Motion Classification Based On Geometrical Features Of Trajectories

Archive ouverte

Nardi, Giacomo | Santos Sano, Matheus | Brelot, Anne | Olivo-Marin, Jean-Christophe | Lagache, Thibault

Edité par CCSD ; IEEE -

International audience. This paper proposes a novel approach for motion clas-sification based on geometrical features computed on tra-jectories. The method follows a machine learning approachtrained and validated on synthetic datasets simulating severalstochastic models. The resulting model enables, in particular,the recognition of different subdiffusive behaviors, offering afiner classification than the standard method based on meansquare displacement. The method is assessed on a biologicaldataset containing trajectories of CCR5 cell receptors.

• Description
• Sujet/Public
• Outil
• Outil d'anticipation
• Mémoire et thèse
• Gestion