Texture based classification of dental cysts

Dental cysts are usually caused due to root infection involving the tooth affected greatly by carious decay. The most common forms of dental cysts include Ameloblastoma, Keratocyst and Dentigerous cyst. The treatment is planned based on the type of cyst. Differentiating Odontogenic Keratocysts and Ameloblastomas from other cystic lesions in the maxillomandibular region is crucial because of their high recurrence rates and to avoid disease progression. Furthermore, Conventional radiography and CT (Computed Tomography) images are inadequate for differential diagnosis. The proposed work carried out using dental radiographs aids the dentist in treatment planning by facilitating the process of categorizing the dental cyst using image processing techniques based on texture information. Initially, the dental radiographs are enhanced using contrast enhancement. Then, Classification is performed using texture parameter estimation based on GLCM approach. The estimated texture features include Contrast, Correlation, Energy, Homogeneity and Mean. These parameters are used for classifying the dental cyst using K-means classifier in feature space. Finally, the accuracy of the proposed work is assessed and the results are presented.