Person Recognition using Gait Energy Image, MobileNetV3Small and Machine Learning

In the past few decades, gait as a biometric attribute has garnered a lot of attention for its application in a broad variety of security and privacy-related scenarios, including identity recognition and authentication. The deployment of surveillance cameras allows for the collection of Gait data, which is a non-invasive form of identification that is based on a biometric trait that can be seen from a distance and can be utilized for identifying reasons. It is difficult to create reliable automatic gait identification systems since Gait can be modified by several covariate parameters. Some of these covariate elements include garments, walking tempo, camera view angle. In this work, a method for recognizing people with a high degree of accuracy is demonstrated. This paper makes use of the Gait Energy Image, MobileNetV3Sma11 model, and Machine Learning to develop an approach to recognize person using Gait pattern. The paper includes experimental findings, including precision, recall, f1-score, and classification accuracy, all of which point to the effectiveness of the suggested method.