Multi-View Stereo for Community Photo Collections

DMCA

Multi-View Stereo for Community Photo Collections

Cached

Download Links

Other Repositories/Bibliography

DBLP

by Michael Goesele

Citations:

186 - 23 self

BibTeX

@MISC{Goesele_multi-viewstereo,
    author = {Michael Goesele},
    title = {Multi-View Stereo for Community Photo Collections},
    year = {}
}

OpenURL

Abstract

We present a multi-view stereo algorithm that addresses the extreme changes in lighting, scale, clutter, and other effects in large online community photo collections. Our idea is to intelligently choose images to match, both at a per-view and per-pixel level. We show that such adaptive view selection enables robust performance even with dramatic appearance variability. The stereo matching technique takes as input sparse 3D points reconstructed from structure-from-motion methods and iteratively grows surfaces from these points. Optimizing for surface normals within a photoconsistency measure significantly improves the matching results. While the focus of our approach is to estimate high-quality depth maps, we also show examples of merging the resulting depth maps into compelling scene reconstructions. We demonstrate our algorithm on standard multi-view stereo datasets and on casually acquired photo collections of famous scenes gathered from the Internet. 1

Keyphrases

multi-view stereo community photo collection structure-from-motion method per-pixel level surface normal standard multi-view stereo datasets high-quality depth map robust performance photoconsistency measure photo collection adaptive view selection multi-view stereo algorithm grows surface dramatic appearance variability matching result large online community photo collection extreme change depth map stereo matching technique famous scene scene reconstruction

DMCA