{"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/end-to-end-learning-of-brain-tissue","title":"End-to-end learning of brain tissue segmentation from imperfect labeling","arxiv_id":"1612.00940","date":"2016-12-03","proceeding":null,"authors":["Alex Fedorov","Jeremy Johnson","Eswar Damaraju","Alexei Ozerin","Vince Calhoun","Sergey Plis"],"abstract":"Segmenting a structural magnetic resonance imaging (MRI) scan is an important\npre-processing step for analytic procedures and subsequent inferences about\nlongitudinal tissue changes. Manual segmentation defines the current gold\nstandard in quality but is prohibitively expensive. Automatic approaches are\ncomputationally intensive, incredibly slow at scale, and error prone due to\nusually involving many potentially faulty intermediate steps. In order to\nstreamline the segmentation, we introduce a deep learning model that is based\non volumetric dilated convolutions, subsequently reducing both processing time\nand errors. Compared to its competitors, the model has a reduced set of\nparameters and thus is easier to train and much faster to execute. The contrast\nin performance between the dilated network and its competitors becomes obvious\nwhen both are tested on a large dataset of unprocessed human brain volumes. The\ndilated network consistently outperforms not only another state-of-the-art deep\nlearning approach, the up convolutional network, but also the ground truth on\nwhich it was trained. Not only can the incredible speed of our model make large\nscale analyses much easier but we also believe it has great potential in a\nclinical setting where, with little to no substantial delay, a patient and\nprovider can go over test results.","url_abs":"http://arxiv.org/abs/1612.00940v2","url_pdf":"http://arxiv.org/pdf/1612.00940v2.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":"end-to-end-learning-of-brain-tissue","repo_url":"https://github.com/Entodi/meshnet-pytorch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"end-to-end-learning-of-brain-tissue","repo_url":"https://github.com/neuroneural/brainchop","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[],"methods":[{"method_slug":"speed","method_name":"SPEED"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}