%0 Generic %1 wang2017understanding %A Wang, Panqu %A Chen, Pengfei %A Yuan, Ye %A Liu, Ding %A Huang, Zehua %A Hou, Xiaodi %A Cottrell, Garrison %D 2017 %K %T Understanding Convolution for Semantic Segmentation %U http://arxiv.org/abs/1702.08502 %X Recent advances in deep learning, especially deep convolutional neural networks (CNNs), have led to significant improvement over previous semantic segmentation systems. Here we show how to improve pixel-wise semantic segmentation by manipulating convolution-related operations that are of both theoretical and practical value. First, we design dense upsampling convolution (DUC) to generate pixel-level prediction, which is able to capture and decode more detailed information that is generally missing in bilinear upsampling. Second, we propose a hybrid dilated convolution (HDC) framework in the encoding phase. This framework 1) effectively enlarges the receptive fields (RF) of the network to aggregate global information; 2) alleviates what we call the "gridding issue" caused by the standard dilated convolution operation. We evaluate our approaches thoroughly on the Cityscapes dataset, and achieve a state-of-art result of 80.1% mIOU in the test set at the time of submission. We also have achieved state-of-the-art overall on the KITTI road estimation benchmark and the PASCAL VOC2012 segmentation task. Our source code can be found at https://github.com/TuSimple/TuSimple-DUC .

@misc{wang2017understanding, abstract = {Recent advances in deep learning, especially deep convolutional neural networks (CNNs), have led to significant improvement over previous semantic segmentation systems. Here we show how to improve pixel-wise semantic segmentation by manipulating convolution-related operations that are of both theoretical and practical value. First, we design dense upsampling convolution (DUC) to generate pixel-level prediction, which is able to capture and decode more detailed information that is generally missing in bilinear upsampling. Second, we propose a hybrid dilated convolution (HDC) framework in the encoding phase. This framework 1) effectively enlarges the receptive fields (RF) of the network to aggregate global information; 2) alleviates what we call the "gridding issue" caused by the standard dilated convolution operation. We evaluate our approaches thoroughly on the Cityscapes dataset, and achieve a state-of-art result of 80.1% mIOU in the test set at the time of submission. We also have achieved state-of-the-art overall on the KITTI road estimation benchmark and the PASCAL VOC2012 segmentation task. Our source code can be found at https://github.com/TuSimple/TuSimple-DUC .}, added-at = {2021-01-20T13:02:59.000+0100}, author = {Wang, Panqu and Chen, Pengfei and Yuan, Ye and Liu, Ding and Huang, Zehua and Hou, Xiaodi and Cottrell, Garrison}, biburl = {https://www.bibsonomy.org/bibtex/2f99c2fe900754f3678d85abbb57bd168/philipphaas}, interhash = {4934f836653c8e5b7cd33e2e317c6788}, intrahash = {f99c2fe900754f3678d85abbb57bd168}, keywords = {}, note = {cite arxiv:1702.08502Comment: WACV 2018. Updated acknowledgements. Source code: https://github.com/TuSimple/TuSimple-DUC}, timestamp = {2021-01-20T13:02:59.000+0100}, title = {Understanding Convolution for Semantic Segmentation}, url = {http://arxiv.org/abs/1702.08502}, year = 2017 }