IEEE 2016 International Conference on Wireless Communications, Signal Processing and Networking (WiSPNET 2016)

246 / 2016-01-27 18:30:05

Automated diagnosis of various dental cysts using soft computing techniques

Dental cysts, GLCM, GLRLM,.SRE

A cyst is a pathological epithelial lined cavity that is filled with fluid or soft material and usually grows from internal pressure generated by fluid cavity from osmosis. The cyst grows from hydraulic pressure which causes the bone around it to be resorbed. The commonly occurring dental cysts include Dentigerous, Kerato, Radicular, Buccal bifurcation, Calcifying and Nasopalentine cyst. The most common treatment for cyst is removal of the cyst region. Classifying the various types of cystic lesions in the maxillomandibular region is essential due to their high recurrence rates. Conventional radiographies such as x-rays and CT scans are limited for differential diagnosis. This paper focuses on automatic classification of dental cysts to enable the diagnostic ease of the dentist for designing an appropriate treatment procedure. Initially, the dental x-ray image is preprocessed using histogram equalization technique to improve its visual quality. Segmentation of the cyst region is performed by template matching and artificial neural network (ANN). Then features such as contrast and correlation are extracted from GLCM matrix of the segmented cyst regions and achieved an accuracy of 72.4% and 83.3% in the classification stage. Similarly, short run emphasis (SRE) value is extracted from GLRLM matrix and obtained 86.6% in the classification stage.