{"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/one-network-to-solve-them-all-solving-linear","title":"One Network to Solve Them All --- Solving Linear Inverse Problems using Deep Projection Models","arxiv_id":"1703.09912","date":"2017-03-29","proceeding":null,"authors":["J. H. Rick Chang","Chun-Liang Li","Barnabas Poczos","B. V. K. Vijaya Kumar","Aswin C. Sankaranarayanan"],"abstract":"While deep learning methods have achieved state-of-the-art performance in\nmany challenging inverse problems like image inpainting and super-resolution,\nthey invariably involve problem-specific training of the networks. Under this\napproach, different problems require different networks. In scenarios where we\nneed to solve a wide variety of problems, e.g., on a mobile camera, it is\ninefficient and costly to use these specially-trained networks. On the other\nhand, traditional methods using signal priors can be used in all linear inverse\nproblems but often have worse performance on challenging tasks. In this work,\nwe provide a middle ground between the two kinds of methods --- we propose a\ngeneral framework to train a single deep neural network that solves arbitrary\nlinear inverse problems. The proposed network acts as a proximal operator for\nan optimization algorithm and projects non-image signals onto the set of\nnatural images defined by the decision boundary of a classifier. In our\nexperiments, the proposed framework demonstrates superior performance over\ntraditional methods using a wavelet sparsity prior and achieves comparable\nperformance of specially-trained networks on tasks including compressive\nsensing and pixel-wise inpainting.","url_abs":"http://arxiv.org/abs/1703.09912v1","url_pdf":"http://arxiv.org/pdf/1703.09912v1.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":"one-network-to-solve-them-all-solving-linear","repo_url":"https://github.com/image-science-lab/OneNet","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"GPL-3.0"}},{"paper_slug":"one-network-to-solve-them-all-solving-linear","repo_url":"https://github.com/rick-chang/OneNet","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"GPL-3.0"}}],"tasks":[{"task_slug":"all","task_name":"All"},{"task_slug":"compressive-sensing","task_name":"Compressive Sensing"},{"task_slug":"image-inpainting","task_name":"Image Inpainting"},{"task_slug":"super-resolution","task_name":"Super-Resolution"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1703.09912","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}