Convolutional cost aggregation for robust stereo matching

Somi Jeong, Seungryong Kim, Bumsub Ham, Kwanghoon Sohn

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

Although convolutional neural network (CNN)-based stereo matching methods have become increasingly popular thanks to their robustness, they primarily have been focused on the matching cost computation. By leveraging CNNs, we present a novel method for matching cost aggregation to boost the stereo matching performance. Our insight is to learn the convolution kernel within CNN architecture for cost aggregation in a fully convolutional manner. Tailored to cost aggregation problem, our method differs from handcrafted methods in terms of its convolutional aggregation through optimally learned CNNs. First, the matching cost is aggregated with cost volume unary network, and then optimized with explicit disparity boundary, estimated through disparity boundary pairwise network, within a global energy minimization. Experiments demonstrate that our method outperforms conventional hand-crafted aggregation methods.

Original languageEnglish
Title of host publication2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings
PublisherIEEE Computer Society
Pages2523-2527
Number of pages5
ISBN (Electronic)9781509021758
DOIs
Publication statusPublished - 2018 Feb 20
Event24th IEEE International Conference on Image Processing, ICIP 2017 - Beijing, China
Duration: 2017 Sep 172017 Sep 20

Publication series

NameProceedings - International Conference on Image Processing, ICIP
Volume2017-September
ISSN (Print)1522-4880

Other

Other24th IEEE International Conference on Image Processing, ICIP 2017
CountryChina
CityBeijing
Period17/9/1717/9/20

Fingerprint

Agglomeration
Costs
Neural networks
Network architecture
Convolution
Experiments

All Science Journal Classification (ASJC) codes

  • Software
  • Computer Vision and Pattern Recognition
  • Signal Processing

Cite this

Jeong, S., Kim, S., Ham, B., & Sohn, K. (2018). Convolutional cost aggregation for robust stereo matching. In 2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings (pp. 2523-2527). (Proceedings - International Conference on Image Processing, ICIP; Vol. 2017-September). IEEE Computer Society. https://doi.org/10.1109/ICIP.2017.8296737
Jeong, Somi ; Kim, Seungryong ; Ham, Bumsub ; Sohn, Kwanghoon. / Convolutional cost aggregation for robust stereo matching. 2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings. IEEE Computer Society, 2018. pp. 2523-2527 (Proceedings - International Conference on Image Processing, ICIP).
@inproceedings{88caac6b762b401d9f65ab116372f0bb,
title = "Convolutional cost aggregation for robust stereo matching",
abstract = "Although convolutional neural network (CNN)-based stereo matching methods have become increasingly popular thanks to their robustness, they primarily have been focused on the matching cost computation. By leveraging CNNs, we present a novel method for matching cost aggregation to boost the stereo matching performance. Our insight is to learn the convolution kernel within CNN architecture for cost aggregation in a fully convolutional manner. Tailored to cost aggregation problem, our method differs from handcrafted methods in terms of its convolutional aggregation through optimally learned CNNs. First, the matching cost is aggregated with cost volume unary network, and then optimized with explicit disparity boundary, estimated through disparity boundary pairwise network, within a global energy minimization. Experiments demonstrate that our method outperforms conventional hand-crafted aggregation methods.",
author = "Somi Jeong and Seungryong Kim and Bumsub Ham and Kwanghoon Sohn",
year = "2018",
month = "2",
day = "20",
doi = "10.1109/ICIP.2017.8296737",
language = "English",
series = "Proceedings - International Conference on Image Processing, ICIP",
publisher = "IEEE Computer Society",
pages = "2523--2527",
booktitle = "2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings",
address = "United States",

}

Jeong, S, Kim, S, Ham, B & Sohn, K 2018, Convolutional cost aggregation for robust stereo matching. in 2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings. Proceedings - International Conference on Image Processing, ICIP, vol. 2017-September, IEEE Computer Society, pp. 2523-2527, 24th IEEE International Conference on Image Processing, ICIP 2017, Beijing, China, 17/9/17. https://doi.org/10.1109/ICIP.2017.8296737

Convolutional cost aggregation for robust stereo matching. / Jeong, Somi; Kim, Seungryong; Ham, Bumsub; Sohn, Kwanghoon.

2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings. IEEE Computer Society, 2018. p. 2523-2527 (Proceedings - International Conference on Image Processing, ICIP; Vol. 2017-September).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Convolutional cost aggregation for robust stereo matching

AU - Jeong, Somi

AU - Kim, Seungryong

AU - Ham, Bumsub

AU - Sohn, Kwanghoon

PY - 2018/2/20

Y1 - 2018/2/20

N2 - Although convolutional neural network (CNN)-based stereo matching methods have become increasingly popular thanks to their robustness, they primarily have been focused on the matching cost computation. By leveraging CNNs, we present a novel method for matching cost aggregation to boost the stereo matching performance. Our insight is to learn the convolution kernel within CNN architecture for cost aggregation in a fully convolutional manner. Tailored to cost aggregation problem, our method differs from handcrafted methods in terms of its convolutional aggregation through optimally learned CNNs. First, the matching cost is aggregated with cost volume unary network, and then optimized with explicit disparity boundary, estimated through disparity boundary pairwise network, within a global energy minimization. Experiments demonstrate that our method outperforms conventional hand-crafted aggregation methods.

AB - Although convolutional neural network (CNN)-based stereo matching methods have become increasingly popular thanks to their robustness, they primarily have been focused on the matching cost computation. By leveraging CNNs, we present a novel method for matching cost aggregation to boost the stereo matching performance. Our insight is to learn the convolution kernel within CNN architecture for cost aggregation in a fully convolutional manner. Tailored to cost aggregation problem, our method differs from handcrafted methods in terms of its convolutional aggregation through optimally learned CNNs. First, the matching cost is aggregated with cost volume unary network, and then optimized with explicit disparity boundary, estimated through disparity boundary pairwise network, within a global energy minimization. Experiments demonstrate that our method outperforms conventional hand-crafted aggregation methods.

UR - http://www.scopus.com/inward/record.url?scp=85045338492&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85045338492&partnerID=8YFLogxK

U2 - 10.1109/ICIP.2017.8296737

DO - 10.1109/ICIP.2017.8296737

M3 - Conference contribution

AN - SCOPUS:85045338492

T3 - Proceedings - International Conference on Image Processing, ICIP

SP - 2523

EP - 2527

BT - 2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings

PB - IEEE Computer Society

ER -

Jeong S, Kim S, Ham B, Sohn K. Convolutional cost aggregation for robust stereo matching. In 2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings. IEEE Computer Society. 2018. p. 2523-2527. (Proceedings - International Conference on Image Processing, ICIP). https://doi.org/10.1109/ICIP.2017.8296737