Abstract | ||
---|---|---|
Capturing the shape and spatially-varying appearance (SVBRDF) of an object from images is a challenging task that has applications in both computer vision and graphics. Traditional optimization-based approaches often need a large number of images taken from multiple views in a controlled environment. Newer deep learning-based approaches require only a few input images, but the reconstruction quality is not on par with optimization techniques. We propose a novel deep learning architecture with a stage-wise estimation of shape and SVBRDF The earlier predictions guide each estimation, and a joint refinement network later refines both SVBRDF and shape. We follow a practical mobile image capture setting and use unaligned twoshot flash and no flash images as input. Both our two-shot image capture and network inference can run on mobile hardware. We also create a large-scale synthetic training dataset with domain-randomized geometry and realistic materials. Extensive experiments on both synthetic and real-world datasets show that our networks trained on a synthetic dataset can generalize well to real-world images. Comparisons with recent approaches demonstrate the superior performance of the proposed approach. |
Year | DOI | Venue |
---|---|---|
2020 | 10.1109/CVPR42600.2020.00404 | 2020 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR) |
DocType | ISSN | Citations |
Conference | 1063-6919 | 0 |
PageRank | References | Authors |
0.34 | 34 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Mark Boss | 1 | 0 | 0.68 |
Varun Jampani | 2 | 184 | 19.44 |
Kihwan Kim | 3 | 409 | 28.22 |
Hendrik P. A. Lensch | 4 | 1471 | 96.59 |
Jan Kautz | 5 | 3615 | 198.77 |