Gait features fusion for efficient automatic age classification

Abstract

Far from the camera, image resolution is significantly degraded and person cannot cooperate with the acquisition equipment. So, the classical intrusive biometrics approach could not be applied. As a non‐intrusive biometric, gait analysis gained the attention of the computer vision community for number of potential applications such as age estimation. Since, that gait is very sensitive to ageing, gait analysis is the suitable solution for age estimation at a great distance from the camera. Given the complexity of this task, the authors propose in this study a new approach based on descriptors cascade. The proposed approach is to use a fusion of some efficient contour and silhouette descriptors. Indeed, they introduce the proposed descriptor based on silhouette projection model (SM) in the first time. In the second time, the proposed descriptor is merged with the best existing ones in order to enhance the classification performances. Despite that age classification using gait is a very challenging task, experiments conducted on OU‐ISIR database show that their proposed descriptors fusion approach enhances considerably the recognition rate.

Keywords

• gait features fusion
• automatic age classification
• nonintrusive biometric
• gait analysis
• computer vision community
• age estimation