Depth Synthesis and Local Warps for Plausible Image-based Navigation
Modern camera calibration and multi-view stereo techniques enable users to smoothly navigate between different views of a scene captured using standard cameras. The underlying automatic 3D reconstruction methods work well for buildings and regular structures but often fail on vegetation, vehicles and other complex geometry present in everyday urban scenes. Consequently, missing depth information makes image-based rendering (IBR) for such scenes very challenging. Our goal is to provide plausible free-viewpoint navigation for such datasets. To do this, we introduce a new IBR algorithm that is robust to missing or unreliable geometry, providing plausible novel views even in regions quite far from the input camera positions. We first over-segment the input images, creating superpixels of homogeneous color content which often tends to preserve depth discontinuities. We then introduce a depth-synthesis approach for poorly reconstructed regions based on a graph structure on the over-segmentation and appropriate traversal of the graph. The superpixels augmented with synthesized depth allow us to define a local shape-preserving warp which compensates for inaccurate depth. Our rendering algorithm blends the warped images, and generates plausible image-based novel views for our challenging target scenes. Our results demonstrate novel view synthesis in real time for multiple challenging scenes with significant depth complexity, providing a convincing immersive navigation experience.
© ACM, 2013. This is authors version of the work posted here for personal use. Not for redistribution. The definitive version was published in ACM Transactions on Graphics.
© ACM, 2013. This is authors version of the work posted here for personal use. Not for redistribution. The definitive version was published in ACM Transactions on Graphics.
Images and movies
See also
NEW!!!!! Source Code and Datasets
Full source code and datasets are available at https://gitlab.inria.fr/sibr/projects/spixelwarp.git , as part of the SIBR system. For full documentation see https://sibr.gitlabpages.inria.fr/ .Video
Supplementary results
Extras
- Museum1 dataset (30 MB)
- Museum2 dataset (28 MB)
- Hugo1 dataset (75 MB)
- University dataset (27 MB)
- Slides (283 MB)
- Free viewpoint paths and camera poses (bundler) for Museum1 and YellowHouse (see here )
Errata
Equation 3 in the original paper contained a typo. The correct version can be found on this version on this page.Code
We would be happy to run your datasets.BibTex references
@Article{CDSD13, author = "Chaurasia, Gaurav and Duch\^ene, Sylvain and Sorkine-Hornung, Olga and Drettakis, George", title = "Depth Synthesis and Local Warps for Plausible Image-based Navigation", journal = "ACM Transactions on Graphics", volume = "32", year = "2013", note = "to be presented at SIGGRAPH 2013", keywords = "Image-based rendering, image warp, superpixels, variational warp, wide baseline, multi-view stereo", url = "http://www-sop.inria.fr/reves/Basilic/2013/CDSD13" }