Learning From Oversampling: A Systematic Exploitation of Oversampling to Address Data Scarcity Issues in Deep Learning- Based Magnetic Resonance Image Reconstruction

• Published in: IEEE access Vol. 12; pp. 97621 - 97629
• Main Authors: Kumara Jalata, Ibsa, Khan, Reeshad, Nakarmi, Ukash
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 2024; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Data acquisition; Data models; Datasets; Deep learning; Image acquisition; Image reconstruction; Imaging; Inverse problems; Magnetic resonance imaging; Mathematical models; Over sampling; Oversampling; Sampled data; Sampling methods; Supervised learning; Training data; unrolled network
• ISSN: 2169-3536; 2169-3536
• DOI: 10.1109/ACCESS.2024.3426362

Data acquisitions in Magnetic Resonance Imaging (MRI) are inherently slow due to sequential acquisition protocol. Image reconstruction from under-sampled data is posed as an inverse problem in traditional model-based learning paradigms. Recent data-centric learning frameworks such as deep learning (DL) frameworks are data hungry, and demand a large, labeled training data sets. To address the lack of large training datasets, in MRI reconstructions, researchers approach the problem in two ways: <xref ref-type="disp-formula" rid="deqn1">(1) unsupervised method where the model is trained without the presence of fully sampled data. <xref ref-type="disp-formula" rid="deqn2">(2) using a method that efficiently use the limited dataset for training purpose. In this paper, we first systematically investigate advantages and limitations of current oversampling methods. Then, we also propose a novel oversampling method and a DL framework that systematically exploits the oversampling technique in the learning process as well as also increase the size of training data set. Essentially, we pose the training data oversampling as a one-to-many mapping function and introduce a new loss function based on similarity metric that can be integrated into a DL framework. Our proposed method not only addresses the training data scarcity in MR image reconstruction and improves reconstruction, but also makes the learned model more robust to different under-sampling techniques