CGIntrinsics: Better Intrinsic Image Decomposition through Physically-Based Rendering
CGIntrinsics: Better Intrinsic Image Decomposition through Physically-Based Rendering
About this item
Full title
Author / Creator
Publisher
Ithaca: Cornell University Library, arXiv.org
Journal title
Language
English
Formats
Publication information
Publisher
Ithaca: Cornell University Library, arXiv.org
Subjects
More information
Scope and Contents
Contents
Intrinsic image decomposition is a challenging, long-standing computer vision problem for which ground truth data is very difficult to acquire. We explore the use of synthetic data for training CNN-based intrinsic image decomposition models, then applying these learned models to real-world images. To that end, we present \ICG, a new, large-scale da...
Alternative Titles
Full title
CGIntrinsics: Better Intrinsic Image Decomposition through Physically-Based Rendering
Authors, Artists and Contributors
Author / Creator
Identifiers
Primary Identifiers
Record Identifier
TN_cdi_proquest_journals_2095190799
Permalink
https://devfeature-collection.sl.nsw.gov.au/record/TN_cdi_proquest_journals_2095190799
Other Identifiers
E-ISSN
2331-8422
