Fuzzy-set based feature extraction for objects of various shapes and appearances

A method to extract functional primitives using fuzzy relations so as to effectively describe objects of various shapes and appearances is proposed. Object surfaces in a range image are segmented into visual surface regions. The geometric features between these regions are represented as fuzzy functions, which are used to extract functional primitives. This method can efficiently deal with such an object that has various shapes or appearances, and increases the performance of image analysis.