Exploiting deep convolutional network and patch-level CRFs for indoor semantic segmentation

Xingming Wu; Mengnan Du; Weihai Chen; Zhengguo Li

doi:10.1109/ICIEA.2016.7603568

Exploiting deep convolutional network and patch-level CRFs for indoor semantic segmentation

Xingming Wu
, Mengnan Du
, Weihai Chen
, Zhengguo Li

School of Automation Science and Electrical Engineering

Beihang University
Agency for Science, Technology and Research, Singapore

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

4 Scopus citations

Abstract

Recent advances in semantic segmentation have mostly been achieved by utilizing deep convolutional neural networks (CNNs). In this paper, a novel indoor semantic segmentation method is proposed by integrating CNN and patch-level Conditional random fields (CRF). Multi-scale images are sent to CNN to capture objects of different sizes as well as to extract features at multiple scales. Patch-level CRF is constructed to further refine the object boundaries localization accuracy. Extensive experiments on the publicly available NYU V2 database demonstrate that the proposed method could obtain state-of-the-art accuracy in terms of four evaluation metrics.

Original language	English
Title of host publication	Proceedings of the 2016 IEEE 11th Conference on Industrial Electronics and Applications, ICIEA 2016
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	150-155
Number of pages	6
ISBN (Electronic)	9781509026050
DOIs	https://doi.org/10.1109/ICIEA.2016.7603568
State	Published - 19 Oct 2016
Event	11th IEEE Conference on Industrial Electronics and Applications, ICIEA 2016 - Hefei, China Duration: 5 Jun 2016 → 7 Jun 2016

Publication series

Name	Proceedings of the 2016 IEEE 11th Conference on Industrial Electronics and Applications, ICIEA 2016

Conference

Conference	11th IEEE Conference on Industrial Electronics and Applications, ICIEA 2016
Country/Territory	China
City	Hefei
Period	5/06/16 → 7/06/16

Access to Document

10.1109/ICIEA.2016.7603568

Cite this

Wu, X., Du, M., Chen, W., & Li, Z. (2016). Exploiting deep convolutional network and patch-level CRFs for indoor semantic segmentation. In Proceedings of the 2016 IEEE 11th Conference on Industrial Electronics and Applications, ICIEA 2016 (pp. 150-155). Article 7603568 (Proceedings of the 2016 IEEE 11th Conference on Industrial Electronics and Applications, ICIEA 2016). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICIEA.2016.7603568

Wu, Xingming ; Du, Mengnan ; Chen, Weihai et al. / Exploiting deep convolutional network and patch-level CRFs for indoor semantic segmentation. Proceedings of the 2016 IEEE 11th Conference on Industrial Electronics and Applications, ICIEA 2016. Institute of Electrical and Electronics Engineers Inc., 2016. pp. 150-155 (Proceedings of the 2016 IEEE 11th Conference on Industrial Electronics and Applications, ICIEA 2016).

@inproceedings{6046c4476fef48c7a60c4bb7f78e50ef,

title = "Exploiting deep convolutional network and patch-level CRFs for indoor semantic segmentation",

abstract = "Recent advances in semantic segmentation have mostly been achieved by utilizing deep convolutional neural networks (CNNs). In this paper, a novel indoor semantic segmentation method is proposed by integrating CNN and patch-level Conditional random fields (CRF). Multi-scale images are sent to CNN to capture objects of different sizes as well as to extract features at multiple scales. Patch-level CRF is constructed to further refine the object boundaries localization accuracy. Extensive experiments on the publicly available NYU V2 database demonstrate that the proposed method could obtain state-of-the-art accuracy in terms of four evaluation metrics.",

author = "Xingming Wu and Mengnan Du and Weihai Chen and Zhengguo Li",

note = "Publisher Copyright: {\textcopyright} 2016 IEEE.; 11th IEEE Conference on Industrial Electronics and Applications, ICIEA 2016 ; Conference date: 05-06-2016 Through 07-06-2016",

year = "2016",

month = oct,

day = "19",

doi = "10.1109/ICIEA.2016.7603568",

language = "英语",

series = "Proceedings of the 2016 IEEE 11th Conference on Industrial Electronics and Applications, ICIEA 2016",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "150--155",

booktitle = "Proceedings of the 2016 IEEE 11th Conference on Industrial Electronics and Applications, ICIEA 2016",

address = "美国",

}

Wu, X, Du, M, Chen, W & Li, Z 2016, Exploiting deep convolutional network and patch-level CRFs for indoor semantic segmentation. in Proceedings of the 2016 IEEE 11th Conference on Industrial Electronics and Applications, ICIEA 2016., 7603568, Proceedings of the 2016 IEEE 11th Conference on Industrial Electronics and Applications, ICIEA 2016, Institute of Electrical and Electronics Engineers Inc., pp. 150-155, 11th IEEE Conference on Industrial Electronics and Applications, ICIEA 2016, Hefei, China, 5/06/16. https://doi.org/10.1109/ICIEA.2016.7603568

Exploiting deep convolutional network and patch-level CRFs for indoor semantic segmentation. / Wu, Xingming; Du, Mengnan; Chen, Weihai et al.
Proceedings of the 2016 IEEE 11th Conference on Industrial Electronics and Applications, ICIEA 2016. Institute of Electrical and Electronics Engineers Inc., 2016. p. 150-155 7603568 (Proceedings of the 2016 IEEE 11th Conference on Industrial Electronics and Applications, ICIEA 2016).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Exploiting deep convolutional network and patch-level CRFs for indoor semantic segmentation

AU - Wu, Xingming

AU - Du, Mengnan

AU - Chen, Weihai

AU - Li, Zhengguo

PY - 2016/10/19

Y1 - 2016/10/19

N2 - Recent advances in semantic segmentation have mostly been achieved by utilizing deep convolutional neural networks (CNNs). In this paper, a novel indoor semantic segmentation method is proposed by integrating CNN and patch-level Conditional random fields (CRF). Multi-scale images are sent to CNN to capture objects of different sizes as well as to extract features at multiple scales. Patch-level CRF is constructed to further refine the object boundaries localization accuracy. Extensive experiments on the publicly available NYU V2 database demonstrate that the proposed method could obtain state-of-the-art accuracy in terms of four evaluation metrics.

AB - Recent advances in semantic segmentation have mostly been achieved by utilizing deep convolutional neural networks (CNNs). In this paper, a novel indoor semantic segmentation method is proposed by integrating CNN and patch-level Conditional random fields (CRF). Multi-scale images are sent to CNN to capture objects of different sizes as well as to extract features at multiple scales. Patch-level CRF is constructed to further refine the object boundaries localization accuracy. Extensive experiments on the publicly available NYU V2 database demonstrate that the proposed method could obtain state-of-the-art accuracy in terms of four evaluation metrics.

UR - https://www.scopus.com/pages/publications/84997217115

U2 - 10.1109/ICIEA.2016.7603568

DO - 10.1109/ICIEA.2016.7603568

M3 - 会议稿件

AN - SCOPUS:84997217115

T3 - Proceedings of the 2016 IEEE 11th Conference on Industrial Electronics and Applications, ICIEA 2016

SP - 150

EP - 155

BT - Proceedings of the 2016 IEEE 11th Conference on Industrial Electronics and Applications, ICIEA 2016

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 11th IEEE Conference on Industrial Electronics and Applications, ICIEA 2016

Y2 - 5 June 2016 through 7 June 2016

ER -

Wu X, Du M, Chen W, Li Z. Exploiting deep convolutional network and patch-level CRFs for indoor semantic segmentation. In Proceedings of the 2016 IEEE 11th Conference on Industrial Electronics and Applications, ICIEA 2016. Institute of Electrical and Electronics Engineers Inc. 2016. p. 150-155. 7603568. (Proceedings of the 2016 IEEE 11th Conference on Industrial Electronics and Applications, ICIEA 2016). doi: 10.1109/ICIEA.2016.7603568

Exploiting deep convolutional network and patch-level CRFs for indoor semantic segmentation

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this