Panoramic Dental X-Ray Image Segmentation and Feature Extraction

Pedro H. M. LiraGilson A. GiraldiLuiz A. P. Neves

Panoramic X-Ray images are very popular as a first toolfor diagnosis in odontological protocols. Automating theprocess of analysis of such images is important in order tohelp dentist procedures. In this process, teeth segmentationfrom the radiographic images and feature extraction are essentialsteps. In this paper, we propose a segmentation approachbased on mathematical morphology, quadtree decompositionfor mask generation, thresholding, snake models.The feature extraction stage is steered by a shape modelbased on Principal Component Analysis (PCA). Firstly, wetake quadtree decomposition after applying low pass filterto smooth the original image and select the smallest blocksto generate a mask. Then, the original image is processedby Otsu's thresholding. The result is improved by morphologicaloperators and the quadtree mask is applied to addressoverlapping, a common problem in x-ray images. Theobtained regions are searched and the larger ones are selectedto find out the teeth candidates. The boundary of theobtained regions are extracted and aligned with the shapemodel in order to recognize the teeth. The selected curvesare used in a search method to initialize a snake technique.Finally, morphometrics data extraction is performed to obtainthe teeth measurements for dentist diagnosis. Preliminaryexperiments show the significance of the proposedmethod to extract teeth from an image.

Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato:
     Mantida por: