Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
410480 | Neurocomputing | 2009 | 10 Pages |
Epipolar geometry relies on the determination of the fundamental matrix. Classical approaches for estimating the fundamental matrix assume that a Gaussian distribution exists in the errors in view of mathematical tractability. However, this assumption will not be justified when the distribution computed is not normally distributed. We propose a new approach that does not make the Gaussian assumption, and so can attain robustness and accuracy in different conditions. The proposed framework, weighted least squares (WLS), is the application of linear mixed-effect models considering the correlation between different data subsamples. It provides an unbiased estimation of the fundamental matrix after mitigating the effects of outliers. We test the new model by using synthetic and real images, and comparing it to standard methods.