1. Deeplab v3+：Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation ECCV 2018

Fig 1

2. BiSeNet: Bilateral Segmentation Network for Real-time Semantic Segmentation ECCV 2018

Fig 2

3. Dual Attention Network for Scene Segmentation Paper Link

Fig 3

4. Dense Decoder Shortcut Connections for Single-Pass Semantic Segmentation CVPR 2018

Fig 4