Recurrent Residual Convolutional Neural Network U-Net (R2U-Net) based Medical Image Segmentation

Abstract

Deep learning (DL)-based semantic segmentation methods have demonstrated impressive performance over the past few years. In particular, these methods have been successfully applied to medical image classification, segmentation, and detection problems. One deep learning technique, U-Net, has become one of the most popular for these applications. In this paper, we propose a Recurrent Residual Convolutional Neural Network (RRCNN) based on U-Net models, which are named RU-Net and R2U-Net respectively. The proposed model leverages the power of U-Net, Residual Network, and RCNN. There are several advantages of these proposed architectures for segmentation tasks. First, the residual block is useful for training deep architectures. Second, the accumulation of features through recurrent residual convolutional layers improves the feature representation for segmentation tasks. Third, it allows us to design a better U-Net architecture with the same number of network parameters and improved performance on medical image segmentation. The proposed model is tested on three benchmark datasets, including retinal image blood vessel segmentation, skin cancer segmentation, and lung lesion segmentation. Experimental results show superior performance in segmentation tasks compared to comparable models such as U-Net and Residual U-Net (ResU-Net).