{"id":1254,"date":"2024-05-07T16:35:40","date_gmt":"2024-05-07T16:35:40","guid":{"rendered":"https:\/\/cchsu.info\/wordpress\/?p=1254"},"modified":"2024-05-07T16:35:40","modified_gmt":"2024-05-07T16:35:40","slug":"award-we-received-top-performance-certificate-in-cvpr-ntire2024-image-super-resolution-track-6-199top-3","status":"publish","type":"post","link":"https:\/\/cchsu.info\/wordpress\/2024\/05\/07\/award-we-received-top-performance-certificate-in-cvpr-ntire2024-image-super-resolution-track-6-199top-3\/","title":{"rendered":"[Award] We received top performance certificate in CVPR NTIRE2024 Image super-resolution track (6\/199=Top 3%)"},"content":{"rendered":"
\n
\u606d\u559c \u674e\u5bb6\u9298<\/span><\/a><\/span> \u5468\u4f9d\u7487<\/span><\/a><\/span> \u53c3\u52a0 CVPR NTIRE2024 Image super-resolution \u7af6\u8cfd\u6210\u679c\u6392\u540d\u5168\u7403 Top 3% (6\/199)\uff0c\u50c5\u6b21\u65bc\u90a3\u4e9b\u5927\u578b\u7814\u7a76\u5718\u968a\u8207\u516c\u53f8\u3002\u66f4\u91cd\u8981\u7684\u662f\uff0c\u6211\u5011\u6b64\u6b21\u7af6\u8cfd\u7684\u7814\u7a76\u8ad6\u6587\u540c\u6a23\u4e5f\u88ab NTIRE \u7372\u5f97\u63a5\u53d7\u3002NTIRE\u662f\u6bcf\u5e74\u90fd\u6703\u5728CVPR\u4e0a\u8209\u8fa6\u7684\u77e5\u540d Workshop\uff0c\u65e8\u5728\u589e\u52a0\u7814\u7a76\u4eba\u54e1\u5c0d\u5f71\u50cf\u5fa9\u539f\u3001\u589e\u5f37\u548c\u4fee\u88dc\u7b49\u95dc\u9375\u96fb\u8166\u8996\u89ba\u4efb\u52d9\u7684\u5ee3\u6cdb\u7406\u89e3\u53ca\u8a0e\u8ad6\u3002\u5177\u9ad4\u4f86\u8aaa\uff0c\u662f\u95dc\u65bc\u5982\u4f55\u6062\u5fa9\u9000\u5316\u7684\u5f71\u50cf\u5167\u5bb9\u4e26\u586b\u5145\u7f3a\u5931\u7684\u4fe1\u606f\uff0c\u6216\u5be6\u73fe\u6240\u9700\u76ee\u6a19\uff08\u589e\u5f37\u95dc\u65bc\u611f\u77e5\u54c1\u8cea\u3001\u5167\u5bb9\u3001\u6216\u8655\u7406\u6b64\u985e\u5f71\u50cf\u7684\u61c9\u7528\u7a0b\u5f0f\u7684\u6027\u80fd\uff09\u3002<\/div>\n<\/div>\n
\n
\u8fd1\u5e74\u4f86\uff0c\u5927\u5bb6\u591a\u900f\u904e Swin-Transformer \u70ba\u57fa\u790e\u6a21\u578b\u53bb\u5617\u8a66\u589e\u5927\u6a21\u578bReceptive field \u6216\u8a2d\u8a08\u8907\u96dc\u7684\u6a21\u584a\u4f86\u589e\u5f37\u6a21\u578b\u8403\u53d6\u7279\u5fb5\u7684\u80fd\u529b\u3002\u7136\u800c\uff0c\u6211\u5011\u89c0\u5bdf\u51fa\u4e86\u4e00\u500b\u666e\u904d\u7684\u73fe\u8c61\uff0c\u5c31\u662f\u7279\u5fb5\u5716\u7684\u5f37\u5ea6\u5206\u5e03\u5f80\u5f80\u6703\u96a8\u8457\u7db2\u8def\u6df1\u5ea6\u589e\u52a0\u800c\u8b8a\u5927\uff0c\u4f46\u5728\u6700\u672b\u7aef\u7684\u6642\u5019\u88ab\u6291\u5236\u5230\u4e00\u500b\u8f03\u5c0f\u7684\u5340\u9593\uff0c\u9019\u53ef\u80fd\u96b1\u5f0f\u5730\u9650\u5236\u4e86\u6a21\u578b\u7684\u6548\u80fd\u4e0a\u754c\uff0c\u56e0\u70ba\u7a7a\u9593\u8a0a\u606f\u6703\u56e0\u70ba\u906d\u9022\u8cc7\u8a0a\u74f6\u9838 (Information bottleneck) \u800c\u640d\u5931\u751a\u81f3\u5e7e\u4e4e\u4e0d\u898b\uff0c\u800c\u5c0e\u81f4\u5927\u5bb6\u9700\u8981\u8a2d\u8a08\u975e\u5e38\u8907\u96dc\u7684\u7db2\u8def\u6216\u6a21\u584a\u624d\u80fd\u628a SR \u6548\u80fd\u518d\u5f80\u4e0a\u63d0\u5347\u3002<\/div>\n
\u6211\u5011\u7814\u7a76\u8ddf\u7af6\u8cfd\u5f80\u5f80\u90fd\u4e3b\u6253\u4e00\u500b\u91cd\u9ede\uff1a\u5be6\u52d9\u8207\u7406\u8ad6\u4f75\u91cd\u3002\u6211\u5011\u63d0\u51fa\u4e86 Dense-residual-connected Transformer (DRCT) \uff0c\u900f\u904e\u7d50\u5408dense-connection\uff0c\u4f86\u7a69\u5b9a\u524d\u5411\u904e\u7a0b\u800c\u4e0d\u7528\u589e\u52a0\u984d\u5916\u6a21\u7d44\uff0c\u8b93SR\u6a21\u578b\u80fd\u5920\u201d\u9060\u96e2\u201d\u8cc7\u8a0a\u74f6\u9838\uff0c\u76f8\u6bd4\u65bc\u73fe\u4ecaSOTA\u6a21\u578b\u7701\u4e8633%\u7684\u53c3\u6578\u91cf\u5c31\u80fd\u5728\u6548\u80fd\u4e0a\u6709\u986f\u8457\u63d0\u5347\uff0c\u9019\u6a23\u7684\u8a2d\u8a08\u5f88\u7b26\u5408\u6211\u5011\u5fc3\u76ee\u4e2d\u7684\u6982\u5ff5\uff0c\u6709\u6548\u3001\u4f46\u4e0d\u8907\u96dc\u3002<\/div>\n<\/div>\n
\n
[Code] https:\/\/github.com\/ming053l\/DRCT<\/a><\/span><\/div>\n
[Paper] https:\/\/arxiv.org\/abs\/2404.00722<\/a><\/span><\/div>\n
[Project Page] https:\/\/allproj002.github.io\/drct.github.io\/<\/a><\/span><\/div>\n<\/div>\n
\n
============<\/div>\n
Congratulations to Chia-Ming Li and Yi-Hsuan Nelly Chou for achieving the Top 3% Performance Award (6\/199) in the CVPR NTIRE2024 Image Super-resolution Competition!<\/div>\n<\/div>\n
\n
This is a remarkable achievement, especially considering that they were competing against large research teams and companies. Moreover, their research paper on this competition has also been accepted by NTIRE. NTIRE is a well-known workshop held annually at CVPR that aims to promote a broader understanding and discussion among researchers on key computer vision tasks such as image restoration, enhancement, and editing. Specifically, it focuses on how to recover degraded image content and fill in missing information, or achieve desired goals (enhance the performance of applications related to perceptual quality, content, or processing such images).<\/div>\n<\/div>\n
\n
In recent years, many have attempted to increase the model’s receptive field or design complex modules to enhance the model’s feature extraction capability by using Swin-Transformer as the base model. However, they have observed a common phenomenon: the intensity distribution of feature maps tends to increase with network depth, but is suppressed to a smaller range at the end stage. This may implicitly limit the upper bound of the model’s performance, as spatial information may be lost or even almost invisible due to the information bottleneck, resulting in the need to design very complex networks or modules to further improve SR performance.<\/div>\n<\/div>\n
\n
Their research and competition have always focused on one key point: the equal importance of practice and theory.<\/div>\n<\/div>\n
\n
They have proposed Dense-residual-connected Transformer (DRCT), which combines dense-connection to stabilize the forward process without adding additional modules, allowing the SR model to “stay away” from the information bottleneck. Compared to current SOTA models, it can significantly improve performance while saving 33% of the number of parameters. This design is in line with their concept of being effective but not complex.<\/div>\n<\/div>\n","protected":false},"excerpt":{"rendered":"

\u606d\u559c \u674e\u5bb6\u9298 \u5468\u4f9d\u7487 \u53c3\u52a0 CVPR NTIRE2024 Image super-resolution \u7af6\u8cfd\u6210 […]<\/p>\n","protected":false},"author":2,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[2],"tags":[],"class_list":["post-1254","post","type-post","status-publish","format-standard","hentry","category-award"],"_links":{"self":[{"href":"https:\/\/cchsu.info\/wordpress\/wp-json\/wp\/v2\/posts\/1254","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/cchsu.info\/wordpress\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/cchsu.info\/wordpress\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/cchsu.info\/wordpress\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/cchsu.info\/wordpress\/wp-json\/wp\/v2\/comments?post=1254"}],"version-history":[{"count":0,"href":"https:\/\/cchsu.info\/wordpress\/wp-json\/wp\/v2\/posts\/1254\/revisions"}],"wp:attachment":[{"href":"https:\/\/cchsu.info\/wordpress\/wp-json\/wp\/v2\/media?parent=1254"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/cchsu.info\/wordpress\/wp-json\/wp\/v2\/categories?post=1254"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/cchsu.info\/wordpress\/wp-json\/wp\/v2\/tags?post=1254"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}