Gait Based Gender Identification Approach

To improve the performance of gait based human identification system, gender can plays an important role in the field of surveillance and monitoring applications. The proposed algorithm consist of four steps. In initial step, silhouette object detection is take place by using background subtraction and morphological operation. In segmentation step, silhouette body is divided into six regions. Then their gait features are extracted by using 2D discrete wavelet transform and finally the K-Nearest Neighbor (KNN) classifier is employed to classify the gender for identification of the person. To evaluate the performance of the proposed algorithm, experiments are conducted on CASIA Gait database. An experimental result shows that the proposed method is more effective for gender identification using gait biometrics. The proposed approach achieved highly competitive performance compare with earlier published methods.