Moving object segmentation using scene understanding

Perera, A. G. A., Brooksby, G., Hoogs, A., Doretto, G., and Kaufhold, J.
Moving object segmentation using scene understanding. Technical Report 2005GRC182, GE Global Research, 2005. Visualization and Computer Vision Laboratory

Download

(unavailable)

Abstract

We present a novel approach to improving moving object detection algorithms for aerial video by incorporating scene understanding. We argue that scene understanding is a valuable input feature for moving object segmentation, and show how it can be used to reduce the false positive rate of any moving object detection algorithm. The approach is especially effective on sequences with large scene depth and much parallax, as often occurs with low, fast-moving sensors. In one difficult sequence, our results show an 8-fold reduction of false positives on average, with essentially no impact on the true positive rate, even when scene understanding is only moderately accurate. We show how scene understanding can be used to increase the accuracy of frame-to-frame homography estimates, and also present a novel approach that uses motion information from the previous frame for outlier rejection at the current frame to further increase accuracy.

@TECHREPORT{pereraBHDK05tr,
  author = {Perera, A. G. A. and Brooksby, G. and Hoogs, A. and Doretto, G. and
	Kaufhold, J.},
  title = {Moving object segmentation using scene understanding},
  institution = {GE Global Research},
  year = {2005},
  number = {2005GRC182},
  address = {Niskayuna, NY, USA},
  month = mar,
  note = {Visualization and Computer Vision Laboratory},
  bib2html_pubtype = {Tech Reports},
  bib2html_rescat = {Video Surveillance, Visual Motion Segmentation},
  abstract = {We present a novel approach to improving moving object detection algorithms
	for aerial video by incorporating scene understanding. We argue that
	scene understanding is a valuable input feature for moving object
	segmentation, and show how it can be used to reduce the false positive
	rate of any moving object detection algorithm. The approach is especially
	effective on sequences with large scene depth and much parallax,
	as often occurs with low, fast-moving sensors. In one difficult sequence,
	our results show an 8-fold reduction of false positives on average,
	with essentially no impact on the true positive rate, even when scene
	understanding is only moderately accurate. We show how scene understanding
	can be used to increase the accuracy of frame-to-frame homography
	estimates, and also present a novel approach that uses motion information
	from the previous frame for outlier rejection at the current frame
	to further increase accuracy.},
  file = {pereraBHDK05tr.pdf:doretto\\report\\pereraBHDK05tr.pdf:PDF},
  keywords = {motion segmentation, moving object detection, scene understanding,
	scene content, recognition},
  owner = {doretto},
  pdf = {doretto\report\pereraBHDK05tr.pdf},
  timestamp = {2006.11.29}
}

Generated by bib2html.pl (written by Patrick Riley )