Human Re-Identification Using Gait Analysis

We live in the generation wherein innovations such as Cancer Detection, Facial Recognition and Data Analysis have become a common place, all thanks to Image processing and Machine Learning. These scientific advancements have led to extensive progress in the medical field. Technology and Medicine coming together is the actual need of the hour. Person reidentification is a fairly researched area. Till now the main focus has been on using physical features like face, fingerprint and iris. Apart from these physical features, other behaviour based features such as signature, voice and gait can also be used for person identification. People often feel that they can recognize a person from a certain distance simply by the way that person walks. This common experience combined with the recent interest in biometrics, has led to the development of gait recognition by creating a deep learning pipeline for the identification of the person based on their walking style. This paper aims to provide a thorough understanding of the subject of human gait and proposes a system for successful re identification of a person. The system works using OpenCV libraries and deep learning models like caffe.