An adaptive algorithm for cardiac MRI image segmentation

In the modern medical image processing community, the medical image segmentation has been a challenging and vital problem. Accurate diagnosis of the diseases and other follow-up treatments rely much on the accuracy and effectiveness of segmentation algorithm. Based on the investigation on human heart MRI image, an adaptive heart segmentation algorithm with improved Mean Shift transformation has been proposed. First, the image contrast was enhanced by using gray stretch transformation, then the image was segmented based on improved Mean Shift transformation. The experimental results show that the proposed algorithm can segment the heart image accurately and it is also robust to noise.