{"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/deepasl-kinetic-model-incorporated-loss-for","title":"DeepASL: Kinetic Model Incorporated Loss for Denoising Arterial Spin Labeled MRI via Deep Residual Learning","arxiv_id":"1804.02755","date":"2018-04-08","proceeding":null,"authors":["Cagdas Ulas","Giles Tetteh","Stephan Kaczmarz","Christine Preibisch","Bjoern H. Menze"],"abstract":"Arterial spin labeling (ASL) allows to quantify the cerebral blood flow (CBF)\nby magnetic labeling of the arterial blood water. ASL is increasingly used in\nclinical studies due to its noninvasiveness, repeatability and benefits in\nquantification. However, ASL suffers from an inherently low-signal-to-noise\nratio (SNR) requiring repeated measurements of control/spin-labeled (C/L) pairs\nto achieve a reasonable image quality, which in return increases motion\nsensitivity. This leads to clinically prolonged scanning times increasing the\nrisk of motion artifacts. Thus, there is an immense need of advanced imaging\nand processing techniques in ASL. In this paper, we propose a novel deep\nlearning based approach to improve the perfusion-weighted image quality\nobtained from a subset of all available pairwise C/L subtractions.\nSpecifically, we train a deep fully convolutional network (FCN) to learn a\nmapping from noisy perfusion-weighted image and its subtraction (residual) from\nthe clean image. Additionally, we incorporate the CBF estimation model in the\nloss function during training, which enables the network to produce high\nquality images while simultaneously enforcing the CBF estimates to be as close\nas reference CBF values. Extensive experiments on synthetic and clinical ASL\ndatasets demonstrate the effectiveness of our method in terms of improved ASL\nimage quality, accurate CBF parameter estimation and considerably small\ncomputation time during testing.","url_abs":"http://arxiv.org/abs/1804.02755v2","url_pdf":"http://arxiv.org/pdf/1804.02755v2.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":"deepasl-kinetic-model-incorporated-loss-for","repo_url":"https://github.com/cagdasulas/ASL_CNN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"denoising","task_name":"Denoising"},{"task_slug":"parameter-estimation","task_name":"parameter estimation"}],"methods":[],"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}