A New Approach For Human Identification using AI

Human Identification is widely used to enhance surveillance systems. CCTV (closed circuit television) system based low-resolution cameras are used almost everywhere but these cameras don’t capture in very good qualities/resolution, thus facial features or other biometrics are impossible to extract. There is a requirement of a technology which implements low-resolution camera feeds to identify humans at a distance for a more effective identification system. Gait is an individual’s pattern of walking, which is unique to every individual. It can be further understood as the pattern of movements of limb made during walking by an individual. When comparing to first-generation biometric modalities, gait has an upper edge as it is unobtrusive, i.e., it doesn’t require contact with the subject, hence making it an ideal biometric modality for human identification at a distance. An artificially intelligent model is proposed which can be made to apply motion tracking and pattern recognition in order to analyze the walking pattern and predict the possibility of whether the individual walking in the feed and the individual whose gait is being analyzed are the same or not.