Segmentation of Bony Tissues from CBCT Images

摘要

A segmentation algorithm of the jawbones and spinal cords is developed from cone beam computed tomographic (CBCT) data sets. Our segmentation algorithm was tested on five CBCT data sets and compared with manual segmentation results. The jawbones are bony tissues consisting of the upper jaw and the lower mandible. The procedure is based on the intensity distribution and space growing to separate the jawbones and cervical spine from their neighboring tissues. The segmentation procedure is to grow and preserve the shapes of these bony tissues. The developed segmentation algorithm separated the jawbones and cervical spine successfully. The results were quite accurate comparable with those by hand. Further, the method was robust and efficient.