CVPR2019論文サマリ

tag: cityscapes

Dual Attention Network for Scene Segmentation
by: Hideki Tsunashima
Semantic Segmentation SOTA Attention Cityscapes PASCAL Context COCO stuff dataset
Decoders Matter for Semantic Segmentation: Data-Dependent Decoding Enables Flexible Feature Aggregation
by: Hideki Tsunashima
semantic segmentation upsampling 1x1 convolution DUpsampling SOTA PASCAL VOC PASCAL Context Cityscapes