{"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/deep-laplacian-pyramid-networks-for-fast-and","title":"Deep Laplacian Pyramid Networks for Fast and Accurate Super-Resolution","arxiv_id":"1704.03915","date":"2017-04-12","proceeding":"CVPR 2017 7","authors":["Wei-Sheng Lai","Jia-Bin Huang","Narendra Ahuja","Ming-Hsuan Yang"],"abstract":"Convolutional neural networks have recently demonstrated high-quality\nreconstruction for single-image super-resolution. In this paper, we propose the\nLaplacian Pyramid Super-Resolution Network (LapSRN) to progressively\nreconstruct the sub-band residuals of high-resolution images. At each pyramid\nlevel, our model takes coarse-resolution feature maps as input, predicts the\nhigh-frequency residuals, and uses transposed convolutions for upsampling to\nthe finer level. Our method does not require the bicubic interpolation as the\npre-processing step and thus dramatically reduces the computational complexity.\nWe train the proposed LapSRN with deep supervision using a robust Charbonnier\nloss function and achieve high-quality reconstruction. Furthermore, our network\ngenerates multi-scale predictions in one feed-forward pass through the\nprogressive reconstruction, thereby facilitates resource-aware applications.\nExtensive quantitative and qualitative evaluations on benchmark datasets show\nthat the proposed algorithm performs favorably against the state-of-the-art\nmethods in terms of speed and accuracy.","url_abs":"http://arxiv.org/abs/1704.03915v2","url_pdf":"http://arxiv.org/pdf/1704.03915v2.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":"deep-laplacian-pyramid-networks-for-fast-and","repo_url":"https://github.com/nhatsmrt/superres","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"image-super-resolution","task_name":"Image Super-Resolution"},{"task_slug":"super-resolution","task_name":"Super-Resolution"}],"methods":[{"method_slug":"speed","method_name":"SPEED"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/image-super-resolution-on-bsd100-4x-upscaling","task":"Image Super-Resolution","dataset":"BSD100 - 4x upscaling","model":"LapSRN","rank_in_archive_order":46,"of":71,"metrics":{"PSNR":"27.32","SSIM":"0.728"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-set14-4x-upscaling","task":"Image Super-Resolution","dataset":"Set14 - 4x upscaling","model":"LapSR","rank_in_archive_order":77,"of":104,"metrics":{"PSNR":"28.19","SSIM":"0.772"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-urban100-4x","task":"Image Super-Resolution","dataset":"Urban100 - 4x upscaling","model":"LapSRN","rank_in_archive_order":55,"of":65,"metrics":{"PSNR":"25.21","SSIM":"0.756"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1704.03915","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}