Textured Object Recognition: Balancing Model Robustness and Complexity

Guido Manfredi (1) , Michel Devy (1) , Daniel Sidobre (2)

1 LAAS-RAP - Équipe Robotique, Action et Perception
2 LAAS-RIS - Équipe Robotique et InteractionS

When it comes to textured object modelling, the standard practice is to use a multiple view approach. The numerous views allow reconstruction and provide robustness to viewpoint change but yield complex models. This paper shows that robustness with lighter models can be achieved through robust descriptors. A comparison between various descriptors allows choosing the one providing the best viewpoint robust-ness, in this case the ASIFT descriptor. Then, using this descriptor, the results show, for a wide variety of object shapes, that as few as seventeen views provide a high level of robustness to viewpoint change while being fast to process and using small memory space. This work concludes advocating in favour of modelling methods using robust descriptors and a small number of views.

Mots clés: object modelling; object recognition; multiple views; robust descriptors

Domaines: Vision par ordinateur et reconnaissance de formes [cs.CV]

Guido Manfredi, Michel Devy, Daniel Sidobre. Textured Object Recognition: Balancing Model Robustness and Complexity. 16th International Conference on Computer Analysis of Images and Patterns (CAIP 2015), Sep 2015, La Valette, Malta. ⟨10.1007/978-3-319-23192-1_5⟩. ⟨hal-01355103⟩