AMLESAC: A New Maximum Likelihood Robust Estimator
Anton Konouchine, Victor Gaganov, Vladimir Veznevets · 2005
Robust parameter estimation methods are the general tool in computer vision, widely used for such tasks as multiple view relation estimation and camera calibration. In this paper, we propose a new general robust maximum-likelihood estimator called AMLESAC, which is a noise adaptive variant of renowned MLESAC estimator. It adopts the same sampling strategy and seeks the solution to maximize the likelihood rather then some heuristic measure, but unlike MLESAC, it simultaneously estimates the outlier share γ and inlier noise level σ. Effective optimization for computation speed-up is also introduced. Results are given for both synthetic and real test data for different types of models. The algorithm is demonstrated to outperform previous approaches for the task of pose estimation and provide results equal or superior to other robust estimators in other tests.