DC Field | Value | Language |
---|---|---|
dc.contributor.author | Choi, Inchang | ko |
dc.contributor.author | Orazio, Gallo | ko |
dc.contributor.author | Alejandro, Troccoli | ko |
dc.contributor.author | Kim, Min Hyuk | ko |
dc.contributor.author | Kautz, Jan | ko |
dc.date.accessioned | 2019-12-13T07:38:49Z | - |
dc.date.available | 2019-12-13T07:38:49Z | - |
dc.date.created | 2019-11-11 | - |
dc.date.created | 2019-11-11 | - |
dc.date.created | 2019-11-11 | - |
dc.date.issued | 2019-11-01 | - |
dc.identifier.citation | IEEE/CVF International Conference on Computer Vision (ICCV), pp.7780 - 7789 | - |
dc.identifier.issn | 1550-5499 | - |
dc.identifier.uri | http://hdl.handle.net/10203/269011 | - |
dc.description.abstract | We present Extreme View Synthesis, a solution for novel view extrapolation that works even when the number of input images is small-as few as two. In this context, occlusions and depth uncertainty are two of the most pressing issues, and worsen as the degree of extrapolation increases. We follow the traditional paradigm of petforming depth-based warping and refinement, with a few key improvements. First,we estimate a depth probability volume, rather than just a single depth value for each pixel of the novel view. This allows us to leverage depth uncertainty in challenging regions, such as depth discontinuities. After using it to get an initial estimate of the novel view, we explicitly combine learned image priors and the depth uncertainty to synthesize a refined image with less artfacts. Our method is the first to show visually pleasing results for baseline magnifications of up to 30x. The code is available at htLtps: //github.comiNVlabs/extreme-view-synth | - |
dc.language | English | - |
dc.publisher | IEEE | - |
dc.title | Extreme View Synthesis | - |
dc.type | Conference | - |
dc.identifier.wosid | 000548549202087 | - |
dc.identifier.scopusid | 2-s2.0-85081929304 | - |
dc.type.rims | CONF | - |
dc.citation.beginningpage | 7780 | - |
dc.citation.endingpage | 7789 | - |
dc.citation.publicationname | IEEE/CVF International Conference on Computer Vision (ICCV) | - |
dc.identifier.conferencecountry | KO | - |
dc.identifier.conferencelocation | COEX Convention Center | - |
dc.identifier.doi | 10.1109/ICCV.2019.00787 | - |
dc.contributor.localauthor | Kim, Min Hyuk | - |
dc.contributor.nonIdAuthor | Orazio, Gallo | - |
dc.contributor.nonIdAuthor | Alejandro, Troccoli | - |
dc.contributor.nonIdAuthor | Kautz, Jan | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.