{"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/image-super-resolution-via-rl-csc-when","title":"Image Super-Resolution via RL-CSC: When Residual Learning Meets Convolutional Sparse Coding","arxiv_id":"1812.11950","date":"2018-12-31","proceeding":null,"authors":["Menglei Zhang","Zhou Liu","Lei Yu"],"abstract":"We propose a simple yet effective model for Single Image Super-Resolution\n(SISR), by combining the merits of Residual Learning and Convolutional Sparse\nCoding (RL-CSC). Our model is inspired by the Learned Iterative\nShrinkage-Threshold Algorithm (LISTA). We extend LISTA to its convolutional\nversion and build the main part of our model by strictly following the\nconvolutional form, which improves the network's interpretability.\nSpecifically, the convolutional sparse codings of input feature maps are\nlearned in a recursive manner, and high-frequency information can be recovered\nfrom these CSCs. More importantly, residual learning is applied to alleviate\nthe training difficulty when the network goes deeper. Extensive experiments on\nbenchmark datasets demonstrate the effectiveness of our method. RL-CSC (30\nlayers) outperforms several recent state-of-the-arts, e.g., DRRN (52 layers)\nand MemNet (80 layers) in both accuracy and visual qualities. Codes and more\nresults are available at https://github.com/axzml/RL-CSC.","url_abs":"http://arxiv.org/abs/1812.11950v1","url_pdf":"http://arxiv.org/pdf/1812.11950v1.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":[],"tasks":[{"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/image-super-resolution-on-bsd100-4x-upscaling","task":"Image Super-Resolution","dataset":"BSD100 - 4x upscaling","model":"RL-CSC","rank_in_archive_order":39,"of":71,"metrics":{"PSNR":"27.44","SSIM":"0.7302"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-set14-4x-upscaling","task":"Image Super-Resolution","dataset":"Set14 - 4x upscaling","model":"RL-CSC","rank_in_archive_order":74,"of":104,"metrics":{"PSNR":"28.29","SSIM":"0.7741"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-urban100-4x","task":"Image Super-Resolution","dataset":"Urban100 - 4x upscaling","model":"RL-CSC","rank_in_archive_order":50,"of":65,"metrics":{"PSNR":"25.59","SSIM":"0.7680"},"uses_additional_data":false}],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}