Balance Error Scoring System Based On Duplex Kinects

The aim of the study is to develop and validate a contactless Balance Error Scoring System(BESS)using duplex Kinects which provide an automated sideline concussion administration method to solve the problems of high cost,complex operation and subjective factors.In this paper,we adopted a system with two Kinects to overcome the problem of selfocclusion,view and artificial vibration caused by single Kinect.In order to analyze the effect of placement angle and distance on the effective working range,we established a model based on the placements of the two Kinects.Then,according to the transformation principle of the spatial coordinate system,the skeleton data collected by duplex Kinects were unified into the same coordinate system.We proposed a skeletal optimization algorithm based on bone length to fuse the two skeleton data collected from duplex-view.Finally,we carried on the feature analysis,extraction and classification to the pose according to the BESS protocol and realized the scoring function by the correction algorithm of repeated counting.At last,the Balance Error Scoring System using duplex Kinects was validated by three complete tests of five subjects.The linear regression analysis method was used to compare the artificial observation data and system measurement data.And the correlation coefficient of the duplex Kinects is 0.793,which is 2.08 times of the single Kinect correlation coefficient of 0.38.