Extraction of the Brain from CT Head Scans Based on Domain Knowledge
Guoyu Qian, Suhuai Luo, Jesse J. Jin, Mi‐Ra Park, Wieslaw Lucjan Nowinski, Tuan D. Pham, Xiaobo Zhou · AIP conference proceedings · 2007
We present an automatic approach for an efficient brain extraction from CT head scans. Regions of interest are first set in each slice by applying thresholding and region growing. Next, the brain candidates are extracted by using three‐dimensional region growing with a variable, anatomy‐dependent structuring element. Domain knowledge, including Hounsfield unit ranges, anatomy, and image acquisition parameters, is applied. The proposed method has been applied automatically to 27 CT normal and pathological scans and has shown promising results. The average sensitivity, specificity and Dice's index for 5 cases are 99.6%, 99.4% and 98.7%, respectively.