Gait recognition using sub-vector quantisation technique

Recognising people from their gait is a challenging problem in biometric research. In this paper, we address the problem of gait identification based on a novel approach of sub-vector quantisation (SVQ) technique. A silhouette-based algorithm is utilised to capture the spatial-temporal information of the gait. A sequence of temporally ordered outer contour widths of binarised silhouettes of a walking person represents the feature vectors set. The feature vectors are segmented into sub vectors and vector quantised independently to represent the gait signatures using low dimensional vectors. Dynamic time warping (DTW) technique is used for gait feature sequence matching. The proposed method is validated on several well known benchmarked databases as well as on our own database. The experimental results confirm the validity and robustness of the proposed SVQ method for gait recognition.