Semantic segmentation using stride spatial pyramid pooling and dual attention decoder

作者：

Highlights：

• We propose an SSPP structure to capture multiscale semantic information.

• Attention mechanism is applied to bridge the information gap in segmentation networks.

• We propose a new decoder to make full use of the low- and high-level feature maps.

• Auxiliary loss is applied to make the network easier to train.

• Our method attains state-of-the-art performance on PASCAL VOC 2012, Cityscapes and COCO-Stuff.

摘要

•We propose an SSPP structure to capture multiscale semantic information.•Attention mechanism is applied to bridge the information gap in segmentation networks.•We propose a new decoder to make full use of the low- and high-level feature maps.•Auxiliary loss is applied to make the network easier to train.•Our method attains state-of-the-art performance on PASCAL VOC 2012, Cityscapes and COCO-Stuff.

论文关键词：Semantic segmentation,Convolutional neural networks,Pyramid pooling,Attention mechanism

论文评审过程：Received 7 February 2019, Revised 24 May 2020, Accepted 12 June 2020, Available online 13 June 2020, Version of Record 18 June 2020.

论文官网地址：https://doi.org/10.1016/j.patcog.2020.107498