Rejecting Mismatches by Correspondence Function

Abstract

A novel method ICF (Identifying point correspondences by Correspondence Function) is proposed for rejecting mismatches from given putative point correspondences. By analyzing the connotation of homography, we introduce a novel concept of correspondence function for two images of a general 3D scene, which captures the relationships between corresponding points by mapping a point in one image to its corresponding point in another. Since the correspondence functions are unknown in real applications, we also study how to estimate them from given putative correspondences, and propose an algorithm IECF (Iteratively Estimate Correspondence Function) based on diagnostic technique and SVM. Then, the proposed ICF method is able to reject the mismatches by checking whether they are consistent with the estimated correspondence functions. Extensive experiments on real images demonstrate the excellent performance of our proposed method. In addition, the ICF is a general method for rejecting mismatches, and it is applicable to images of rigid objects or images of nonrigid objects with unknown deformation.

Robust orientation estimate via inertial guided visual sample consensus

“…Least mean squares (LMS) and least median mean squares (LMedS) [24] belong to these type and are well known for their simplicity and computational effectiveness. However, the outliers are not thoroughly removed in the putative correspondences, which will result in false or even unacceptable model parameterizations in the subsequent procedures, i.e., the mistaken essential matrix from the epipolar constraints using the matches mixed with inliers and outliers.…”

Side-Scan Sonar Image Mosaic Using Couple Feature Points with Constraint of Track Line Positions

“…The reason may be that, in addition to the correct CFPs in the overlapping areas, there also exist more wrong ones. The RANSAC algorithm is sensitive to the high outlier rate and unable to eliminate these wrong CFPs when they account for a large proportion [35,36]. Thus, 40% is set as a typical overlapping ratio for the SSS measurement task [19,25].…”

Accurately estimating rigid transformations in registration using a boosting-inspired mechanism

“…To overcome this problem, some means must be found to evaluate the reliability of each match: either to classify them as correct or false, or to characterize the extent to which they are believed to be correct. Li and Hu published the first paper [20] dedicated to the evaluation of the established point matches; other papers have since followed [1,28,29]. Existing methods for point match evaluation are based on one or more of the following: (i) structural consistency, (ii) transformation consistency, (iii) robust statistics, and (iv) heuristics.…”

“…A Hough transform is used in [25] to eliminate mismatches in Hough space. It is assumed in [20] that the PPMs are associated via two correspondence functions: one associates points in the first shape to those in the second and the other associates the points in the second shape to those in the first, these functions are estimated using the subspace projection support vector machine regression method. If two matched points fail to satisfy any of these two functions, they will be rejected as a mismatch.…”