Title :
Estimation of trifocal tensor using GMM
Author :
Hu, Mingxing ; Yuan, Baozong
Author_Institution :
Inst. of Inf. Sci., Northern Jiaotong Univ., Beijing, China
fDate :
9/12/2002 12:00:00 AM
Abstract :
A novel estimation of a trifocal tensor based on the Gaussian mixture model (GMM) is presented. The mixture model is built assuming that the residuals of inliers and outliers belong to different Gaussian distributions. The Bayesian rule is then employed to detect the inliers for re-estimation. Experiments show that the presented method is more precise and relatively unaffected by outliers
Keywords :
Bayes methods; Gaussian distribution; covariance matrices; image motion analysis; tensors; Bayesian rule; GMM; Gaussian distributions; Gaussian mixture model; image analysis; image data; inliers; motion analysis; outliers; trifocal tensor estimation;
Journal_Title :
Electronics Letters
DOI :
10.1049/el:20020755
