Inferior maxillary bone tissue classification in 3D CT images

Abstract

This paper presents a method for segmenting the inferior maxillary bone in CT images and a technique to automatically classify bone tissue without requiring a training stage. These methods are used to measure the mean density of seven main anatomical zones of the mandible, making the difference between cortical and cancellous bone. The results lead to determine the normal density values in each region of the inferior maxillary bone and help to evaluate the success of the bone regeneration process. The proposed method was validated on ten axial slices from different zones of a patient mandible, by comparing automatic classification results with those obtained by expert manual classification. A 4% mean difference was found between percentages of bone tissue types, and the mean difference between mean density values was of 88 HU. Once the method was validated, it was applied to measure density in the seven anatomical zones of the inferior maxillary bone. © 2010 Springer-Verlag Berlin Heidelberg.