University of Hertfordshire

Joint optimisation for object class segmentation and dense stereo reconstruction

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

Standard

Joint optimisation for object class segmentation and dense stereo reconstruction. / Ladicky, L.; Sengupta, S.; Russell, C.; Sturgess, P.; Bastanlar, Yalin; Clocksin, William; Torr, P.H.S.

Proceedings of the British Machine Vision Conference 2010: BMVC. BMVA Press, 2010. p. paper 104.

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

Harvard

Ladicky, L, Sengupta, S, Russell, C, Sturgess, P, Bastanlar, Y, Clocksin, W & Torr, PHS 2010, Joint optimisation for object class segmentation and dense stereo reconstruction. in Proceedings of the British Machine Vision Conference 2010: BMVC. BMVA Press, pp. paper 104, British Machine Vision Conf 2010, Aberystwyth, United Kingdom, 31/08/10. https://doi.org/10.5244/C.24.104

APA

Ladicky, L., Sengupta, S., Russell, C., Sturgess, P., Bastanlar, Y., Clocksin, W., & Torr, P. H. S. (2010). Joint optimisation for object class segmentation and dense stereo reconstruction. In Proceedings of the British Machine Vision Conference 2010: BMVC (pp. paper 104). BMVA Press. https://doi.org/10.5244/C.24.104

Vancouver

Ladicky L, Sengupta S, Russell C, Sturgess P, Bastanlar Y, Clocksin W et al. Joint optimisation for object class segmentation and dense stereo reconstruction. In Proceedings of the British Machine Vision Conference 2010: BMVC. BMVA Press. 2010. p. paper 104 https://doi.org/10.5244/C.24.104

Author

Ladicky, L. ; Sengupta, S. ; Russell, C. ; Sturgess, P. ; Bastanlar, Yalin ; Clocksin, William ; Torr, P.H.S. / Joint optimisation for object class segmentation and dense stereo reconstruction. Proceedings of the British Machine Vision Conference 2010: BMVC. BMVA Press, 2010. pp. paper 104

Bibtex

@inproceedings{381fd92490444021b0b649a02912fab8,
title = "Joint optimisation for object class segmentation and dense stereo reconstruction",
abstract = "The problems of dense stereo reconstruction and object class segmentation can both be formulated as Conditional Random Field based labelling problems, in which every pixel in the image is assigned a label corresponding to either its disparity, or an object class such as road or building. While these two problems are mutually informative, no attempt has been made to jointly optimise their labellings. In this work we provide a principled energy minimisation framework that unifies the two problems and demonstrate that, by resolving ambiguities in real world data, joint optimisation of the two problems substantially improves performance. To evaluate our method, we augment the street view Leuven data set, producing 70 hand labelled object class and disparity maps. We hope that the release of these annotations will stimulate further work in the challenging domain of street-view analysis.",
author = "L. Ladicky and S. Sengupta and C. Russell and P. Sturgess and Yalin Bastanlar and William Clocksin and P.H.S. Torr",
note = "This work is supported by EPSRC research grants, HMGCC, TUBITAK researcher exchange grant, the IST Programme of the European Community, under the PASCAL2 Network of Excellence, IST-2007-216886.; British Machine Vision Conf 2010 ; Conference date: 31-08-2010 Through 03-09-2010",
year = "2010",
doi = "10.5244/C.24.104",
language = "English",
pages = "paper 104",
booktitle = "Proceedings of the British Machine Vision Conference 2010",
publisher = "BMVA Press",

}

RIS

TY - GEN

T1 - Joint optimisation for object class segmentation and dense stereo reconstruction

AU - Ladicky, L.

AU - Sengupta, S.

AU - Russell, C.

AU - Sturgess, P.

AU - Bastanlar, Yalin

AU - Clocksin, William

AU - Torr, P.H.S.

N1 - This work is supported by EPSRC research grants, HMGCC, TUBITAK researcher exchange grant, the IST Programme of the European Community, under the PASCAL2 Network of Excellence, IST-2007-216886.

PY - 2010

Y1 - 2010

N2 - The problems of dense stereo reconstruction and object class segmentation can both be formulated as Conditional Random Field based labelling problems, in which every pixel in the image is assigned a label corresponding to either its disparity, or an object class such as road or building. While these two problems are mutually informative, no attempt has been made to jointly optimise their labellings. In this work we provide a principled energy minimisation framework that unifies the two problems and demonstrate that, by resolving ambiguities in real world data, joint optimisation of the two problems substantially improves performance. To evaluate our method, we augment the street view Leuven data set, producing 70 hand labelled object class and disparity maps. We hope that the release of these annotations will stimulate further work in the challenging domain of street-view analysis.

AB - The problems of dense stereo reconstruction and object class segmentation can both be formulated as Conditional Random Field based labelling problems, in which every pixel in the image is assigned a label corresponding to either its disparity, or an object class such as road or building. While these two problems are mutually informative, no attempt has been made to jointly optimise their labellings. In this work we provide a principled energy minimisation framework that unifies the two problems and demonstrate that, by resolving ambiguities in real world data, joint optimisation of the two problems substantially improves performance. To evaluate our method, we augment the street view Leuven data set, producing 70 hand labelled object class and disparity maps. We hope that the release of these annotations will stimulate further work in the challenging domain of street-view analysis.

U2 - 10.5244/C.24.104

DO - 10.5244/C.24.104

M3 - Conference contribution

SP - paper 104

BT - Proceedings of the British Machine Vision Conference 2010

PB - BMVA Press

T2 - British Machine Vision Conf 2010

Y2 - 31 August 2010 through 3 September 2010

ER -