Notice of Retraction Gait recognition using the temporal information of leg angles

摘要: Gait recognition is of increasing interest recently. A gait method using the temporal information leg angles proposed, based on fact that changes can reflect main trait a man's moving. First pendulum model referred to for extracting parts, and least-square used boundary fitting, obtain angle about thigh shin. Then, acquired expanded in Fourier series form, consideration walking activity's periodic property. Genetic algorithm subsequently adopted search harmonic coefficients, which are normalized produce signature. Finally, NN ENN classifiers both applied do sorting task. Experimental results CMU database show proposed has high rates. Furthermore, self-occlusion problem, may often be encountered recognition, well handled by this method.