Automatic segmentation of jaw from panoramic dental X-ray images using GVF snakes

Panoramic radiographs or dental X-rays are still a prominent tool used by the dentists for dental diagnosis. Accurate segmentation and reconstruction of each tooth from panoramic radiographs can play very significant role in early diagnosis and treatment that helps physicians to make a better decision. Panoramic radiographs contain regions outside of the jaw, which make segmentation of teeth very time consuming and complex. Automatic segmentation of jaw is necessary to avoid such complexity and to reduce time. Due to the absence of substantial difference of pixel intensity between different structures and absence of strong edges in panoramic radiographs, segmentation of jaw is still perplexing. In this paper, we use the gradient information to segment the jaw regardless of the position and size of the jaw. Our algorithm for segmentation of jaw has four steps: k-means clustering, detection of points around the jaw (proposed), gradient vector flow (GVF) snakes and correction by making the shape of the segmented area oval. K-means clustering followed by thresholding and detection of points around the jaw (proposed) is used to find suitable points for initialization of the GVF snakes. Then, GVF snakes are initialized with the detected points to segment the jaw. Finally, correction is performed by making the shape of the segmented area oval.