Hypersphere Fitting: Model, Algorithms and Future Work
Speaker: Jean-Yves Tourneret (Professeur à l'INP-ENSEEIHT, Toulouse)
Title: Hypersphere Fitting: Model, Algorithms and Future Work
Abstract: In this talk, I will describe some expectation maximisation (EM) algorithms allowing Lidar point clouds associated with mixtures of hyperspheres to be fitted. The first algorithm is adapted to a unique hypersphere. it relies on the introduction of latent vectors having a priori independent von-Mises Fisher distributions defined on the hypersphere. This statistical model leads to a complete data likelihood whose expected value, conditioned on the observed data, has a Von Mises-Fisher distribution. As a result, the inference problem can be solved by a simple EM algorithm. A robust version of this algorithm allows the presence of outliers to be mitigated. A generalization to mixtures of hypersheres will be finally discussed, with an application to the calibration of several Lidars. Some future works will conclude this talk.
More information: https://perso.tesa.prd.fr/jyt/index.html
Talk in room M7 101 (Campus Monod, ENS de Lyon)