Computer aided classification using support vector machines in detecting cysts of jaws

Abstract

Jaw cyst is one of the most common pathology observed in the field of dentistry. Early detection of the cystic lesion will help the surgeons to take appropriate therapeutic measures after a thorough diagnostic procedure. One of the challenging task for surgeons is to differentiate the cysts from the other pathologies. The appearance of these pathologies on a radiograph is a complex and confusing task due to the close similarity between the cysts and tumors which is a difficult to differentiate just by its appearance. Hence to resolve this problem, a computer aided classification algorithm is needed for accurate classification of cysts. The work presents a new approach for the determination of the presence or severity of the jaw bone disease aiding the diagnosis and radiotherapy planning. This paper presents texture characterization for the dental panoramic images. The transposed images are analyzed using Gray level co-occurrence matrix(GLCM). The textural properties such as entropy, contrast, correlation, energy and homogeneity are determined for both cyst and non-cystic images. The results obtained are fed to the classification model to classify the given image into normal or abnormal images containing cyst. Support vector machines are chosen for image classification. Image dataset of 30 were used in training and validation. The image set consists of 20 abnormal images and 10 normal images used in image classification.