Preemptive RANSAC for Live Structure and Motion Estimation

  • Authors:
  • David Nistér

  • Affiliations:
  • -

  • Venue:
  • ICCV '03 Proceedings of the Ninth IEEE International Conference on Computer Vision - Volume 2
  • Year:
  • 2003

Quantified Score

Hi-index 0.00

Visualization

Abstract

A system capable of performing robust live ego-motion estimationfor perspective cameras is presented. The system is powered byrandom sample consensus with preemptive scoring of the motionhypotheses. A general statement of the problem of efficientpreemptive scoring is given. Then a theoretical investigation ofpreemptive scoring under a simple inlier-outlier model isperformed. A practical preemption scheme is proposed and it isshown that the preemption is powerful enough to enable robust livestructure and motion estimation.