A More General Robust Loss Function
Preprint
Abstract
We present a two-parameter loss function which can be viewed as a generalization of many popular loss functions used in robust statistics: the Cauchy/Lorentzian, Geman-McClure, Welsch, and generalized Charbonnier loss functions(and by transitivity the L2, L1, L1-L2, and pseudo-Huber/Charbonnier loss functions). We describe and visualize this penalty, and document several of its useful properties.Keywords
All Related Versions
This publication has 10 references indexed in Scilit:
- Fast MRF Optimization with Application to Depth Reconstruction2014
- Efficient Nonlocal Regularization for Optical Flow2012
- Secrets of optical flow estimation and their principles2010
- Two deterministic half-quadratic regularization algorithms for computed imaging2002
- Parameter estimation techniques: a tutorial with application to conic fittingImage and Vision Computing, 1997
- On the unification of line processes, outlier rejection, and robust statistics with applications in early visionInternational Journal of Computer Vision, 1996
- The Robust Estimation of Multiple Motions: Parametric and Piecewise-Smooth Flow FieldsComputer Vision and Image Understanding, 1996
- Constructing simple stable descriptions for image partitioningInternational Journal of Computer Vision, 1989
- Techniques for nonlinear least squares and robust regressionCommunications in Statistics - Simulation and Computation, 1978
- Robust Estimation of a Location ParameterThe Annals of Mathematical Statistics, 1964