Feature Extraction and Ranking for End To End Person Re Identification System Using Deep Convolutional Neural Network

Abstract

Surveillance cameras are installed in homes and public places like hospitals, malls, railway stations, airport etc. to ensure safety and security of the public and their belongings. The surveillance footage helps the professionals to investigate and identify the criminals in nearly seven out of ten murders in the last few decades. Matching a human image across humans captured in different cameras at different places were used to find the person of suspect in many criminal cases. The process of finding matches for a person of suspect across multiple cameras is known as Person Re-IDentification. newlineThe Person Re-IDentification system (PRID) helps to keep track of human entries filmed by various cameras at various locations. The PRID system retrieves relevant images of the same person acquired by dissimilar cameras at various locations and at different times as a response to an input image of the person captured by a camera (called a probe image). It automates the process of locating a person or target in one camera view who has been tracked and lost in another camera view by comparing the photos of the same individual newline