Classification of Digital Intra Oral Periapical Radiographs by Selecting a Feature Vector using Hybrid Method for Selection of Features

Objectives: Study the characteristics of Intra Oral Periapical Radiographs (IOPA) and design a feature vector to automatically classify individual tooth as either healthy or non healthy using perception. Methods/Statistical Analysis: The Intra Oral Periapical Radiograph (IOPA) is segmented so that each image contains exactly one tooth. In this paper input for classification algorithm is the segmented IOPA. Feature vector is generated for each image. While determining feature vector, statistical as well as structural properties of the image are considered. Feature vector input to the classifier is multidimensional. The classification is carried out so that input IOPA is classified as an image of a healthy tooth or a non-healthy tooth. Findings: Algorithm is tested on 50 radiographic images containing healthy as well as diseased teeth. The classification algorithm presented in this research work is a two class classifier. This algorithm can be very easily adopted for multi class classification by calling the same repetitively using the classification strategy one against rest. Applications/Improvements: Selection of a feature vector and classification based on this vector is fully automatic. The algorithm incorporates hybrid method for feature selection and uses perception for classification based on feature vector. The algorithm can be improved by designing multi-class classifier for the same data.