Fundamental Matrix Estimation using Generalized Least Squares

H. Zhou, P.R. Green, A.M. Wallace, and S. Xu (UK)


Fundamental matrix; Epipolar geometry; Leastsquare; Outlier.


Classical approaches for estimating the fundamental matrix assume that Gaussian noise is contained in the estimates in view of mathematical tractability. How ever, this assumption will not be justified when the distribution computed is not normally distributed. We propose a new approach that does not make the Gaus sian assumption, and so can attain robustness and ac curacy in different conditions. The proposed frame work, generalized least squares (GLS), is the extension of linear mixed-effects models considering the correla tion between different data subsamples. We test the new model by using synthetic and real images, com paring it to the least median of squares technique .

Important Links:

Go Back