Implementation of fuzzy similarity methods for manipulative hand posture evaluation

In this paper, an implementation of fuzzy similarity methods for model based manipulative hand postures is presented. The model used to represent manipulative hand postures is based on a human hand kinematical model. In the model, the posture is formed as a set of hand joint variables. Joint angles of the posture are fuzzified to obtain a Fuzzy superset, which is used to obtain the symbolic posture pattern. Evaluation of different types of fuzzy membership functions is performed for fuzzification. Fuzzy similarity measures are adapted and analyzed as a distinction tool for a possible hand posture identification and classification.