A Teeth Identification Method Based on Fuzzy Recognition

This paper presents a new method based on template recognition and fuzzy recognization. HDM (Hybrid Differential Method) is used to locate the teeth area from a face image, and fuzzy recognition is used to decide if a new teeth image belongs to the face image. 5-6 times of training is done to build up teeth feature set. A relative simple tooth feature-the lower part width of central incisor and their relative ratio are adopted as the indicators for locating a baseline. Totally 4 teeth are considered here as the identification targets. The intersections of these 4 incisors with the baseline are calculated and used as the major features of a person´s teeth. Experimental membership function and weight factor are used teeth image is preprocessed to get its configuration and to facilitate the Matlab auto calculation. Three groups of experiment samples are presented here and results are satisfied.