{"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/inverserendernet-learning-single-image","title":"InverseRenderNet: Learning single image inverse rendering","arxiv_id":"1811.12328","date":"2018-11-29","proceeding":"CVPR 2019 6","authors":["Ye Yu","William A. P. Smith"],"abstract":"We show how to train a fully convolutional neural network to perform inverse\nrendering from a single, uncontrolled image. The network takes an RGB image as\ninput, regresses albedo and normal maps from which we compute lighting\ncoefficients. Our network is trained using large uncontrolled image collections\nwithout ground truth. By incorporating a differentiable renderer, our network\ncan learn from self-supervision. Since the problem is ill-posed we introduce\nadditional supervision: 1. We learn a statistical natural illumination prior,\n2. Our key insight is to perform offline multiview stereo (MVS) on images\ncontaining rich illumination variation. From the MVS pose and depth maps, we\ncan cross project between overlapping views such that Siamese training can be\nused to ensure consistent estimation of photometric invariants. MVS depth also\nprovides direct coarse supervision for normal map estimation. We believe this\nis the first attempt to use MVS supervision for learning inverse rendering.","url_abs":"http://arxiv.org/abs/1811.12328v1","url_pdf":"http://arxiv.org/pdf/1811.12328v1.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":"inverserendernet-learning-single-image","repo_url":"https://github.com/YeeU/InverseRenderNet","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"tf","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"inverse-rendering","task_name":"Inverse Rendering"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1811.12328","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}