{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/cnn-ps-cnn-based-photometric-stereo-for","title":"CNN-PS: CNN-based Photometric Stereo for General Non-Convex Surfaces","arxiv_id":"1808.10093","date":"2018-08-30","proceeding":"ECCV 2018 9","authors":["Satoshi Ikehata"],"abstract":"Most conventional photometric stereo algorithms inversely solve a BRDF-based\nimage formation model. However, the actual imaging process is often far more\ncomplex due to the global light transport on the non-convex surfaces. This\npaper presents a photometric stereo network that directly learns relationships\nbetween the photometric stereo input and surface normals of a scene. For\nhandling unordered, arbitrary number of input images, we merge all the input\ndata to the intermediate representation called {\\it observation map} that has a\nfixed shape, is able to be fed into a CNN. To improve both training and\nprediction, we take into account the rotational pseudo-invariance of the\nobservation map that is derived from the isotropic constraint. For training the\nnetwork, we create a synthetic photometric stereo dataset that is generated by\na physics-based renderer, therefore the global light transport is considered.\nOur experimental results on both synthetic and real datasets show that our\nmethod outperforms conventional BRDF-based photometric stereo algorithms\nespecially when scenes are highly non-convex.","url_abs":"http://arxiv.org/abs/1808.10093v1","url_pdf":"http://arxiv.org/pdf/1808.10093v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"cnn-ps-cnn-based-photometric-stereo-for","repo_url":"https://github.com/satoshi-ikehata/cnn-ps","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1808.10093","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}