Identifying the left ventricle optimally in cardiac mr images by comparing state-of-the-art segmentation methods

In medical diagnosis, the movement of the left ventricle (LV) could be used to estimate the volume of the left ventricle and the dyssynchrony of the heart, which can provide the basis for diagnosis of heart diseases. Identification of the LV endocardium, especially the images with poor image quality and images in apical or basal slices, is still a very challenging problem. In this paper, an automatic segmentation method based on threshold is proposed. This method works well in image both with good quality and bad quality. We tested the proposed SDD method with other 15 state-of-the-art segmentation methods by 104 frames of testing Cardiac MR images from Computing and Computer Assisted Intervention (MICCAI) 2009 challenge. Finally, we assessed the deviation between the automatically segmented and benchmark manual contours. The proposed method achieved 0.9172 average Dice metric, 1.9817 mm average perpendicular distance (APD). These results compared with other methods indicate that the proposed SDD method is an effective and viable method to identify the boundary of left ventricle