Abstract

A new algorithm is presented that estimates the displacement vector field from two successive image frames. In the case where the sequence is severely corrupted by additive (Gaussian or not, colored) noise of unknown covariance, then second-order statistics methods do not work well. However, we have studied this topic from a viewpoint different from the above to explore the fundamental limits in image optimal flow estimation. Our scheme is based on subpixel optimal flow estimation using the bispectrum in the parametric domain. The displacement vector of a moving object is estimated by solving linear equations involving third-order holograms and the matrix containing the Dirac delta function. To prove the feasibility of the proposed method, we compared it with a phase correlation technique and the nonparametric bispectrum method described in Res. Lett. Signal Process., ID 417915 (2008) . Our results show that our method is considerably more immune to the presence of noise.

© 2009 Optical Society of America

Shift- and rotation-invariant object reconstruction using the bispectrum

Brian M. Sadler and Georgios B. Giannakis
J. Opt. Soc. Am. A 9(1) 57-69 (1992)

Noise-resilient estimation of optical flow by use of overlapped basis functions

Sridhar Srinivasan and Rama Chellappa
J. Opt. Soc. Am. A 16(3) 493-507 (1999)

Image-reconstruction algorithm using Kalman filter weighting of the bispectrum

Dan Newman and Randy C. Van Vrankent
J. Opt. Soc. Am. A 9(10) 1724-1739 (1992)

Scene-based nonuniformity correction and enhancement: pixel statistics and subpixel motion

Wenyi Zhao and Chao Zhang
J. Opt. Soc. Am. A 25(7) 1668-1681 (2008)

Statistical performance modeling for superresolution: a discrete data-continuous reconstruction framework

Frédéric Champagnat, Guy Le Besnerais, and Caroline Kulcsár
J. Opt. Soc. Am. A 26(7) 1730-1746 (2009)