A Correlative Study of Contrary Image Segmentation Methods Appending Dental Panoramic X-ray Images to Detect Jawbone Disorders

Dental radiographs have been widely used by dentists to detect any bony pathology which is difficult to diagnose solely by clinical examination. The usage of dental X-rays images has brought about a great improvement in clinical diagnosis due to its immediate availability and relatively lesser radiation dose. Orthopantomograms (OPG) or panoramic imaging is one of the imaging modality frequently used in dentistry to detect any dental anomaly. But the dental panoramic images suffer from varying superimposition of lot of anatomical structures and have an inherent technical issue which leads to lot of ambiguity when applied as an aid to diagnosis. This paper presents the systematic review of image segmentation algorithms applied on dental X-ray images and its results with the supervision of radiologists. A generic comparison of segmentation algorithms has been discussed for the cysts and lesion segmentation in distinction to panoramic image. Thresholding watershed and level sets methods were chosen for segmenting the desired region of cysts and to study the various characteristics of cystic region. Level sets segmentation produces the better results in segmenting the cyst/tumors. The shape descriptors obtained for the region of cysts could conceivably be used as feature vectors in image classification where the classifiers can automatically detect the abnormal tissues or tumors from OPG images helping in its diagnosis and treatment.