Study of color matching system for porcelain teeth

Objective: To design and build up a computer-aided color matching system for porcelain tooth, which could offer dentists an objective suggestion for selecting the color of the patient´s tooth. Methods: The color matching system is mainly based on image analyzing and processing techniques and pattern recognition methods. This system uses HSI color space to compare and calculate colors. The system consists of five parts, which are image acquisition facility, image processing part, tooth color classification module, color mixture module and a template library. This system uses image analyzing technology to find out tooth image´s color feature and then look it up in the template library to get the standard term of this tooth´s color using pattern recognition methods. Results and discussions: Firstly, we use the minimum distance classifier and do two experiments with three tooth models of the identifier 0M1, 3M2 and 5M1. In the first experiment, we take 20 pictures as samples for each tooth model, setting the first 5 pictures as training samples and the last 5 pictures as testing samples. The recognition rate is 66.7%. In the second experiment, we use the former 10 pictures of each tooth model as training samples and the other 10 pictures as testing samples. The recognition is 90%. In addition, we use KNN classifier to test the above test datasets, but the recognition rates are obviously lower than those obtained by the minimum distance classifier. Conclusions: The tools and processing platform used in this experiment are simple and efficient. The recognition rate of this color matching system is good and accepted.