Automatic MRI brain tissue extraction algorithm based on three-dimensional gray-scale transformation model

In this paper, an automatic MRI brain tissue extraction method is proposed. Firstly, the thresholds of the brain skull, brain tissue, and cerebrospinal fluid are computed by the adaptive threshold algorithm. Then the thresholds are adopted to estimate the center of gravity (COG) and radius of the brain tissue in MRI images. Thirdly, a grayscale transformation model is proposed to enhance the contrast of the boundary between the brain and non-brain tissue. Finally, a pre-judgment 3D region growing algorithm is used to extract the brain tissue, which effectively reduces the number of segmentation errors. The results show that the proposed method is more accurate and robust than BET (Brain extraction Tool).