{"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/beyond-deep-residual-learning-for-image","title":"Beyond Deep Residual Learning for Image Restoration: Persistent Homology-Guided Manifold Simplification","arxiv_id":"1611.06345","date":"2016-11-19","proceeding":null,"authors":["Woong Bae","Jaejun Yoo","Jong Chul Ye"],"abstract":"The latest deep learning approaches perform better than the state-of-the-art\nsignal processing approaches in various image restoration tasks. However, if an\nimage contains many patterns and structures, the performance of these CNNs is\nstill inferior. To address this issue, here we propose a novel feature space\ndeep residual learning algorithm that outperforms the existing residual\nlearning. The main idea is originated from the observation that the performance\nof a learning algorithm can be improved if the input and/or label manifolds can\nbe made topologically simpler by an analytic mapping to a feature space. Our\nextensive numerical studies using denoising experiments and NTIRE single-image\nsuper-resolution (SISR) competition demonstrate that the proposed feature space\nresidual learning outperforms the existing state-of-the-art approaches.\nMoreover, our algorithm was ranked third in NTIRE competition with 5-10 times\nfaster computational time compared to the top ranked teams. The source code is\navailable on page : https://github.com/iorism/CNN.git","url_abs":"http://arxiv.org/abs/1611.06345v4","url_pdf":"http://arxiv.org/pdf/1611.06345v4.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":"beyond-deep-residual-learning-for-image","repo_url":"https://github.com/iorism/CNN","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"color-image-denoising","task_name":"Color Image Denoising"},{"task_slug":"denoising","task_name":"Denoising"},{"task_slug":"image-restoration","task_name":"Image Restoration"},{"task_slug":"image-super-resolution","task_name":"Image Super-Resolution"},{"task_slug":"super-resolution","task_name":"Super-Resolution"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/color-image-denoising-on-cbsd68-sigma50","task":"Color Image Denoising","dataset":"CBSD68 sigma50","model":"DnCNN","rank_in_archive_order":10,"of":18,"metrics":{"PSNR":"28.01"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-bsd100-4x-upscaling","task":"Image Super-Resolution","dataset":"BSD100 - 4x upscaling","model":"Manifold Simplification","rank_in_archive_order":29,"of":71,"metrics":{"PSNR":"27.66","SSIM":"0.7380"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-set14-4x-upscaling","task":"Image Super-Resolution","dataset":"Set14 - 4x upscaling","model":"Manifold Simplification","rank_in_archive_order":51,"of":104,"metrics":{"PSNR":"28.80","SSIM":"0.7856"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-urban100-4x","task":"Image Super-Resolution","dataset":"Urban100 - 4x upscaling","model":"Manifold Simplification","rank_in_archive_order":39,"of":65,"metrics":{"PSNR":"26.42","SSIM":"0.7940"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/1611.06345","atlas_url":"https://app.syntology.ai/?focus=1611.06345","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}