Complexities of deep learning-based undersampled MR image reconstruction

European Radiology Experimental

Narrative Review

Constant Richard Noordman¹  Joeran Bosma¹  Henkjan Huisman²  Frank Frederikus Jacobus Simonis³  Derya Yakar⁴ 

[1] Diagnostic Image Analysis Group, Department of Medical Imaging, Radboud University Medical Center, 6525 GA, Nijmegen, The Netherlands;Diagnostic Image Analysis Group, Department of Medical Imaging, Radboud University Medical Center, 6525 GA, Nijmegen, The Netherlands;Department of Circulation and Medical Imaging, Norwegian University of Science and Technology, 7030, Trondheim, Norway;Magnetic Detection and Imaging Group, Technical Medical Centre, University of Twente, 7522 NB, Enschede, The Netherlands;Medical Imaging Center, Departments of Radiology, Nuclear Medicine and Molecular Imaging, University Medical Center Groningen, 9700 RB, Groningen, The Netherlands;

关键词: Algorithm;    Artificial intelligence;    Deep learning;    Image processing (computer-assisted);    Magnetic resonance imaging;   

DOI  :  10.1186/s41747-023-00372-7

 received in 2023-04-12, accepted in 2023-08-01,  发布年份 2023

来源: Springer

【 摘 要 】

Artificial intelligence has opened a new path of innovation in magnetic resonance (MR) image reconstruction of undersampled k-space acquisitions. This review offers readers an analysis of the current deep learning-based MR image reconstruction methods. The literature in this field shows exponential growth, both in volume and complexity, as the capabilities of machine learning in solving inverse problems such as image reconstruction are explored. We review the latest developments, aiming to assist researchers and radiologists who are developing new methods or seeking to provide valuable feedback. We shed light on key concepts by exploring the technical intricacies of MR image reconstruction, highlighting the importance of raw datasets and the difficulty of evaluating diagnostic value using standard metrics.Relevance statement Increasingly complex algorithms output reconstructed images that are difficult to assess for robustness and diagnostic quality, necessitating high-quality datasets and collaboration with radiologists.Key points• Deep learning-based image reconstruction algorithms are increasing both in complexity and performance.• The evaluation of reconstructed images may mistake perceived image quality for diagnostic value.• Collaboration with radiologists is crucial for advancing deep learning technology.Graphical Abstract

【 授权许可】

CC BY   
© European Society of Radiology (ESR) 2023

【 图 表 】

Fig. 3