End-to-end depth map compression framework via RGB-to-depth structure priors learning

Minghui CHEN; Pingping ZHANG; Zhuo CHEN; Yun ZHANG; Xu WAGN; Sam KWONG

doi:10.1109/ICIP46576.2022.9898073

End-to-end depth map compression framework via RGB-to-depth structure priors learning

Minghui CHEN, Pingping ZHANG, Zhuo CHEN, Yun ZHANG, Xu WAGN, 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 novel framework to exploit and utilize the shared information inner RGB-D data for efficient depth map compression. Two main codecs, designed based on the existing end-to-end image compression network, are adopted for RGB image compression and enhanced depth image compression with RGB-to-Depth structure prior, respectively. In particular, we propose a Structure Prior Fusion (SPF) module to extract the structure information from both RGB and depth codecs at multi-scale feature levels and fuse the cross-modal feature to generate more efficient structure priors for depth compression. Extensive experiments show that the proposed framework can achieve competitive rate-distortion performance as well as RGB-D task-specific performance at depth map compression compared with the direct compression scheme.

Original language	English
Title of host publication	Proceedings of the 2022 IEEE International Conference on Image Processing
Publisher	IEEE
Pages	3206-3210
Number of pages	5
ISBN (Print)	9781665496209
DOIs	https://doi.org/10.1109/ICIP46576.2022.9898073
Publication status	Published - Oct 2022
Externally published	Yes
Event	29th IEEE International Conference on Image Processing, ICIP 2022 - Bordeaux, France Duration: 16 Oct 2022 → 19 Oct 2022

Conference

Conference	29th IEEE International Conference on Image Processing, ICIP 2022
Country/Territory	France
City	Bordeaux
Period	16/10/22 → 19/10/22

Bibliographical note

This work was supported in part by the National Natural Science Foundation of China (Grant 61871270), in part by the Shenzhen Natural Science Foundation under Grants JCYJ20200109110410133 and 20200812110350001, in part by the A*STAR under YIRG project No. A2084c0176 and AHSF project No. C211118005.

Keywords

cross-modal
Depth map compression
feature fusion

Access to Document

10.1109/ICIP46576.2022.9898073

Cite this

@inproceedings{93e3a79678a4495c8142bdc032339caf,

title = "End-to-end depth map compression framework via RGB-to-depth structure priors learning",

abstract = "In this paper, we propose a novel framework to exploit and utilize the shared information inner RGB-D data for efficient depth map compression. Two main codecs, designed based on the existing end-to-end image compression network, are adopted for RGB image compression and enhanced depth image compression with RGB-to-Depth structure prior, respectively. In particular, we propose a Structure Prior Fusion (SPF) module to extract the structure information from both RGB and depth codecs at multi-scale feature levels and fuse the cross-modal feature to generate more efficient structure priors for depth compression. Extensive experiments show that the proposed framework can achieve competitive rate-distortion performance as well as RGB-D task-specific performance at depth map compression compared with the direct compression scheme.",

keywords = "cross-modal, Depth map compression, feature fusion",

author = "Minghui CHEN and Pingping ZHANG and Zhuo CHEN and Yun ZHANG and Xu WAGN and Sam KWONG",

note = "This work was supported in part by the National Natural Science Foundation of China (Grant 61871270), in part by the Shenzhen Natural Science Foundation under Grants JCYJ20200109110410133 and 20200812110350001, in part by the A*STAR under YIRG project No. A2084c0176 and AHSF project No. C211118005.; 29th IEEE International Conference on Image Processing, ICIP 2022 ; Conference date: 16-10-2022 Through 19-10-2022",

year = "2022",

month = oct,

doi = "10.1109/ICIP46576.2022.9898073",

language = "English",

isbn = "9781665496209",

pages = "3206--3210",

booktitle = "Proceedings of the 2022 IEEE International Conference on Image Processing",

publisher = "IEEE",

}

CHEN, M, ZHANG, P, CHEN, Z, ZHANG, Y, WAGN, X & KWONG, S 2022, End-to-end depth map compression framework via RGB-to-depth structure priors learning. in Proceedings of the 2022 IEEE International Conference on Image Processing. IEEE, pp. 3206-3210, 29th IEEE International Conference on Image Processing, ICIP 2022, Bordeaux, France, 16/10/22. https://doi.org/10.1109/ICIP46576.2022.9898073

TY - GEN

T1 - End-to-end depth map compression framework via RGB-to-depth structure priors learning

AU - CHEN, Minghui

AU - ZHANG, Pingping

AU - CHEN, Zhuo

AU - ZHANG, Yun

AU - WAGN, Xu

AU - KWONG, Sam

N1 - This work was supported in part by the National Natural Science Foundation of China (Grant 61871270), in part by the Shenzhen Natural Science Foundation under Grants JCYJ20200109110410133 and 20200812110350001, in part by the A*STAR under YIRG project No. A2084c0176 and AHSF project No. C211118005.

PY - 2022/10

Y1 - 2022/10

N2 - In this paper, we propose a novel framework to exploit and utilize the shared information inner RGB-D data for efficient depth map compression. Two main codecs, designed based on the existing end-to-end image compression network, are adopted for RGB image compression and enhanced depth image compression with RGB-to-Depth structure prior, respectively. In particular, we propose a Structure Prior Fusion (SPF) module to extract the structure information from both RGB and depth codecs at multi-scale feature levels and fuse the cross-modal feature to generate more efficient structure priors for depth compression. Extensive experiments show that the proposed framework can achieve competitive rate-distortion performance as well as RGB-D task-specific performance at depth map compression compared with the direct compression scheme.

AB - In this paper, we propose a novel framework to exploit and utilize the shared information inner RGB-D data for efficient depth map compression. Two main codecs, designed based on the existing end-to-end image compression network, are adopted for RGB image compression and enhanced depth image compression with RGB-to-Depth structure prior, respectively. In particular, we propose a Structure Prior Fusion (SPF) module to extract the structure information from both RGB and depth codecs at multi-scale feature levels and fuse the cross-modal feature to generate more efficient structure priors for depth compression. Extensive experiments show that the proposed framework can achieve competitive rate-distortion performance as well as RGB-D task-specific performance at depth map compression compared with the direct compression scheme.

KW - cross-modal

KW - Depth map compression

KW - feature fusion

UR - http://www.scopus.com/inward/record.url?scp=85146644314&partnerID=8YFLogxK

U2 - 10.1109/ICIP46576.2022.9898073

DO - 10.1109/ICIP46576.2022.9898073

M3 - Conference paper (refereed)

SN - 9781665496209

SP - 3206

EP - 3210

BT - Proceedings of the 2022 IEEE International Conference on Image Processing

PB - IEEE

T2 - 29th IEEE International Conference on Image Processing, ICIP 2022

Y2 - 16 October 2022 through 19 October 2022

ER -

End-to-end depth map compression framework via RGB-to-depth structure priors learning

Abstract

Conference

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this