Model based clustering for 3D directional features: application to depth image analysis

Abul Hasnat (1) , Olivier Alata (1) , A. Trémeau (1)

1 LHC - Laboratoire Hubert Curien

Model based clustering (MBC) is a method that selects an op- timal clustering solution from a set of candidate solutions. In this paper, we propose a novel MBC method to cluster three dimensional directional features. We assume that the features are generated from a finite statistical mixture model based on the von Mises-Fisher (vMF) distribution. The core elements of our proposed method are: (a) generate a set of vMF Mixture Models (vMFMM) and (b) select the optimal model using a parsimony based approach with information criteria. We empirically validate our proposed method by applying it on simulated data. Next, we apply it to cluster image normals in order to perform depth image analysis.

Keywords: Model based clustering; Model selection; von Mises-Fisher distribution; Mixture model; Depth image analysis

Domains: Image Processing [eess.IV]

Abul Hasnat, Olivier Alata, A. Trémeau. Model based clustering for 3D directional features: application to depth image analysis. IEEE International Conference on Image Processing (ICIP) 2014, Oct 2014, Paris, France. pp.1-5. ⟨ujm-01005184⟩