Comparison Between Different Methods of Segmentation in Dental Image Processing

Abstract

This paper presents three different methods of image segmentation and analysis of various panoramic dental radiographies. The images are analyzed from different perspectives depending on the dental condition of each patient’s set of images, so the necessary information for establishing a correct diagnostic are extracted and analyzed. The algorithms are developed using Matlab work environment, so that if an image is hard to interpret, firstly is pre-processed using the Image Segmentation toolbox. Image segmentation methods are applied for analyzing the intra oral configuration of each patient based on the contour, size and intensity of specific surfaces identified by the dental specialist as being problematic. Each method will be detailed in this article and a comparison will be made between the three segmentation methods. Furthermore, another aim of this paper is find the best solution for processing this type of images, which are considered very hard to interpret using an automated tool and also to find even the hidden issues from a panoramic dental radiography. Some diagnostics can be identified from the first look of a specialist, but some images are very hard to interpret, so an automated tool is helpful in this cases. As a conclusion, the solution of processing such an image is developed, tested and finally, it will aid in the analysis of panoramic dental radiography.