CT Image Segmentation Based on Order Morphological Gradient Map and Improved Region Growing Arithmetic

The segmentation of pathological interesting region is the key point in medical image segmentation, because the interesting regions include important diagnostic information. This paper proposed a new segmentation method that is based on the order morphological gradient map and the improved region growing arithmetic. The order morphological gradient map can be obtained by the order morphological transformation. Using improved region growing method, the gradient map can be segmented and the interesting region can be extracted by choosing suitable threshold value. The threshold is selected according to the method of floating region value. In order to verify the effective of this method, the paper made the simulation and analyzed the result. The results show that the edge of the interesting region using this method is exact and clear, and to noised image, this method also can work effectively and has good robust property.