DC Field | Value | Language |
---|---|---|
dc.contributor.author | Cho, In-Young | ko |
dc.contributor.author | Huo, Yuchi | ko |
dc.contributor.author | Yoon, Sung-Eui | ko |
dc.date.accessioned | 2021-08-25T02:30:59Z | - |
dc.date.available | 2021-08-25T02:30:59Z | - |
dc.date.created | 2021-08-24 | - |
dc.date.created | 2021-08-24 | - |
dc.date.issued | 2021-07 | - |
dc.identifier.citation | ACM TRANSACTIONS ON GRAPHICS, v.40, no.4, pp.1 - 14 | - |
dc.identifier.issn | 0730-0301 | - |
dc.identifier.uri | http://hdl.handle.net/10203/287422 | - |
dc.description.abstract | Image-space auxiliary features such as surface normal have significantly contributed to the recent success of Monte Carlo (MC) reconstruction networks. However, path-space features, another essential piece of light propagation, have not yet been sufficiently explored. Due to the curse of dimensionality, information flow between a regression loss and high-dimensional path-space features is sparse, leading to difficult training and inefficient usage of path-space features in a typical reconstruction framework. This paper introduces a contrastive manifold learning framework to utilize path-space features effectively. The proposed framework employs weakly-supervised learning that converts reference pixel colors to dense pseudo labels for light paths. A convolutional path-embedding network then induces a low-dimensional manifold of paths by iteratively clustering intra-class embeddings, while discriminating inter-class embeddings using gradient descent. The proposed framework facilitates path-space exploration of reconstruction networks by extracting low-dimensional yet meaningful embeddings within the features. We apply our framework to the recent image- and sample-space models and demonstrate considerable improvements, especially on the sample space. | - |
dc.language | English | - |
dc.publisher | ASSOC COMPUTING MACHINERY | - |
dc.title | Weakly-Supervised Contrastive Learning in Path Manifold for Monte Carlo Image Reconstruction | - |
dc.type | Article | - |
dc.identifier.wosid | 000674930900005 | - |
dc.identifier.scopusid | 2-s2.0-85109850181 | - |
dc.type.rims | ART | - |
dc.citation.volume | 40 | - |
dc.citation.issue | 4 | - |
dc.citation.beginningpage | 1 | - |
dc.citation.endingpage | 14 | - |
dc.citation.publicationname | ACM TRANSACTIONS ON GRAPHICS | - |
dc.identifier.doi | 10.1145/3450626.3459876 | - |
dc.contributor.localauthor | Yoon, Sung-Eui | - |
dc.contributor.nonIdAuthor | Cho, In-Young | - |
dc.description.isOpenAccess | N | - |
dc.type.journalArticle | Article | - |
dc.subject.keywordAuthor | Monte Carlo image reconstruction | - |
dc.subject.keywordAuthor | contrastive learning | - |
dc.subject.keywordAuthor | weakly-supervised learning | - |
dc.subject.keywordPlus | REPRESENTATION | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.