CONSAC: Robust Multi-Model Fitting by Conditional Sample Consensus
This addresses the challenge of detecting multiple models in computer vision tasks like scene analysis, offering a learned alternative to hand-crafted methods.
The paper tackles the problem of robustly fitting multiple parametric models to noisy measurements, such as vanishing points or homographies, by learning a search strategy from data. The proposed method, CONSAC, outperforms existing robust estimators and specialized algorithms, achieving superior accuracy in both supervised vanishing point estimation and self-supervised multi-homography estimation.
We present a robust estimator for fitting multiple parametric models of the same form to noisy measurements. Applications include finding multiple vanishing points in man-made scenes, fitting planes to architectural imagery, or estimating multiple rigid motions within the same sequence. In contrast to previous works, which resorted to hand-crafted search strategies for multiple model detection, we learn the search strategy from data. A neural network conditioned on previously detected models guides a RANSAC estimator to different subsets of all measurements, thereby finding model instances one after another. We train our method supervised as well as self-supervised. For supervised training of the search strategy, we contribute a new dataset for vanishing point estimation. Leveraging this dataset, the proposed algorithm is superior with respect to other robust estimators as well as to designated vanishing point estimation algorithms. For self-supervised learning of the search, we evaluate the proposed algorithm on multi-homography estimation and demonstrate an accuracy that is superior to state-of-the-art methods.