The Study Of The Gait Recognition Basis On Human Body Jiont Proportion Model

Objective: This paper establish a gait recognition method basedon the human body joint model. The dynamic characteristic curve isformed by marking the joint points of the body in the image，then weuse the coincidence method of the dynamic characteristic curve torecognite the gait.Methods: Method and the main innovation of this study are asfollows: Firstly, to establish the China’s population joint scale models.Totally331people (205male and126female) were enrolled, the bodyheight and the distance from the body joints points were measured byindirect measurement method. To establish21joint proportion index.The characteristics of joint proportion of chinese is analysised bycomparising joints proportions between men and women, chinesepeople and foreigners, and comparising individual joint proportions ofpeople who are at the same height.Secondly, the feature extraction.The video images cutting to25frame/SEC, we marking the jointspoints of human body from sequent images directly. The x axis is thehorizontal displacement of the body joints points, the y axis is the vertical displacement, the z axis is the time change.Then the dynamiccharacteristic curve of the body joints points is formed by theaxis.Thirdly, the classification of the gait recognition. Based on thecharacteristics of the proportion of the human body joint, we use thecoincidence method to contrast.Two pieces of gait graph are coincidedto analysis, from three aspects to compared, the difference of the bodyjoints proportion, the difference of the curve shape of the gait cycle,and the similarity of the gait curve.Results: We realized that the human body joints proportion indexis the only through the statistical analysis, the people who have thesame height but the have the different joint proportion index.,the jointproportion index of the men and women has the obviousdifference(p<0.05), the body joints proportions are different betweenchinese people and foreigners. The gait recognition basis on the thedynamic characteristic curve by coincidence contrast has highrecognition rate and robustness.Conclusion: The method of the dynamic characteristic curvecoincidence basis on the human body joint model has high recognitionrate and robustness.