In the paper, the main objective is to develop a gait-based gender classiﬁcation in video surveillance applications that can run in real time. The Most important contribution of this work is a new fast feature extraction strategy that uses the 2D point cloud obtained from the frames in a gait cycle. For each frame, these points are aligned according to their centroid and arranged into groups. After that, they are projected into their PCA plane, obtaining a representation of the cycle particularly robust against view changes. By analysing the discriminatory capability of different body components, it is observed that hair, back, chest and thigh components are more distinct than other components. Then, ﬁnal discriminative features are computed and KNN tool is used as a classiﬁer. Experiments over auditorium datasets, CASIA-B database and TUM-IITKGP dataset are veriﬁed and the proposed method can essential characters determine the target, efﬁciently label moving targets, and classify the gender.
Digital Object Identifier (DOI)
KalaiSelvan, C. and Sivanantha Raja, A.
"Robust Gait-Based Gender Classiﬁcation for Video Surveillance Applications,"
Applied Mathematics & Information Sciences: Vol. 11:
4, Article 28.
Available at: https://digitalcommons.aaru.edu.jo/amis/vol11/iss4/28