Learning to find occlusion regions

Ahmad Humayun, Oisin Mac Aodha, Gabriel J. Brostow

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

Abstract

For two consecutive frames in a video, we identify which pixels in the first frame become occluded in the second. Such general-purpose detection of occlusion regions is difficult and important because one-to-one correspondence of imaged scene points is needed for many tracking, video segmentation, and reconstruction algorithms. Our hypothesisis that an effective trained occlusion detector can be generated on the basis of i) a broad spectrum of visual features, and ii) representative but synthetic training sequences. By using a Random Forest based framework for feature selection and training, we found that the proposed feature set was sufficient to frequently assign a high probability of occlusion to just the pixels that were indeed becoming occluded. Our extensive experiments on many sequences support this finding, and while accuracy is certainly still scene-dependent, the proposed classifier could be a useful preprocessing step to exploit temporal information in video.
Original languageEnglish
Title of host publicationCVPR 2011
PublisherInstitute of Electrical and Electronics Engineers (IEEE)
Pages2161-2168
Number of pages8
ISBN (Electronic)978-1-4577-0395-9
ISBN (Print)978-1-4577-0394-2
DOIs
Publication statusPublished - 22 Aug 2011
Event2011 IEEE Computer Vision and Pattern Recognition - Colorado Springs, United States
Duration: 20 Jun 201125 Jun 2015
http://tab.computer.org/pamitc/archive/cvpr2011/index.html

Publication series

Name
PublisherInstitute of Electrical and Electronics Engineers (IEEE)
ISSN (Print)1063-6919
ISSN (Electronic)1063-6919

Conference

Conference2011 IEEE Computer Vision and Pattern Recognition
Abbreviated titleCVPR 2011
Country/TerritoryUnited States
CityColorado Springs
Period20/06/1125/06/15
Internet address

Keywords

  • feature extraction
  • video signal processing
  • feature training
  • feature selection
  • random forest
  • visual features
  • imaged scene points
  • general-purpose occlusion region detection
  • pixels
  • video frames
  • Motion segmentation
  • Training
  • Image color analysis
  • Optical imaging
  • Image edge detection
  • Optical variables control
  • Computer vision

Fingerprint

Dive into the research topics of 'Learning to find occlusion regions'. Together they form a unique fingerprint.

Cite this