TY - GEN
T1 - RANet: Region attention network for semantic segmentation
AU - Shen, Dingguo
AU - Ji, Yuanfeng
AU - Li, Ping
AU - Wang, Yi
AU - Lin, Di
N1 - Funding Information:
We thank the anonymous reviewers and editors for their constructive suggestions. This work was supported in parts by NSFC (61702338, 61701312), Natural Science Foundation of Guangdong Province (2019A1515010847), the Macau Science and Technology Development Fund under grant (0027/2018/A1), and The Hong Kong Polytechnic University under grants (P0030419, P0030929).
Publisher Copyright:
© 2020 Neural information processing systems foundation. All rights reserved.
PY - 2020/12
Y1 - 2020/12
N2 - Recent semantic segmentation methods model the relationship between pixels to construct the contextual representations. In this paper, we introduce the Region Attention Network (RANet), a novel attention network for modeling the relationship between object regions. RANet divides the image into object regions, where we select the representative information. In contrast to the previous methods, RANet configures the information pathways between the pixels in different regions, enabling the region interaction to exchange the regional context for enhancing all of the pixels in the image. We train the construction of object regions, the selection of the representative regional contents, the configuration of information pathways and the context exchange between pixels, jointly, to improve the segmentation accuracy. We extensively evaluate our method on the challenging segmentation benchmarks, demonstrating that RANet effectively helps to achieve the state-of-the-art results. Code will be available at: https://github.com/dingguo1996/RANet.
AB - Recent semantic segmentation methods model the relationship between pixels to construct the contextual representations. In this paper, we introduce the Region Attention Network (RANet), a novel attention network for modeling the relationship between object regions. RANet divides the image into object regions, where we select the representative information. In contrast to the previous methods, RANet configures the information pathways between the pixels in different regions, enabling the region interaction to exchange the regional context for enhancing all of the pixels in the image. We train the construction of object regions, the selection of the representative regional contents, the configuration of information pathways and the context exchange between pixels, jointly, to improve the segmentation accuracy. We extensively evaluate our method on the challenging segmentation benchmarks, demonstrating that RANet effectively helps to achieve the state-of-the-art results. Code will be available at: https://github.com/dingguo1996/RANet.
UR - http://www.scopus.com/inward/record.url?scp=85101834481&partnerID=8YFLogxK
M3 - Conference article published in proceeding or book
AN - SCOPUS:85101834481
VL - 33
T3 - Advances in Neural Information Processing Systems
SP - 13927
EP - 13938
BT - Advances in Neural Information Processing Systems
T2 - 34th Conference on Neural Information Processing Systems, NeurIPS 2020
Y2 - 6 December 2020 through 12 December 2020
ER -