VISION BASED GAIT RECOGNITION IN FORENSICS

GAIT recognition technology is rapidly developing biometric tech- nique that identifies individuals based solely on the walking patterns, establishing it as a useful resource for forensic applications. Unlike traditional biometrics, GAIT recognition offers a non intrusive, long range, and contactless means of identification, that is particularly useful in forensic scenarios that other bio- metric identifiers may not be available. An in depth analysis of vision based GAIT recognition techniques are provided, focusing exclusively on GAIT analysis without incorporating additional biometric modalities. Various approaches, including silhouette based, model based, and Deep Learning driven methods, are explored, highlighting the strengths and limitations in forensic investigations. It discusses key challenges such as GAIT variabil- ity due to environmental conditions, clothing, and occlusions, as well as the impact of camera view points on recognition performance. Recent advancements in feature extraction, cross view recognition, and temporal GAIT analysis are also reviewed to improve robustness and accuracy in forensic settings. Future directions emphasize the need for forensic specific GAIT datasets, real time implementation, and improved adaptability to uncon- trolled environments. It aims to provide forensic researchers with insights into the cutting edge of GAIT recognition and its potential for real world forensic applications.