Skip to Main content Skip to Navigation
New interface
Conference papers

Discriminant random field and patch-based redundancy analysis for image change detection

Abstract : To develop better image change detection algorithms, new models able to capture all the spatio-temporal regularities and geometries seen in an image pair are needed. In con- trast to the usual pixel-wise methods, we propose a patch- based formulation for modeling semi-local interactions and detecting occlusions and other local or regional changes in an image pair. To this end, the image redundancy property is exploited to detect unusual spatio-temporal patterns in the scene. We first define adaptive detectors of changes between two given image patches and combine locally in space and scale such detectors. The resulting score at a given loca- tion is exploited within a discriminant Markov random field (DRF) whose global optimization flags out changes with no optical flow computation. Experimental results on several applications demonstrate that the method performs well at detecting occlusions and meaningful regional changes and is especially robust in the case of low signal-to-noise ratios.
Complete list of metadata

Cited literature [33 references]  Display  Hide  Download
Contributor : Charles Kervrann Connect in order to contact the contributor
Submitted on : Tuesday, February 26, 2013 - 4:29:50 PM
Last modification on : Monday, November 7, 2022 - 2:24:09 PM
Long-term archiving on: : Sunday, April 2, 2017 - 5:39:57 AM


Files produced by the author(s)


  • HAL Id : hal-00794897, version 1


Charles Kervrann, Jérôme Boulanger, Thierry Pecot, Patrick Pérez. Discriminant random field and patch-based redundancy analysis for image change detection. EEE Int. Workshop on Machine Learning for Signal Processing, Sep 2009, Grenoble, France. ⟨hal-00794897⟩



Record views


Files downloads