A novel method of CT brain images segmentation

In this paper, an algorithm is developed for anatomical structures segmentation based on CT head images. The segmented structure can be used for image guide surgery navigations. In our method, intensity rescaling, the threshold algorithm, region growing method, fast-marching method (FMM) and mathematical morphology are combined and used systematically. Due to the low contrast of the CT images, intensity rescaling is applied to the images to enhance the contrast. Threshold algorithm is used to extract the intracranial area from the enhanced images. Then, FMM is adopted to segment the intracranial images of brain matter and cerebrospinal fluid (CSF). Mathematical morphology is used to correct the pre-calculated images and obtain accurate brain matter and CSF segmentations. The system has been tested with a number of real CT head images and has achieved some promising results.