Consistent and Asymptotically Statistically-Efficient Solution to Camera Motion Estimation

arXiv:2403.01174v1 Announce Type: new
Abstract: Given 2D point correspondences between an image pair, inferring the camera motion is a fundamental issue in the computer vision community. The existing works generally set out from the epipolar constraint and estimate the essential matrix, which is not optimal in the maximum likelihood (ML) sense. In this paper, we dive into the original measurement model with respect to the rotation matrix and normalized translation vector and formulate the ML problem. We then propose a …

abstract arxiv community computer computer vision consistent cs.cv image issue likelihood matrix sense set solution type vision