Oct 25, 2012
Multi-view stereo usually aims at reconstructing photo-consistent 3D models from images. This is typically achieved by optimizing some photo-consistency measure, relating the reconstructed shape, and possibly a reconstructed appearance, to the input images. In this talk, we consider how to optimize the most intuitive photo-consistency measure, which directly compares input images to images rendered from the reconstructed 3D model. Principled optimization of photo-consistency is shown to be applicable to various 3D modeling scenarios, from the classical multi-view stereo or shape-from-shading settings for objects with Lambertian reflectance properties, to the modeling of surfaces with partly specular reflectance.
Peter Sturm
Date:
25-10-12
Time:
14:15 -15:45
Location:
TU Vienna
Zemanek Lecture Room (Room Number: HHEG01)
HS 13 Ernst Melan
1040 Wien, Karlsplatz 13 (Main building, Stiege VII, 2nd floor)