Zobrazeno 1 - 10
of 123
pro vyhledávání: '"Quéau, Yvain"'
Advanced machine learning methods, and more prominently neural networks, have become standard to solve inverse problems over the last years. However, the theoretical recovery guarantees of such methods are still scarce and difficult to achieve. Only
Externí odkaz:
http://arxiv.org/abs/2403.05395
Autor:
Brument, Baptiste, Bruneau, Robin, Quéau, Yvain, Mélou, Jean, Lauze, François Bernard, Jean-Denis, Durou, Jean-Denis, Calvet, Lilian
This paper introduces a versatile paradigm for integrating multi-view reflectance (optional) and normal maps acquired through photometric stereo. Our approach employs a pixel-wise joint re-parameterization of reflectance and normal, considering them
Externí odkaz:
http://arxiv.org/abs/2312.01215
Neural networks have become a prominent approach to solve inverse problems in recent years. While a plethora of such methods was developed to solve inverse problems empirically, we are still lacking clear theoretical guarantees for these methods. On
Externí odkaz:
http://arxiv.org/abs/2309.12128
Neural networks have become a prominent approach to solve inverse problems in recent years. Amongst the different existing methods, the Deep Image/Inverse Priors (DIPs) technique is an unsupervised approach that optimizes a highly overparametrized ne
Externí odkaz:
http://arxiv.org/abs/2303.11265
The photometric stereo (PS) problem consists in reconstructing the 3D-surface of an object, thanks to a set of photographs taken under different lighting directions. In this paper, we propose a multi-scale architecture for PS which, combined with a n
Externí odkaz:
http://arxiv.org/abs/2211.14118
Publikováno v:
In Computer Vision and Image Understanding November 2024 248
Uncalibrated photometric stereo aims at estimating the 3D-shape of a surface, given a set of images captured from the same viewing angle, but under unknown, varying illumination. While the theoretical foundations of this inverse problem under directi
Externí odkaz:
http://arxiv.org/abs/1911.07268
Akademický článek
Tento výsledek nelze pro nepřihlášené uživatele zobrazit.
K zobrazení výsledku je třeba se přihlásit.
K zobrazení výsledku je třeba se přihlásit.
Publikováno v:
The IEEE International Conference on Computer Vision (ICCV), 2019
Photometric stereo (PS) techniques nowadays remain constrained to an ideal laboratory setup where modeling and calibration of lighting is amenable. To eliminate such restrictions, we propose an efficient principled variational approach to uncalibrate
Externí odkaz:
http://arxiv.org/abs/1904.03942
This study is concerned with the design of a Mueller imaging polarimeter for the visualization of spatially-varying Mueller matrix fields. A simplified calibration procedure is advocated, where all the optical elements are calibrated simultaneously r
Externí odkaz:
http://arxiv.org/abs/1811.02683