In this paper, we propose a novel medical image segmentation using iterativedeep learning framework. We have combined an iterative learning approach and anencoder-decoder network to improve segmentation results, which enables toprecisely localize the regions of interest (ROIs) including complex shapes ordetailed textures of medical images in an iterative manner. The proposediterative deep convolutional encoder-decoder network consists of two mainpaths: convolutional encoder path and convolutional decoder path with iterativelearning. Experimental results show that the proposed iterative deep learningframework is able to yield excellent medical image segmentation performancesfor various medical images. The effectiveness of the proposed method has beenproved by comparing with other state-of-the-art medical image segmentationmethods.
展开▼
