Kernel fuzzy clustering based classification of Ancient-Ceramic fragments

In this paper, we propose a method for automatic classifying fragments in computer-aided restoration of ceramic cultural relics, using surface texture clustering. Firstly, center of clustering will be initialized by re-division on multi-variant finite model. Then kernel-based fuzzy clustering algorithm is applied and parameters of difference degree are used to control iterations of clustering. The experiment result shows that this scheme is characterized by better automatic classifying rate for Ancient-Ceramic fragments.