{"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/fully-convolutional-network-with-multi-step","title":"Fully Convolutional Network with Multi-Step Reinforcement Learning for Image Processing","arxiv_id":"1811.04323","date":"2018-11-10","proceeding":null,"authors":["Ryosuke Furuta","Naoto Inoue","Toshihiko Yamasaki"],"abstract":"This paper tackles a new problem setting: reinforcement learning with\npixel-wise rewards (pixelRL) for image processing. After the introduction of\nthe deep Q-network, deep RL has been achieving great success. However, the\napplications of deep RL for image processing are still limited. Therefore, we\nextend deep RL to pixelRL for various image processing applications. In\npixelRL, each pixel has an agent, and the agent changes the pixel value by\ntaking an action. We also propose an effective learning method for pixelRL that\nsignificantly improves the performance by considering not only the future\nstates of the own pixel but also those of the neighbor pixels. The proposed\nmethod can be applied to some image processing tasks that require pixel-wise\nmanipulations, where deep RL has never been applied. We apply the proposed\nmethod to three image processing tasks: image denoising, image restoration, and\nlocal color enhancement. Our experimental results demonstrate that the proposed\nmethod achieves comparable or better performance, compared with the\nstate-of-the-art methods based on supervised learning.","url_abs":"http://arxiv.org/abs/1811.04323v2","url_pdf":"http://arxiv.org/pdf/1811.04323v2.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":"fully-convolutional-network-with-multi-step","repo_url":"https://github.com/rfuruta/pixelRL","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}}],"tasks":[{"task_slug":"denoising","task_name":"Denoising"},{"task_slug":"image-denoising","task_name":"Image Denoising"},{"task_slug":"image-restoration","task_name":"Image Restoration"},{"task_slug":"local-color-enhancement","task_name":"Local Color Enhancement"},{"task_slug":"reinforcement-learning","task_name":"Reinforcement Learning"},{"task_slug":"reinforcement-learning-1","task_name":"Reinforcement Learning (RL)"},{"task_slug":"reinforcement-learning-2","task_name":"reinforcement-learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1811.04323","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}