Efficient semantic segmentation of man-made scenes using fully-connected conditional random field

Weihao Li, Michael Ying Yang*

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

10 Citations (Scopus)
17 Downloads (Pure)

Abstract

In this paper we explore semantic segmentation of man-made scenes using fully connected conditional random field (CRF). Images of man-made scenes display strong contextual dependencies in the spatial structures. Fully connected CRFs can model long-range connections within the image of man-made scenes and make use of contextual information of scene structures. The pairwise edge potentials of fully connected CRF models are defined by a linear combination of Gaussian kernels. Using filter-based mean field algorithm, the inference is very efficient. Our experimental results demonstrate that fully connected CRF performs better than previous state-of-The-Art approaches on both eTRIMS dataset and LabelMeFacade dataset.

Original languageEnglish
Title of host publicationThe International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Subtitle of host publicationXXIII ISPRS Congress
Place of PublicationPrague
PublisherInternational Society for Photogrammetry and Remote Sensing (ISPRS)
Pages633-640
Number of pages8
Volume41
EditionB3
DOIs
Publication statusPublished - Jul 2016
Event23rd International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences Congress, ISPRS 2016 - Prague, Czech Republic
Duration: 12 Jul 201619 Jul 2016
Conference number: 23
http://www.isprs.org/publications/archives.aspx (Full text Open Access proceedings)

Publication series

NameInternational Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences - ISPRS Archives
PublisherCopernicus
ISSN (Print)1682-1750

Conference

Conference23rd International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences Congress, ISPRS 2016
Abbreviated titleISPRS 2016
CountryCzech Republic
CityPrague
Period12/07/1619/07/16
Internet address

Keywords

  • Fully Connected CRFs
  • Man-made Scene
  • Mean Field Inference
  • Semantic Segmentation

Fingerprint Dive into the research topics of 'Efficient semantic segmentation of man-made scenes using fully-connected conditional random field'. Together they form a unique fingerprint.

  • Cite this

    Li, W., & Yang, M. Y. (2016). Efficient semantic segmentation of man-made scenes using fully-connected conditional random field. In The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences: XXIII ISPRS Congress (B3 ed., Vol. 41, pp. 633-640). (International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences - ISPRS Archives). Prague: International Society for Photogrammetry and Remote Sensing (ISPRS). https://doi.org/10.5194/isprsarchives-XLI-B3-633-2016