Compression Artifacts Reduction for Depth Map by Deep Intensity Guidance

Pingping ZHANG; Xu WANG; Yun ZHANG; Lin MA; Jianmin JIANG; Sam KWONG

doi:10.1007/978-3-319-77380-3_83

Compression Artifacts Reduction for Depth Map by Deep Intensity Guidance

Pingping ZHANG, Xu WANG, Yun ZHANG, Lin MA, Jianmin JIANG, Sam KWONG

Research output: Book Chapters | Papers in Conference Proceedings › Conference paper (refereed) › Research › peer-review

1 Citation (Scopus)

Abstract

In this paper, we propose a intensity guided CNN (IG-Net) model, which learns an end-to-end mapping between the intensity image and distorted depth map to the uncompressed depth map. To eliminate the undesired blocking artifacts such as discontinuities around object boundary, two branches are designed to extract the high-frequency information from intensity image and depth map, respectively. Multi-scale feature fusion and enhancement layers are introduced in the main branch to strength the edge information of the restored depth map. Performance evaluation on JPEG compression artifacts shows the effectiveness and superiority of our proposed model compared with state-of-the-art methods.

Original language	English
Title of host publication	Advances in Multimedia Information Processing : PCM 2017
Editors	Bing ZENG, Qingming HUANG, Abdulmotaleb El SADDIK, Hongliang LI, Shuqiang JIANG, Xiaopeng FAN
Publisher	Springer, Cham
Pages	863-872
Number of pages	10
ISBN (Electronic)	9783319773803
ISBN (Print)	9783319773797
DOIs	https://doi.org/10.1007/978-3-319-77380-3_83
Publication status	Published - 2018
Externally published	Yes
Event	18th Pacific-Rim Conference on Multimedia, PCM 2017 - Harbin, China Duration: 28 Sept 2017 → 29 Sept 2017

Publication series

Name	Lecture Notes in Computer Science (including subseries Information Systems and Applications, incl. Internet/Web, and HCI)
Publisher	Springer, Cham
Volume	10735
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	18th Pacific-Rim Conference on Multimedia, PCM 2017
Country/Territory	China
City	Harbin
Period	28/09/17 → 29/09/17

Funding

This work was supported in part by the National Natural Science Foundation of China under Grant 61501299, 61471348, 61672443 and 61620106008, in part by the Guangdong Nature Science Foundation under Grant 2016A030310058, in part by the Shenzhen Emerging Industries of the Strategic Basic Research Project under Grants JCYJ20150525092941043, JCYJ20160226191842793, in part by the Project 2016049 supported by SZU R/D Fund, and in part by the Tencent “Rhinoceros Birds”-Scientific Research Foundation for Young Teachers of Shenzhen University.

Keywords

Compression artifacts
Convolutional neural network
JPEG compression

Access to Document

10.1007/978-3-319-77380-3_83

Cite this

ZHANG, P., WANG, X., ZHANG, Y., MA, L., JIANG, J., & KWONG, S. (2018). Compression Artifacts Reduction for Depth Map by Deep Intensity Guidance. In B. ZENG, Q. HUANG, A. E. SADDIK, H. LI, S. JIANG, & X. FAN (Eds.), Advances in Multimedia Information Processing : PCM 2017 (pp. 863-872). (Lecture Notes in Computer Science (including subseries Information Systems and Applications, incl. Internet/Web, and HCI); Vol. 10735). Springer, Cham. https://doi.org/10.1007/978-3-319-77380-3_83

ZHANG, Pingping ; WANG, Xu ; ZHANG, Yun et al. / Compression Artifacts Reduction for Depth Map by Deep Intensity Guidance. Advances in Multimedia Information Processing : PCM 2017. editor / Bing ZENG ; Qingming HUANG ; Abdulmotaleb El SADDIK ; Hongliang LI ; Shuqiang JIANG ; Xiaopeng FAN. Springer, Cham, 2018. pp. 863-872 (Lecture Notes in Computer Science (including subseries Information Systems and Applications, incl. Internet/Web, and HCI)).

@inproceedings{e33bd5a20e1143718a789f9fc59916d5,

title = "Compression Artifacts Reduction for Depth Map by Deep Intensity Guidance",

abstract = "In this paper, we propose a intensity guided CNN (IG-Net) model, which learns an end-to-end mapping between the intensity image and distorted depth map to the uncompressed depth map. To eliminate the undesired blocking artifacts such as discontinuities around object boundary, two branches are designed to extract the high-frequency information from intensity image and depth map, respectively. Multi-scale feature fusion and enhancement layers are introduced in the main branch to strength the edge information of the restored depth map. Performance evaluation on JPEG compression artifacts shows the effectiveness and superiority of our proposed model compared with state-of-the-art methods.",

keywords = "Compression artifacts, Convolutional neural network, JPEG compression",

author = "Pingping ZHANG and Xu WANG and Yun ZHANG and Lin MA and Jianmin JIANG and Sam KWONG",

year = "2018",

doi = "10.1007/978-3-319-77380-3_83",

language = "English",

isbn = "9783319773797",

series = "Lecture Notes in Computer Science (including subseries Information Systems and Applications, incl. Internet/Web, and HCI)",

publisher = "Springer, Cham",

pages = "863--872",

editor = "Bing ZENG and Qingming HUANG and SADDIK, {Abdulmotaleb El} and LI, {Hongliang } and Shuqiang JIANG and Xiaopeng FAN",

booktitle = "Advances in Multimedia Information Processing : PCM 2017",

note = "18th Pacific-Rim Conference on Multimedia, PCM 2017 ; Conference date: 28-09-2017 Through 29-09-2017",

}

ZHANG, P, WANG, X, ZHANG, Y, MA, L, JIANG, J & KWONG, S 2018, Compression Artifacts Reduction for Depth Map by Deep Intensity Guidance. in B ZENG, Q HUANG, AE SADDIK, H LI, S JIANG & X FAN (eds), Advances in Multimedia Information Processing : PCM 2017. Lecture Notes in Computer Science (including subseries Information Systems and Applications, incl. Internet/Web, and HCI), vol. 10735, Springer, Cham, pp. 863-872, 18th Pacific-Rim Conference on Multimedia, PCM 2017, Harbin, China, 28/09/17. https://doi.org/10.1007/978-3-319-77380-3_83

Compression Artifacts Reduction for Depth Map by Deep Intensity Guidance. / ZHANG, Pingping; WANG, Xu; ZHANG, Yun et al.
Advances in Multimedia Information Processing : PCM 2017. ed. / Bing ZENG; Qingming HUANG; Abdulmotaleb El SADDIK; Hongliang LI; Shuqiang JIANG; Xiaopeng FAN. Springer, Cham, 2018. p. 863-872 (Lecture Notes in Computer Science (including subseries Information Systems and Applications, incl. Internet/Web, and HCI); Vol. 10735).

Research output: Book Chapters | Papers in Conference Proceedings › Conference paper (refereed) › Research › peer-review

TY - GEN

T1 - Compression Artifacts Reduction for Depth Map by Deep Intensity Guidance

AU - ZHANG, Pingping

AU - WANG, Xu

AU - ZHANG, Yun

AU - MA, Lin

AU - JIANG, Jianmin

AU - KWONG, Sam

PY - 2018

Y1 - 2018

N2 - In this paper, we propose a intensity guided CNN (IG-Net) model, which learns an end-to-end mapping between the intensity image and distorted depth map to the uncompressed depth map. To eliminate the undesired blocking artifacts such as discontinuities around object boundary, two branches are designed to extract the high-frequency information from intensity image and depth map, respectively. Multi-scale feature fusion and enhancement layers are introduced in the main branch to strength the edge information of the restored depth map. Performance evaluation on JPEG compression artifacts shows the effectiveness and superiority of our proposed model compared with state-of-the-art methods.

AB - In this paper, we propose a intensity guided CNN (IG-Net) model, which learns an end-to-end mapping between the intensity image and distorted depth map to the uncompressed depth map. To eliminate the undesired blocking artifacts such as discontinuities around object boundary, two branches are designed to extract the high-frequency information from intensity image and depth map, respectively. Multi-scale feature fusion and enhancement layers are introduced in the main branch to strength the edge information of the restored depth map. Performance evaluation on JPEG compression artifacts shows the effectiveness and superiority of our proposed model compared with state-of-the-art methods.

KW - Compression artifacts

KW - Convolutional neural network

KW - JPEG compression

UR - http://www.scopus.com/inward/record.url?scp=85047467579&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-77380-3_83

DO - 10.1007/978-3-319-77380-3_83

M3 - Conference paper (refereed)

SN - 9783319773797

T3 - Lecture Notes in Computer Science (including subseries Information Systems and Applications, incl. Internet/Web, and HCI)

SP - 863

EP - 872

BT - Advances in Multimedia Information Processing : PCM 2017

A2 - ZENG, Bing

A2 - HUANG, Qingming

A2 - SADDIK, Abdulmotaleb El

A2 - LI, Hongliang

A2 - JIANG, Shuqiang

A2 - FAN, Xiaopeng

PB - Springer, Cham

T2 - 18th Pacific-Rim Conference on Multimedia, PCM 2017

Y2 - 28 September 2017 through 29 September 2017

ER -

ZHANG P, WANG X, ZHANG Y, MA L, JIANG J, KWONG S. Compression Artifacts Reduction for Depth Map by Deep Intensity Guidance. In ZENG B, HUANG Q, SADDIK AE, LI H, JIANG S, FAN X, editors, Advances in Multimedia Information Processing : PCM 2017. Springer, Cham. 2018. p. 863-872. (Lecture Notes in Computer Science (including subseries Information Systems and Applications, incl. Internet/Web, and HCI)). doi: 10.1007/978-3-319-77380-3_83

Compression Artifacts Reduction for Depth Map by Deep Intensity Guidance

Abstract

Publication series

Conference

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this