Detection of dental Fluorosis using enhanced K-means and Fuzzy C means

Abstract

Detection and prognosis of Dental fluorosis is foremost for the doctors, the researchers as well as for the sufferers. Dental fluorosis is the presence of vague white streaks or patches on the tooth area that only happen when there is utilisation of fluoride from any source, over extended periods when teeth are getting mature. The existing research works and techniques mainly depends on the visual ability of the dentist to detect or to separate the areas of fluorosis. So to overcome the issue, this paper proposed a solution for getting the severity level of dental fluorosis without human intervention using enhanced K-Means unsupervised clustering algorithm and fuzzy C-means on HSV (hue, saturation and value) colour model. Using the clustering algorithm helped us out to successfully distinguish the fluorosis affected area. The area percentage was calculated and accordingly images were categorized with different level of severity level. The level was compared with that given by expert on each image. The results obtained are far better than the conventional approach [1].