{"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/ml-craist-multi-scale-low-high-frequency","title":"ML-CrAIST: Multi-scale Low-high Frequency Information-based Cross black Attention with Image Super-resolving Transformer","arxiv_id":"2408.09940","date":"2024-08-19","proceeding":null,"authors":["Alik Pramanick","Utsav Bheda","Arijit Sur"],"abstract":"Recently, transformers have captured significant interest in the area of single-image super-resolution tasks, demonstrating substantial gains in performance. Current models heavily depend on the network's extensive ability to extract high-level semantic details from images while overlooking the effective utilization of multi-scale image details and intermediate information within the network. Furthermore, it has been observed that high-frequency areas in images present significant complexity for super-resolution compared to low-frequency areas. This work proposes a transformer-based super-resolution architecture called ML-CrAIST that addresses this gap by utilizing low-high frequency information in multiple scales. Unlike most of the previous work (either spatial or channel), we operate spatial and channel self-attention, which concurrently model pixel interaction from both spatial and channel dimensions, exploiting the inherent correlations across spatial and channel axis. Further, we devise a cross-attention block for super-resolution, which explores the correlations between low and high-frequency information. Quantitative and qualitative assessments indicate that our proposed ML-CrAIST surpasses state-of-the-art super-resolution methods (e.g., 0.15 dB gain @Manga109 $\\times$4). Code is available on: https://github.com/Alik033/ML-CrAIST.","url_abs":"https://arxiv.org/abs/2408.09940v1","url_pdf":"https://arxiv.org/pdf/2408.09940v1.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":"ml-craist-multi-scale-low-high-frequency","repo_url":"https://github.com/alik033/ml-craist","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"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-2x-upscaling","task":"Image Super-Resolution","dataset":"2x upscaling","model":"ML-CrAIST","rank_in_archive_order":1,"of":2,"metrics":{"#params (K)":"1259","FLOPs(G)":"165.7"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-2x-upscaling","task":"Image Super-Resolution","dataset":"2x upscaling","model":"ML-CrAIST-Li","rank_in_archive_order":2,"of":2,"metrics":{"#params (K)":"743","FLOPs(G)":"97.2"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-3x-upscaling","task":"Image Super-Resolution","dataset":"3x upscaling","model":"ML-CrAIST","rank_in_archive_order":1,"of":2,"metrics":{"#params (K)":"1268","FLOPs(G)":"84.1"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-3x-upscaling","task":"Image Super-Resolution","dataset":"3x upscaling","model":"ML-CrAIST-Li","rank_in_archive_order":2,"of":2,"metrics":{"#params (K)":"749","FLOPs(G)":"49.6"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-4x-upscaling","task":"Image Super-Resolution","dataset":"4x upscaling","model":"ML-CrAIST","rank_in_archive_order":1,"of":2,"metrics":{"#params (K)":"1280","FLOPs(G)":"42.9"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-4x-upscaling","task":"Image Super-Resolution","dataset":"4x upscaling","model":"ML-CrAIST-Li","rank_in_archive_order":2,"of":2,"metrics":{"#params (K)":"758","FLOPs(G)":"25.5"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-b100-2x-upscaling","task":"Image Super-Resolution","dataset":"B100 - 2x upscaling","model":"ML-CrAIST","rank_in_archive_order":1,"of":2,"metrics":{"SSIM":"0.9022"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-b100-2x-upscaling","task":"Image Super-Resolution","dataset":"B100 - 2x upscaling","model":"ML-CrAIST-Li","rank_in_archive_order":2,"of":2,"metrics":{"PSNR":"32.36","SSIM":"0.902"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-b100-3x-upscaling","task":"Image Super-Resolution","dataset":"B100 - 3x upscaling","model":"ML-CrAIST","rank_in_archive_order":1,"of":2,"metrics":{"SSIM":"0.8111"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-b100-3x-upscaling","task":"Image Super-Resolution","dataset":"B100 - 3x upscaling","model":"ML-CrAIST-Li","rank_in_archive_order":2,"of":2,"metrics":{"PSNR":"29.28","SSIM":"0.8106"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-b100-4x-upscaling","task":"Image Super-Resolution","dataset":"B100 - 4x upscaling","model":"ML-CrAIST","rank_in_archive_order":1,"of":2,"metrics":{"PSNR":"27.78","SSIM":"0.7446"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-b100-4x-upscaling","task":"Image Super-Resolution","dataset":"B100 - 4x upscaling","model":"ML-CrAIST-Li","rank_in_archive_order":2,"of":2,"metrics":{"PSNR":"27.73"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-manga109-2x","task":"Image Super-Resolution","dataset":"Manga109 - 2x upscaling","model":"ML-CrAIST","rank_in_archive_order":16,"of":21,"metrics":{"PSNR":"39.26","SSIM":"0.9786"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-manga109-2x","task":"Image Super-Resolution","dataset":"Manga109 - 2x upscaling","model":"ML-CrAIST-Li","rank_in_archive_order":17,"of":21,"metrics":{"PSNR":"39.23","SSIM":"0.9785"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-manga109-3x","task":"Image Super-Resolution","dataset":"Manga109 - 3x upscaling","model":"ML-CrAIST","rank_in_archive_order":12,"of":17,"metrics":{"PSNR":"34.42","SSIM":"0.9501"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-manga109-3x","task":"Image Super-Resolution","dataset":"Manga109 - 3x upscaling","model":"ML-CrAIST-Li","rank_in_archive_order":13,"of":17,"metrics":{"PSNR":"34.26","SSIM":"0.9492"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-manga109-4x","task":"Image Super-Resolution","dataset":"Manga109 - 4x upscaling","model":"ML-CrAIST","rank_in_archive_order":29,"of":50,"metrics":{"PSNR":"31.17","SSIM":"0.9176"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-manga109-4x","task":"Image Super-Resolution","dataset":"Manga109 - 4x upscaling","model":"ML-CrAIST-Li","rank_in_archive_order":32,"of":50,"metrics":{"PSNR":"31.11","SSIM":"0.9162"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-set14-2x-upscaling","task":"Image Super-Resolution","dataset":"Set14 - 2x upscaling","model":"ML-CrAIST","rank_in_archive_order":21,"of":35,"metrics":{"PSNR":"33.77","SSIM":"0.922"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-set14-2x-upscaling","task":"Image Super-Resolution","dataset":"Set14 - 2x upscaling","model":"ML-CrAIST-Li","rank_in_archive_order":23,"of":35,"metrics":{"PSNR":"33.64","SSIM":"0.9213"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-set14-3x-upscaling","task":"Image Super-Resolution","dataset":"Set14 - 3x upscaling","model":"ML-CrAIST","rank_in_archive_order":15,"of":24,"metrics":{"PSNR":"30.39","SSIM":"0.8488"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-set14-3x-upscaling","task":"Image Super-Resolution","dataset":"Set14 - 3x upscaling","model":"ML-CrAIST-Li","rank_in_archive_order":17,"of":24,"metrics":{"PSNR":"30.23","SSIM":"0.8474"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-set14-4x-upscaling","task":"Image Super-Resolution","dataset":"Set14 - 4x upscaling","model":"ML-CrAIST","rank_in_archive_order":61,"of":104,"metrics":{"PSNR":"28.53","SSIM":"0.7895"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-set14-4x-upscaling","task":"Image Super-Resolution","dataset":"Set14 - 4x upscaling","model":"ML-CrAIST-Li","rank_in_archive_order":68,"of":104,"metrics":{"PSNR":"28.4","SSIM":"0.7863"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-set5-2x-upscaling","task":"Image Super-Resolution","dataset":"Set5 - 2x upscaling","model":"ML-CrAIST","rank_in_archive_order":21,"of":41,"metrics":{"PSNR":"38.19","SSIM":"0.9617"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-set5-2x-upscaling","task":"Image Super-Resolution","dataset":"Set5 - 2x upscaling","model":"ML-CrAIST-Li","rank_in_archive_order":22,"of":41,"metrics":{"PSNR":"38.15","SSIM":"0.9615"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-set5-3x-upscaling","task":"Image Super-Resolution","dataset":"Set5 - 3x upscaling","model":"ML-CrAIST","rank_in_archive_order":17,"of":32,"metrics":{"PSNR":"34.7","SSIM":"0.9302"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-set5-3x-upscaling","task":"Image Super-Resolution","dataset":"Set5 - 3x upscaling","model":"ML-CrAIST-Li","rank_in_archive_order":21,"of":32,"metrics":{"PSNR":"34.58","SSIM":"0.9294"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-set5-4x-upscaling","task":"Image Super-Resolution","dataset":"Set5 - 4x upscaling","model":"ML-CrAIST","rank_in_archive_order":9,"of":12,"metrics":{"PSNR":"32.36","SSIM":"0.8984"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-set5-4x-upscaling","task":"Image Super-Resolution","dataset":"Set5 - 4x upscaling","model":"ML-CrAIST-Li","rank_in_archive_order":10,"of":12,"metrics":{"PSNR":"32.15","SSIM":"0.8962"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-urban100-2x","task":"Image Super-Resolution","dataset":"Urban100 - 2x upscaling","model":"ML-CrAIST","rank_in_archive_order":16,"of":29,"metrics":{"PSNR":"33.04","SSIM":"0.937"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-urban100-2x","task":"Image Super-Resolution","dataset":"Urban100 - 2x upscaling","model":"ML-CrAIST-Li","rank_in_archive_order":17,"of":29,"metrics":{"PSNR":"32.93","SSIM":"0.9361"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-urban100-3x","task":"Image Super-Resolution","dataset":"Urban100 - 3x upscaling","model":"ML-CrAIST","rank_in_archive_order":14,"of":22,"metrics":{"PSNR":"28.89","SSIM":"0.8676"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-urban100-3x","task":"Image Super-Resolution","dataset":"Urban100 - 3x upscaling","model":"ML-CrAIST-Li","rank_in_archive_order":17,"of":22,"metrics":{"PSNR":"28.73","SSIM":"0.8651"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-urban100-4x","task":"Image Super-Resolution","dataset":"Urban100 - 4x upscaling","model":"ML-CrAIST","rank_in_archive_order":30,"of":65,"metrics":{"PSNR":"26.68","SSIM":"0.8057"},"uses_additional_data":false},{"leaderboard":"/sota/image-super-resolution-on-urban100-4x","task":"Image Super-Resolution","dataset":"Urban100 - 4x upscaling","model":"ML-CrAIST-Li","rank_in_archive_order":36,"of":65,"metrics":{"PSNR":"26.53","SSIM":"0.8019"},"uses_additional_data":false}],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}