Non-smooth M-estimator for maximum consensus estimation
Document Type
Conference Proceeding
Publisher
BMVA Press
School
School of Science
RAS ID
30763
Funders
Australian Research Council
Grant Number
ARC Number : FT170100072, ARC Number : FT140101229, ARC Number : DP160103490
Grant Link
http://purl.org/au-research/grants/arc/FT170100072
Abstract
This paper revisits the application of M-estimators for a spectrum of robust estimation problems in computer vision, particularly with the maximum consensus criterion. Current practice makes use of smooth robust loss functions, e.g. Huber loss, which enables M-estimators to be tackled by such well-known optimization techniques as Iteratively Re-weighted Least Square (IRLS). When consensus maximization is used as loss function for M-estimators, however, the optimization problem becomes non-smooth. Our paper proposes an approach to resolve this issue. Based on the Alternating Direction Method of Multiplier (ADMM) technique, we develop a deterministic algorithm that is provably convergent, which enables the maximum consensus problem to be solved in the context of M-estimator. We further show that our algorithm outperforms other differentiable robust loss functions that are currently used by many practitioners. Notably, the proposed method allows the sub-problems to be solved efficiently in parallel, thus entails it to be implemented in distributed settings.
Access Rights
subscription content
Comments
Le, H., Eriksson, A., Milford, M., Do, T. T., Chin, T. J., & Suter, D. (2018). Non-smooth M-estimator for maximum consensus estimation supplementary material. In British Machine Vision Conference 2018. Newcastle, United Kingdom: Northumbria University. Available here