一种改进变换网络的域自适应语义分割网络

张峻宁, 苏群星, 王成, 徐超, 李一宁

A Domain Adaptive Semantic Segmentation Network Based on Improved Transformation Network

ZHANG Junning, SU Qunxing, WANG Cheng, XU Chao, LI Yining

表4 不同算法在GTA5→Cityscapes数据集上的语义分割结果比较

Tab.4 Comparison of semantic segmentation results of different algorithms in GTA5→Cityscapes datasets

网络 算法 road sidewalk Building wall fence pole t-light t-sign wegetation terrain DeepLab v2[14] Cycada 85.2 37.2 76.5 21.8 15.0 23.8 22.9 21.5 80.5 31.3 DCAN 82.3 26.7 77.4 23.7 20.5 20.4 30.3 15.9 80.9 25.4 CLAN 88.0 30.6 79.2 23.4 20.5 26.1 23.0 14.8 81.6 34.5 BLD 89.2 40.9 81.2 29.1 19.2 14.2 29.0 19.6 83.7 35.9 本文 90.9 42.3 82.1 30.8 18.5 16.7 31.5 20.8 85.9 33.7 Resnet[24] Cycada 86.7 35.6 80.1 19.8 17.5 38.0 39.9 41.5 82.7 27.9 DCAN 85.0 30.8 81.3 25.8 21.2 22.2 25.4 26.6 83.4 36.7 CLAN 87.0 27.1 79.6 27.3 23.3 28.3 35.5 24.2 83.6 27.4 BLD 91.0 44.7 84.2 34.6 27.6 30.2 36.0 36.0 85.0 43.6 本文 92.2 42.3 83.5 36.2 28.0 31.8 36.7 36.2 85.7 44.6 网络 算法 sky person rider car truck bus train motorbike bicycle mIoU DeepLab v2[14] Cycada 60.7 50.5 9.0 76.9 17.1 28.2 4.5 9.8 0 35.4 DCAN 69.5 52.6 11.1 79.6 24.9 21.2 1.30 17.0 6.70 36.2 CLAN 72.0 45.8 7.9 80.5 26.6 29.9 0.0 10.7 0.0 36.6 BLD 80.7 54.7 23.3 82.7 25.8 28.0 2.3 25.7 19.9 41.3 本文 83.3 55.9 23.6 82.1 27.7 29.4 2.2 26.8 21.2 42.4 Resnet[24] Cycada 73.6 64.9 19 65.0 12.0 28.6 4.5 31.1 42.0 42.7 DCAN 76.2 58.9 24.9 80.7 29.5 42.9 2.50 26.9 11.6 41.7 CLAN 74.2 58.6 28.0 76.2 33.1 36.7 6.7 31.9 31.4 43.2 BLD 83.0 58.6 31.6 83.3 35.3 49.7 3.3 28.8 35.6 48.5 本文 81.2 59.8 32.7 84.1 36.3 49.9 3.0 30.7 37.4 49.1