Behavior analysis of Evolution Strategy Sample Consensus
Yuichiro Toda, Naoyuki Kubota · 2014
Recently, robust estimators are expected in various fields such as signal processing and machine learning. In our previous work, we proposed Evolution Strategy Sample Consensus (ESSAC) as a new robust estimator method and improved a trade off between a calculation time and stability of SAmple Consensus (SAC) algorithms. In this paper, we show several experiments for behavior analysis of ESSAC in order to discuss why ESSAC enable to search stably in the dataset including the huge number of noises. and analyze several experiments related with the fitness function of SAC and the behavior of ESSAC.