Paper Title
A Comparative Study of Gait Analysis for Human Biometric Estimation

Abstract
For identity verification, biometric data—such as fingerprint, iris, face, and speech recognition—is frequently used in security authentication systems. The non-intrusive collection of human gait is an advantage over other biometric data. From a kinematic perspective, one can see that every individual has a distinct and odd gait. The concept behind gait recognition is that subtle differences in a person’s gait can be utilized as a biometric identifier to identify them. In order to detect human gaits, the research provides a novel method that involves extracting gaits from modeled body parts and applying them to human detection. The objective is to use the developed approach to detect distinct walking people with maximum accuracy. The proposed work entails recording the subject’s video, obtaining its skeletal data, and calculating the subject’s likelihood of identification. Combined, the Siamese method, CNN classifiers and LSTM classifiers enable some of the highest detection accuracy for walking people. Keywords - Human Gait, Time normalisation technique, Sil- houettes, static and dynamic gait features, Principal Component Analysis(PCA), Histogram of Oriented gradients(HOG), Gait cycles(GCS), CNN and LSTM classifier, Siamese Technology.