Fields of experts for image-based rendering
Image priors for novel view synthesis have traditionally been non-parametric models based on large libraries of image patch exemplars, producing high-quality results but making inference very slow. Recently a parametric framework, called Fields of Experts, has been proposed for image restoration tha...
| Main Authors: | , , , |
|---|---|
| Format: | Conference item |
| Language: | English |
| Published: |
British Machine Vision Association
2006
|
| _version_ | 1817931215037005824 |
|---|---|
| author | Woodford, OJ Reid, ID Fitzgibbon, AW Torr, PHS |
| author_facet | Woodford, OJ Reid, ID Fitzgibbon, AW Torr, PHS |
| author_sort | Woodford, OJ |
| collection | OXFORD |
| description | Image priors for novel view synthesis have traditionally been non-parametric models based on large libraries of image patch exemplars, producing high-quality results but making inference very slow. Recently a parametric framework, called Fields of Experts, has been proposed for image restoration that promises to speed up inference dramatically. In this paper we apply Fields of Experts for the first time to the problem of novel view synthesis, posed as a Markov random field labelling problem with very large cliques. Additionally, we introduce to computer vision for the first time a new optimization algorithm from statistical physics which reaches better minima than the ICM and simulated annealing algorithms to which such large-clique problems have previously been restricted. |
| first_indexed | 2024-12-09T03:18:28Z |
| format | Conference item |
| id | oxford-uuid:66dec7f3-058e-4762-b103-441192467b9a |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-12-09T03:18:28Z |
| publishDate | 2006 |
| publisher | British Machine Vision Association |
| record_format | dspace |
| spelling | oxford-uuid:66dec7f3-058e-4762-b103-441192467b9a2024-11-04T16:09:38ZFields of experts for image-based renderingConference itemhttp://purl.org/coar/resource_type/c_5794uuid:66dec7f3-058e-4762-b103-441192467b9aEnglishSymplectic ElementsBritish Machine Vision Association2006Woodford, OJReid, IDFitzgibbon, AWTorr, PHSImage priors for novel view synthesis have traditionally been non-parametric models based on large libraries of image patch exemplars, producing high-quality results but making inference very slow. Recently a parametric framework, called Fields of Experts, has been proposed for image restoration that promises to speed up inference dramatically. In this paper we apply Fields of Experts for the first time to the problem of novel view synthesis, posed as a Markov random field labelling problem with very large cliques. Additionally, we introduce to computer vision for the first time a new optimization algorithm from statistical physics which reaches better minima than the ICM and simulated annealing algorithms to which such large-clique problems have previously been restricted. |
| spellingShingle | Woodford, OJ Reid, ID Fitzgibbon, AW Torr, PHS Fields of experts for image-based rendering |
| title | Fields of experts for image-based rendering |
| title_full | Fields of experts for image-based rendering |
| title_fullStr | Fields of experts for image-based rendering |
| title_full_unstemmed | Fields of experts for image-based rendering |
| title_short | Fields of experts for image-based rendering |
| title_sort | fields of experts for image based rendering |
| work_keys_str_mv | AT woodfordoj fieldsofexpertsforimagebasedrendering AT reidid fieldsofexpertsforimagebasedrendering AT fitzgibbonaw fieldsofexpertsforimagebasedrendering AT torrphs fieldsofexpertsforimagebasedrendering |