Identity Recognition of Multi-Haptic Pressure Feature based on Sparse Representation

Abstract

Herein, a new identity recognition method of multi-haptic pressure feature based on sparse representation was investigated. According to the common dynamic features, the regional feature and the ratio of length vs. width of external bounding rectangle (extracted by using the least area method) were extracted. The subset of dynamic feature was optimized by correlation criterion, the sparse representation of haptic pressure was obtained according to the sparse basis (i.e., wavelet basis), and the sparse feature vector was calculated by the Topelitz measurement matrix. After that, the haptic pressure feature set was created by combining dynamic feature subset and sparse feature subset linearly. Furthermore, Support Vector Machine (SVM) classifier identified more than two objects following the one to many rule and output the identification result according to the rule of majority voting, and the stability of features is studied by calculating the intraclass correlation coefficient (ICC) and coefficient of variation (C.V). Overall, the improved acuracy of identity recognition demonstrating the effectiveness and stability of the multihaptic pressure feature.