Edge-Preserving Integration of a Normal Field: Weighted Least Squares and L1 Approaches

Yvain Quéau (1) , Jean-Denis Durou (1)

1 IRIT-REVA - Real Expression Artificial Life

We introduce several new functionals, inspired from variational image denoising models, for recovering a piecewise-smooth surface from a dense estimation of its normal field. In the weighted least-squares approach, the non-differentiable elements of the surface are a priori detected so as to weight the least-squares model. To avoid this detection step, we introduce reweighted least-squares for minimising an isotropic TV-like functional, and split-Bregman iterations for L1 minimisation.

Mots clés: Photometric stereo; Integration; Shape-from-gradient

Domaines: Synthèse d'image et réalité virtuelle [cs.GR]; Traitement du signal et de l'image [eess.SP]; Traitement des images [eess.IV]; Vision par ordinateur et reconnaissance de formes [cs.CV]; Intelligence artificielle [cs.AI]

Yvain Quéau, Jean-Denis Durou. Edge-Preserving Integration of a Normal Field: Weighted Least Squares and L1 Approaches. 5th International Conference on Scale Space and Variational Methods in Computer Vision (SSVM 2015), May 2015, Lège Cap Ferret, France. pp. 576-588. ⟨hal-01360871⟩