incorporates metric learning techniques to improve matching accuracy and robustness under varying conditions such as lighting changes, occlusions, and different camera angles. Additionally, the platform integrates real-time video processing using surveillance feeds from railway stations, enabling continuous monitoring and rapid identification of persons of interest. Experimental results demonstrate that the proposed system achieves high accuracy in person re-identification tasks, significantly improving tracking efficiency compared to conventional methods. The system can be used for applications such as identifying missing persons, tracking suspects, and enhancing crowd management. Furthermore, the integration of scalable infrastructure ensures that the system can handle large volumes of data across multiple stations. In conclusion, the proposed deep learning-based person re-identification system offers a powerful and scalable solution for improving public safety in Indian Railways, enabling faster response times and more effective surveillance operations.
