Jon André Ottesen1,2, Matthan W A Caan3,4, Inge Rasmus Groote3,5, Atle Bjørnerud3,6. 1. Computational Radiology & Artificial Intelligence (CRAI) Unit, Division of Radiology and Nuclear Medicine, Oslo University Hospital, Oslo, Norway. jon.a.ottesen@gmail.com. 2. Department of Physics, Faculty of Mathematics and Natural Sciences, University of Oslo, Oslo, Norway. jon.a.ottesen@gmail.com. 3. Computational Radiology & Artificial Intelligence (CRAI) Unit, Division of Radiology and Nuclear Medicine, Oslo University Hospital, Oslo, Norway. 4. Amsterdam UMC, Biomedical Engineering and Physics, University of Amsterdam, Amsterdam, Netherlands. 5. Department of Radiology, Vestfold Hospital Trust, Tønsberg, Norway. 6. Department of Physics, Faculty of Mathematics and Natural Sciences, University of Oslo, Oslo, Norway.
Abstract
OBJECTIVE: To improve accelerated MRI reconstruction through a densely connected cascading deep learning reconstruction framework. MATERIALS AND METHODS: A cascading deep learning reconstruction framework (reference model) was modified by applying three architectural modifications: input-level dense connections between cascade inputs and outputs, an improved deep learning sub-network, and long-range skip-connections between subsequent deep learning networks. An ablation study was performed, where five model configurations were trained on the NYU fastMRI neuro dataset with an end-to-end scheme conjunct on four- and eightfold acceleration. The trained models were evaluated by comparing their respective structural similarity index measure (SSIM), normalized mean square error (NMSE), and peak signal to noise ratio (PSNR). RESULTS: The proposed densely interconnected residual cascading network (DIRCN), utilizing all three suggested modifications achieved a SSIM improvement of 8% and 11%, a NMSE improvement of 14% and 23%, and a PSNR improvement of 2% and 3% for four- and eightfold acceleration, respectively. In an ablation study, the individual architectural modifications all contributed to this improvement for both acceleration factors, by improving the SSIM, NMSE, and PSNR with approximately 2-4%, 4-9%, and 0.5-1%, respectively. CONCLUSION: The proposed architectural modifications allow for simple adjustments on an already existing cascading framework to further improve the resulting reconstructions.
OBJECTIVE: To improve accelerated MRI reconstruction through a densely connected cascading deep learning reconstruction framework. MATERIALS AND METHODS: A cascading deep learning reconstruction framework (reference model) was modified by applying three architectural modifications: input-level dense connections between cascade inputs and outputs, an improved deep learning sub-network, and long-range skip-connections between subsequent deep learning networks. An ablation study was performed, where five model configurations were trained on the NYU fastMRI neuro dataset with an end-to-end scheme conjunct on four- and eightfold acceleration. The trained models were evaluated by comparing their respective structural similarity index measure (SSIM), normalized mean square error (NMSE), and peak signal to noise ratio (PSNR). RESULTS: The proposed densely interconnected residual cascading network (DIRCN), utilizing all three suggested modifications achieved a SSIM improvement of 8% and 11%, a NMSE improvement of 14% and 23%, and a PSNR improvement of 2% and 3% for four- and eightfold acceleration, respectively. In an ablation study, the individual architectural modifications all contributed to this improvement for both acceleration factors, by improving the SSIM, NMSE, and PSNR with approximately 2-4%, 4-9%, and 0.5-1%, respectively. CONCLUSION: The proposed architectural modifications allow for simple adjustments on an already existing cascading framework to further improve the resulting reconstructions.
Authors: Yilmaz Korkmaz; Salman U H Dar; Mahmut Yurt; Muzaffer Ozbey; Tolga Cukur Journal: IEEE Trans Med Imaging Date: 2022-06-30 Impact factor: 11.037