Diagnosing Interdental Decays in Mouth Radiography Images using Kernel Fuzzy C means Segmentation and Cascade Object Detector

Mouth radiography is one of the common ways of diagnosing tooth decays. Especially for interdental decays which are hard to be examined by naked eyes. In this paper, we present a method for diagnosing internal decay in real word mouth radiography images which have been gathered in Tabirz Sina dental clinic. Firstly, we will use Kernel Fuzzy C-Means (KFCM) algorithm, which is modifying the objective function in the fuzzy C-means algorithm using a kernel-induced distance metric, as an image segmentation method. Then, the processed images are labelled with decay and are then employed to a cascade object detector for diagnosing purposes. In order to show the efficiency of the employed method the performance is tested on testing mouth radiography image data set. The results indicate that this method composed of KFCM and cascade object detector structures is successful in detecting interdental decays.