Wavelet improved GAN for MRI reconstruction

Wavelet improved GAN for MRI reconstruction. Chen, Y., Firmin, D., & Yang, G. 2021.

Background: Compressed sensing magnetic resonance imaging (CS-MRI) is an important technique of accel- erating the acquisition process of magnetic resonance (MR) images by undersampling. It has the potential of reducing MR scanning time and costs, thus minimising patient discomfort. Motivation: One of the successful CS-MRI techniques to recover the original image from undersampled images is generative adversarial network (GAN). However, GAN-based techniques suffer from three key limitations: training instability, slow convergence and input size constraints. Method and Result: In this study, we propose a novel GAN-based CS-MRI technique: WPD-DAGAN (Wavelet Packet Decomposition Improved de-aliaising GAN). We incorporate Wasserstein loss function and a novel structure based on wavelet packet decomposition (WPD) into the de-aliaising GAN (DAGAN) architecture, which is a well established GAN-based CS-MRI technique. We show that the proposed network architecture achieves a significant performance improvement over the state-of-the-art CS-MRI techniques.

@misc{
 title = {Wavelet improved GAN for MRI reconstruction},
 type = {misc},
 year = {2021},
 source = {SPIE Medical Imaging 2021},
 keywords = {compressed sensing,generative adversarial network,mri,wavelet packet decomposition},
 pages = {37},
 volume = {1159513},
 issue = {February},
 id = {4486d93b-5a8f-3731-9017-3114f2877841},
 created = {2024-01-13T06:15:54.120Z},
 file_attached = {true},
 profile_id = {d5b9f595-4104-33d2-b640-e84dee1b7ded},
 last_modified = {2025-02-21T12:00:33.307Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {true},
 hidden = {false},
 folder_uuids = {ded4b384-1a9c-491d-88dd-c5e6a88d479d},
 private_publication = {false},
 abstract = {Background: Compressed sensing magnetic resonance imaging (CS-MRI) is an important technique of accel- erating the acquisition process of magnetic resonance (MR) images by undersampling. It has the potential of reducing MR scanning time and costs, thus minimising patient discomfort. Motivation: One of the successful CS-MRI techniques to recover the original image from undersampled images is generative adversarial network (GAN). However, GAN-based techniques suffer from three key limitations: training instability, slow convergence and input size constraints. Method and Result: In this study, we propose a novel GAN-based CS-MRI technique: WPD-DAGAN (Wavelet Packet Decomposition Improved de-aliaising GAN). We incorporate Wasserstein loss function and a novel structure based on wavelet packet decomposition (WPD) into the de-aliaising GAN (DAGAN) architecture, which is a well established GAN-based CS-MRI technique. We show that the proposed network architecture achieves a significant performance improvement over the state-of-the-art CS-MRI techniques.},
 bibtype = {misc},
 author = {Chen, Yutong and Firmin, David and Yang, Guang},
 doi = {10.1117/12.2581004}
}

Downloads: 0

{"_id":"89rFbepyPkuNLExzj","bibbaseid":"chen-firmin-yang-waveletimprovedganformrireconstruction-2021","author_short":["Chen, Y.","Firmin, D.","Yang, G."],"bibdata":{"title":"Wavelet improved GAN for MRI reconstruction","type":"misc","year":"2021","source":"SPIE Medical Imaging 2021","keywords":"compressed sensing,generative adversarial network,mri,wavelet packet decomposition","pages":"37","volume":"1159513","issue":"February","id":"4486d93b-5a8f-3731-9017-3114f2877841","created":"2024-01-13T06:15:54.120Z","file_attached":"true","profile_id":"d5b9f595-4104-33d2-b640-e84dee1b7ded","last_modified":"2025-02-21T12:00:33.307Z","read":false,"starred":false,"authored":"true","confirmed":"true","hidden":false,"folder_uuids":"ded4b384-1a9c-491d-88dd-c5e6a88d479d","private_publication":false,"abstract":"Background: Compressed sensing magnetic resonance imaging (CS-MRI) is an important technique of accel- erating the acquisition process of magnetic resonance (MR) images by undersampling. It has the potential of reducing MR scanning time and costs, thus minimising patient discomfort. Motivation: One of the successful CS-MRI techniques to recover the original image from undersampled images is generative adversarial network (GAN). However, GAN-based techniques suffer from three key limitations: training instability, slow convergence and input size constraints. Method and Result: In this study, we propose a novel GAN-based CS-MRI technique: WPD-DAGAN (Wavelet Packet Decomposition Improved de-aliaising GAN). We incorporate Wasserstein loss function and a novel structure based on wavelet packet decomposition (WPD) into the de-aliaising GAN (DAGAN) architecture, which is a well established GAN-based CS-MRI technique. We show that the proposed network architecture achieves a significant performance improvement over the state-of-the-art CS-MRI techniques.","bibtype":"misc","author":"Chen, Yutong and Firmin, David and Yang, Guang","doi":"10.1117/12.2581004","bibtex":"@misc{\n title = {Wavelet improved GAN for MRI reconstruction},\n type = {misc},\n year = {2021},\n source = {SPIE Medical Imaging 2021},\n keywords = {compressed sensing,generative adversarial network,mri,wavelet packet decomposition},\n pages = {37},\n volume = {1159513},\n issue = {February},\n id = {4486d93b-5a8f-3731-9017-3114f2877841},\n created = {2024-01-13T06:15:54.120Z},\n file_attached = {true},\n profile_id = {d5b9f595-4104-33d2-b640-e84dee1b7ded},\n last_modified = {2025-02-21T12:00:33.307Z},\n read = {false},\n starred = {false},\n authored = {true},\n confirmed = {true},\n hidden = {false},\n folder_uuids = {ded4b384-1a9c-491d-88dd-c5e6a88d479d},\n private_publication = {false},\n abstract = {Background: Compressed sensing magnetic resonance imaging (CS-MRI) is an important technique of accel- erating the acquisition process of magnetic resonance (MR) images by undersampling. It has the potential of reducing MR scanning time and costs, thus minimising patient discomfort. Motivation: One of the successful CS-MRI techniques to recover the original image from undersampled images is generative adversarial network (GAN). However, GAN-based techniques suffer from three key limitations: training instability, slow convergence and input size constraints. Method and Result: In this study, we propose a novel GAN-based CS-MRI technique: WPD-DAGAN (Wavelet Packet Decomposition Improved de-aliaising GAN). We incorporate Wasserstein loss function and a novel structure based on wavelet packet decomposition (WPD) into the de-aliaising GAN (DAGAN) architecture, which is a well established GAN-based CS-MRI technique. We show that the proposed network architecture achieves a significant performance improvement over the state-of-the-art CS-MRI techniques.},\n bibtype = {misc},\n author = {Chen, Yutong and Firmin, David and Yang, Guang},\n doi = {10.1117/12.2581004}\n}","author_short":["Chen, Y.","Firmin, D.","Yang, G."],"urls":{"Paper":"https://bibbase.org/service/mendeley/d5b9f595-4104-33d2-b640-e84dee1b7ded/file/59a3c90e-820d-c73d-681f-cf6306dd1b37/1159513.pdf.pdf"},"biburl":"https://bibbase.org/service/mendeley/d5b9f595-4104-33d2-b640-e84dee1b7ded","bibbaseid":"chen-firmin-yang-waveletimprovedganformrireconstruction-2021","role":"author","keyword":["compressed sensing","generative adversarial network","mri","wavelet packet decomposition"],"metadata":{"authorlinks":{}},"downloads":0},"bibtype":"misc","biburl":"https://bibbase.org/service/mendeley/d5b9f595-4104-33d2-b640-e84dee1b7ded","dataSources":["2252seNhipfTmjEBQ"],"keywords":["compressed sensing","generative adversarial network","mri","wavelet packet decomposition"],"search_terms":["wavelet","improved","gan","mri","reconstruction","chen","firmin","yang"],"title":"Wavelet improved GAN for MRI reconstruction","year":2021}