Learning single-image depth from videos using quality assessment networks

Weifeng Chen; Shengyi Qian; Jia Deng

doi:10.1109/CVPR.2019.00575

Learning single-image depth from videos using quality assessment networks

Weifeng Chen, Shengyi Qian, Jia Deng

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

57 Scopus citations

Abstract

Depth estimation from a single image in the wild remains a challenging problem. One main obstacle is the lack of high-quality training data for images in the wild. In this paper we propose a method to automatically generate such data through Structure-from-Motion (SfM) on Internet videos. The core of this method is a Quality Assessment Network that identifies high-quality reconstructions obtained from SfM. Using this method, we collect single-view depth training data from a large number of YouTube videos and construct a new dataset called YouTube3D. Experiments show that YouTube3D is useful in training depth estimation networks and advances the state of the art of single-view depTH ESTIMAtion in the wild.

Original language	English (US)
Title of host publication	2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
Publisher	IEEE
Pages	5597-5606
Number of pages	10
ISBN (Print)	9781728132938
DOIs	https://doi.org/10.1109/CVPR.2019.00575
State	Published - Jan 9 2020
Externally published	Yes

Bibliographical note

KAUST Repository Item: Exported on 2022-06-30
Acknowledged KAUST grant number(s): OSR-2015-CRG4-2639
Acknowledgements: This publication is based upon work partially supported by National Science Foundation under Grant No. 1617767, the King Abdullah University of Science and Technology (KAUST) Office of Sponsored Research (OSR) under Award No. OSR-2015-CRG4-2639 and a gift from Google.
This publication acknowledges KAUST support, but has no KAUST affiliated authors.

Access to Document

10.1109/CVPR.2019.00575

Cite this

@inproceedings{823430a145744f02840cf584893bb07e,

title = "Learning single-image depth from videos using quality assessment networks",

abstract = "Depth estimation from a single image in the wild remains a challenging problem. One main obstacle is the lack of high-quality training data for images in the wild. In this paper we propose a method to automatically generate such data through Structure-from-Motion (SfM) on Internet videos. The core of this method is a Quality Assessment Network that identifies high-quality reconstructions obtained from SfM. Using this method, we collect single-view depth training data from a large number of YouTube videos and construct a new dataset called YouTube3D. Experiments show that YouTube3D is useful in training depth estimation networks and advances the state of the art of single-view depTH ESTIMAtion in the wild.",

author = "Weifeng Chen and Shengyi Qian and Jia Deng",

note = "KAUST Repository Item: Exported on 2022-06-30 Acknowledged KAUST grant number(s): OSR-2015-CRG4-2639 Acknowledgements: This publication is based upon work partially supported by National Science Foundation under Grant No. 1617767, the King Abdullah University of Science and Technology (KAUST) Office of Sponsored Research (OSR) under Award No. OSR-2015-CRG4-2639 and a gift from Google. This publication acknowledges KAUST support, but has no KAUST affiliated authors.",

year = "2020",

month = jan,

day = "9",

doi = "10.1109/CVPR.2019.00575",

language = "English (US)",

isbn = "9781728132938",

pages = "5597--5606",

booktitle = "2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)",

publisher = "IEEE",

}

TY - GEN

T1 - Learning single-image depth from videos using quality assessment networks

AU - Chen, Weifeng

AU - Qian, Shengyi

AU - Deng, Jia

N1 - KAUST Repository Item: Exported on 2022-06-30 Acknowledged KAUST grant number(s): OSR-2015-CRG4-2639 Acknowledgements: This publication is based upon work partially supported by National Science Foundation under Grant No. 1617767, the King Abdullah University of Science and Technology (KAUST) Office of Sponsored Research (OSR) under Award No. OSR-2015-CRG4-2639 and a gift from Google. This publication acknowledges KAUST support, but has no KAUST affiliated authors.

PY - 2020/1/9

Y1 - 2020/1/9

N2 - Depth estimation from a single image in the wild remains a challenging problem. One main obstacle is the lack of high-quality training data for images in the wild. In this paper we propose a method to automatically generate such data through Structure-from-Motion (SfM) on Internet videos. The core of this method is a Quality Assessment Network that identifies high-quality reconstructions obtained from SfM. Using this method, we collect single-view depth training data from a large number of YouTube videos and construct a new dataset called YouTube3D. Experiments show that YouTube3D is useful in training depth estimation networks and advances the state of the art of single-view depTH ESTIMAtion in the wild.

AB - Depth estimation from a single image in the wild remains a challenging problem. One main obstacle is the lack of high-quality training data for images in the wild. In this paper we propose a method to automatically generate such data through Structure-from-Motion (SfM) on Internet videos. The core of this method is a Quality Assessment Network that identifies high-quality reconstructions obtained from SfM. Using this method, we collect single-view depth training data from a large number of YouTube videos and construct a new dataset called YouTube3D. Experiments show that YouTube3D is useful in training depth estimation networks and advances the state of the art of single-view depTH ESTIMAtion in the wild.

UR - http://hdl.handle.net/10754/679477

UR - https://ieeexplore.ieee.org/document/8953964/

UR - http://www.scopus.com/inward/record.url?scp=85078792988&partnerID=8YFLogxK

U2 - 10.1109/CVPR.2019.00575

DO - 10.1109/CVPR.2019.00575

M3 - Conference contribution

SN - 9781728132938

SP - 5597

EP - 5606

BT - 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

PB - IEEE

ER -

Learning single-image depth from videos using quality assessment networks

Abstract

Bibliographical note

Access to Document

Other files and links

Fingerprint

Cite this