Paper Title
MORe-Unet: Multi-Outputs Recurrent Unet for Medical Image Segmentation

Medical image segmentation has been an interesting topic in recent years. Deep learning is showing a superiority to traditional methods. The disadvantage of deep learning networks used standard convolution neural network is that the size of the model is often large, resulting in a lot of memory consumption and large computation time. In this paper, we propose a new model named MORe-Unet, which combines the use of Recurrent convolutional neural network (RCNN) based on Unet architecture to reduce the network size with exploitation of all output feature maps in convolution units of the network nodes. We evaluated the effectiveness of the models on liver segmentation application on LiTS 2017 dataset and Spleen segmentation on Medical Segmentation Decathlon Challenge 2018 datasets. Experimental results show that the proposed model has smaller size and improved performance compared to current models. Keywords - Recurrent Convolutional Neural Network, Medical Image Segmentation, Liver Segmentation, Spleen Segmentation, Unet Architecture.