MAGSAC++, a fast, reliable and accurate robust estimator

TL;DR

MAGSAC++ introduces a novel robust estimation method that improves accuracy and speed by eliminating the need for inlier-outlier decisions, utilizing a new scoring function, a marginalization approach, and an innovative sampling technique.

Contribution

It presents a new model quality function, a marginalization procedure, and the Progressive NAPSAC sampler, enhancing robustness and efficiency over existing methods.

Findings

Outperforms state-of-the-art methods on real datasets

Faster and more accurate than previous robust estimators

Less prone to failure in practical applications

Abstract

A new method for robust estimation, MAGSAC++, is proposed. It introduces a new model quality (scoring) function that does not require the inlier-outlier decision, and a novel marginalization procedure formulated as an iteratively re-weighted least-squares approach. We also propose a new sampler, Progressive NAPSAC, for RANSAC-like robust estimators. Exploiting the fact that nearby points often originate from the same model in real-world data, it finds local structures earlier than global samplers. The progressive transition from local to global sampling does not suffer from the weaknesses of purely localized samplers. On six publicly available real-world datasets for homography and fundamental matrix fitting, MAGSAC++ produces results superior to state-of-the-art robust methods. It is faster, more geometrically accurate and fails less often.