A robust regression scale of residual estimator: SSAC

摘要

The random sample consensus (RANSAC) has been established as the standard method for model estimation in the presence of outliers. RANSAC is an iterative method, and the number of iterations necessary to find a correct solution is exponential with the minimum number of data points needed to estimate the model. It is therefore of utmost importance to find the minimal parameterization of the model to estimate. The random sample consensus (SAC) regression estimates are associated with M-scale. The weakness of M estimation is the lack of consideration of the distribution of data and it is not a function of the overall data, because it uses the median as the weight value. This paper proposes a new sample consensus estimator based on S-estimator, namely S-estimator SAmple consensus (SSAC). The accuracy of the proposed method has been studied through simulation study with existing algorithms.