The paper "Road Extraction by Deep Residual U-Net" by Zhengxin Zhang, Qingjie Liu, and Yunhong Wang proposes a semantic segmentation neural network that combines the strengths of residual learning and U-Net for road area extraction from high-resolution remote sensing images. The proposed network, called Deep ResUnet, uses residual units instead of plain neural units, which facilitates training and information propagation. The network architecture is similar to U-Net but with fewer parameters, as it eliminates the cropping operation. The Deep ResUnet is tested on the Massachusetts roads dataset and compared with U-Net and other state-of-the-art methods. The results show that the Deep ResUnet outperforms all compared methods in terms of relaxed precision and recall, demonstrating its superior performance and efficiency. The network's ability to handle context information and robustness to occlusions are highlighted, making it a promising approach for road extraction tasks.The paper "Road Extraction by Deep Residual U-Net" by Zhengxin Zhang, Qingjie Liu, and Yunhong Wang proposes a semantic segmentation neural network that combines the strengths of residual learning and U-Net for road area extraction from high-resolution remote sensing images. The proposed network, called Deep ResUnet, uses residual units instead of plain neural units, which facilitates training and information propagation. The network architecture is similar to U-Net but with fewer parameters, as it eliminates the cropping operation. The Deep ResUnet is tested on the Massachusetts roads dataset and compared with U-Net and other state-of-the-art methods. The results show that the Deep ResUnet outperforms all compared methods in terms of relaxed precision and recall, demonstrating its superior performance and efficiency. The network's ability to handle context information and robustness to occlusions are highlighted, making it a promising approach for road extraction tasks.