Fuzzy Clustering of Female Body Shape and Identification of Side Parts Characteristics

摘要

The development and application of virtual human body reconstruction required distinguish human bodily form accurately. At present, most of the analysis of the human body was based on body cross-section and circumference characteristics, which can not well reflect the human body curves. This paper analyzed the characteristics of the human body side, using two indexes reflect female body side parts characteristics. According to statistical analysis, the two indexes relative membership functions were established. This paper classified 20 samples by the fuzzy classification and recognition model, and obtained the optimal fuzzy recognition matrix, the optimal fuzzy clustering central matrix and the reasonable weight vector. Then a model of the fuzzy clustering and recognition of female bodily form were built based on lateral part characteristic. Moreover, 25 female body data was analyzed with this model. Calculation results showed that using fuzzy partition clustering method to female side parts form characteristics of classification and recognition of the method was feasible.