Multi-View Inverse Rendering

Introduction of Multi-View Inverse Rendering (MVIR)

Multi-View Inverse Rendering (MVIR) aims at recovering detailed 3D shape from multi-view images under unconstrained conditions. To achieve this, 3D shape, albedo and illumination consisting a scene are modelled to compose an inverse rendering framework. An optimization problem is solved to estimated detailed 3D shape, as well as albedo and illumination.

Spectral Multi-View Inverse Rendering (Spectral MVIR) [Project Page]

The flowchart of Spectral MVIR using multi-view RGB images captured with a standard camera and a low-cost LED bulb.

Publication:
Spectral MVIR: Joint Reconstruction of 3D Shape and Spectral Reflectance
Chunyu Li, Yusuke Monno, Masatoshi Okutomi
IEEE International Conference on Computational Photography (ICCP 2021)

Polarimetric Multi-View Inverse Rendering (Polarimetric MVIR) [Project Page]

The flowchart of Polarimetric MVIR using multi-view RGB and AoP images.

Publication:
Polarimetric Multi-View Inverse Rendering
Jinyu Zhao, Yusuke Monno, Masatoshi Okutomi
European Conference on Computer Vision (ECCV 2020)

Multi-View Inverse Rendering (MVIR) [Project Page]

The flowchart of MVIR under arbitrary illumination and albedo.

Publication:
Multi-View Inverse Rendering under Arbitrary Illumination and Albedo
Kichang Kim, Akihiko Torii, Masatoshi Okutomi
European Conference on Computer Vision (ECCV 2016), pp. 750–767, 2016.