{"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/hmanet-hybrid-multi-axis-aggregation-network","title":"HMANet: Hybrid Multi-Axis Aggregation Network for Image Super-Resolution","arxiv_id":"2405.05001","date":"2024-05-08","proceeding":null,"authors":["Shu-Chuan Chu","Zhi-chao Dou","Jeng-Shyang Pan","Shaowei Weng","JunBao Li"],"abstract":"Transformer-based methods have demonstrated excellent performance on super-resolution visual tasks, surpassing conventional convolutional neural networks. However, existing work typically restricts self-attention computation to non-overlapping windows to save computational costs. This means that Transformer-based networks can only use input information from a limited spatial range. Therefore, a novel Hybrid Multi-Axis Aggregation network (HMA) is proposed in this paper to exploit feature potential information better. HMA is constructed by stacking Residual Hybrid Transformer Blocks(RHTB) and Grid Attention Blocks(GAB). On the one side, RHTB combines channel attention and self-attention to enhance non-local feature fusion and produce more attractive visual results. Conversely, GAB is used in cross-domain information interaction to jointly model similar features and obtain a larger perceptual field. For the super-resolution task in the training phase, a novel pre-training method is designed to enhance the model representation capabilities further and validate the proposed model's effectiveness through many experiments. The experimental results show that HMA outperforms the state-of-the-art methods on the benchmark dataset. We provide code and models at https://github.com/korouuuuu/HMA.","url_abs":"https://arxiv.org/abs/2405.05001v1","url_pdf":"https://arxiv.org/pdf/2405.05001v1.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":"hmanet-hybrid-multi-axis-aggregation-network","repo_url":"https://github.com/korouuuuu/hma","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"image-super-resolution","task_name":"Image Super-Resolution"}],"methods":[{"method_slug":"absolute-position-encodings","method_name":"Absolute Position Encodings"},{"method_slug":"adam","method_name":"Adam"},{"method_slug":"attention","method_name":"Attention"},{"method_slug":"bpe","method_name":"BPE"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"label-smoothing","method_name":"Label Smoothing"},{"method_slug":"layer-normalization","method_name":"Layer Normalization"},{"method_slug":"linear-layer","method_name":"Linear Layer"},{"method_slug":"multi-head-attention","method_name":"Multi-Head Attention"},{"method_slug":"position-wise-feed-forward-layer","method_name":"Position-Wise Feed-Forward Layer"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"transformer","method_name":"Transformer"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/image-super-resolution-on-bsd100-2x-upscaling","task":"Image Super-Resolution","dataset":"BSD100 - 2x upscaling","model":"HMA†","rank_in_archive_order":4,"of":30,"metrics":{"PSNR":"32.79","SSIM":"0.9071"},"uses_additional_data":true},{"leaderboard":"/sota/image-super-resolution-on-bsd100-3x-upscaling","task":"Image Super-Resolution","dataset":"BSD100 - 3x upscaling","model":"HMA†","rank_in_archive_order":2,"of":21,"metrics":{"PSNR":"29.66","SSIM":"0.8196"},"uses_additional_data":true},{"leaderboard":"/sota/image-super-resolution-on-bsd100-4x-upscaling","task":"Image Super-Resolution","dataset":"BSD100 - 4x upscaling","model":"HMA†","rank_in_archive_order":3,"of":71,"metrics":{"PSNR":"28.13","SSIM":"0.7562"},"uses_additional_data":true},{"leaderboard":"/sota/image-super-resolution-on-manga109-2x","task":"Image Super-Resolution","dataset":"Manga109 - 2x upscaling","model":"HMA†","rank_in_archive_order":3,"of":21,"metrics":{"PSNR":"41.13","SSIM":"0.9836"},"uses_additional_data":true},{"leaderboard":"/sota/image-super-resolution-on-manga109-3x","task":"Image Super-Resolution","dataset":"Manga109 - 3x upscaling","model":"HMA†","rank_in_archive_order":2,"of":17,"metrics":{"PSNR":"36.10","SSIM":"0.9580"},"uses_additional_data":true},{"leaderboard":"/sota/image-super-resolution-on-manga109-4x","task":"Image Super-Resolution","dataset":"Manga109 - 4x upscaling","model":"HMA†","rank_in_archive_order":3,"of":50,"metrics":{"PSNR":"33.19","SSIM":"0.9344"},"uses_additional_data":true},{"leaderboard":"/sota/image-super-resolution-on-set14-2x-upscaling","task":"Image Super-Resolution","dataset":"Set14 - 2x upscaling","model":"HMA†","rank_in_archive_order":2,"of":35,"metrics":{"PSNR":"35.33","SSIM":"0.9297"},"uses_additional_data":true},{"leaderboard":"/sota/image-super-resolution-on-set14-3x-upscaling","task":"Image Super-Resolution","dataset":"Set14 - 3x upscaling","model":"HMA†","rank_in_archive_order":2,"of":24,"metrics":{"PSNR":"31.47","SSIM":"0.8585"},"uses_additional_data":true},{"leaderboard":"/sota/image-super-resolution-on-set14-4x-upscaling","task":"Image Super-Resolution","dataset":"Set14 - 4x upscaling","model":"HMA†","rank_in_archive_order":2,"of":104,"metrics":{"PSNR":"29.51","SSIM":"0.8019"},"uses_additional_data":true},{"leaderboard":"/sota/image-super-resolution-on-set5-2x-upscaling","task":"Image Super-Resolution","dataset":"Set5 - 2x upscaling","model":"HMA†","rank_in_archive_order":2,"of":41,"metrics":{"PSNR":"38.95","SSIM":"0.9649"},"uses_additional_data":true},{"leaderboard":"/sota/image-super-resolution-on-set5-3x-upscaling","task":"Image Super-Resolution","dataset":"Set5 - 3x upscaling","model":"HMA†","rank_in_archive_order":1,"of":32,"metrics":{"PSNR":"35.35","SSIM":"0.9347"},"uses_additional_data":true},{"leaderboard":"/sota/image-super-resolution-on-set5-4x-upscaling","task":"Image Super-Resolution","dataset":"Set5 - 4x upscaling","model":"HMA†","rank_in_archive_order":1,"of":12,"metrics":{"PSNR":"33.38","SSIM":"0.9089"},"uses_additional_data":true},{"leaderboard":"/sota/image-super-resolution-on-urban100-2x","task":"Image Super-Resolution","dataset":"Urban100 - 2x upscaling","model":"HMA†","rank_in_archive_order":1,"of":29,"metrics":{"PSNR":"35.24","SSIM":"0.9513"},"uses_additional_data":true},{"leaderboard":"/sota/image-super-resolution-on-urban100-3x","task":"Image Super-Resolution","dataset":"Urban100 - 3x upscaling","model":"HMA†","rank_in_archive_order":2,"of":22,"metrics":{"PSNR":"31.00","SSIM":"0.8984"},"uses_additional_data":true},{"leaderboard":"/sota/image-super-resolution-on-urban100-4x","task":"Image Super-Resolution","dataset":"Urban100 - 4x upscaling","model":"HMA†","rank_in_archive_order":3,"of":65,"metrics":{"PSNR":"28.69","SSIM":"0.8512"},"uses_additional_data":true}],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}