{"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/dense-transformer-networks","title":"Dense Transformer Networks","arxiv_id":"1705.08881","date":"2017-05-24","proceeding":null,"authors":["Jun Li","Yongjun Chen","Lei Cai","Ian Davidson","Shuiwang Ji"],"abstract":"The key idea of current deep learning methods for dense prediction is to\napply a model on a regular patch centered on each pixel to make pixel-wise\npredictions. These methods are limited in the sense that the patches are\ndetermined by network architecture instead of learned from data. In this work,\nwe propose the dense transformer networks, which can learn the shapes and sizes\nof patches from data. The dense transformer networks employ an encoder-decoder\narchitecture, and a pair of dense transformer modules are inserted into each of\nthe encoder and decoder paths. The novelty of this work is that we provide\ntechnical solutions for learning the shapes and sizes of patches from data and\nefficiently restoring the spatial correspondence required for dense prediction.\nThe proposed dense transformer modules are differentiable, thus the entire\nnetwork can be trained. We apply the proposed networks on natural and\nbiological image segmentation tasks and show superior performance is achieved\nin comparison to baseline methods.","url_abs":"http://arxiv.org/abs/1705.08881v2","url_pdf":"http://arxiv.org/pdf/1705.08881v2.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":"dense-transformer-networks","repo_url":"https://github.com/zhengyang-wang/Unet_3D","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"decoder","task_name":"Decoder"},{"task_slug":"image-segmentation","task_name":"Image Segmentation"},{"task_slug":"semantic-segmentation","task_name":"Semantic Segmentation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1705.08881","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}