Do We Really Have to Consider Covariance Matrices for Image Feature Points?

概要

論文の詳細を見る
We have explored various statistical optimization techniques based on covariance matrices that characterize the uncertainty of the positions of feature points in the images. We first describe how to compute the covariance matrix of a feature point from the gray levels by integrating existing methods. Then, we experimentally examine if thus computed covariance matrices really reflect the accuracy of the feature positions. For this purpose, we observe the correlation between the feature covariance and the amount of sub-pixel correction resulting from variable template matching, using real images. We also test if the accuracy of computing the homography and the fundamental matrices from two images can be really improved by statistical optimization based on the covariance matrices.
社団法人電子情報通信学会の論文
2002-02-01