SCU-Net: Semantic Segmentation Network for Learning Channel Information on Remote Sensing Images

Figure 10

Prediction maps of the SCU-net-102-A and DFCN121C (the ground truth map is divided into 7 classes: 0 for background, 1 for woodland, 2 for bare land or wasteland, 3 for waters, 4 for buildings, 5 for roads, and 6 for furrows. The first row is the original high-resolution remote sensing image, the second row is the ground truth, the third row is the DFCN121C prediction map, and the fourth row is the SCU-net-102-A prediction map).