Feature matching constrained by cross ratio invariance

The main aim of this work is to propose a new technique to solve the well-k nown feature correspondence problem for motion estimation. The problem is f ormulated as an optimization process whose energy function includes constra ints based on projective invariance of cross-ratio of five coplanar points. Starting from some approximated correspondences, estimated by radiometric similarity, for features with high directional variance, optimal matches ar e obtained through an optimization technique. The new contribution of this work consists of a matching process, refining the raw measurements, based o n an energy function minimization technique converging to an optimal soluti on for most of the features by taking advantage of some good initial guess, and in the use of cross ratio as geometrical invariant constraint to detec t and correct the mismatches due to wrong radiometric similarity measures. Though the method is based on geometrical invariance of coplanar points, it is not required that all features have to be coplanar or to preprocess the images to detect the planar regions. Experimental results are presented fo r real and synthetic images, and the performance of the novel approach is e valuated on different image sequences and compared to well-known techniques . (C) 2000 Pattern Recognition Soceity. Published by Elsevier Science Ltd. All rights reserved.