Iterative training of robust k-space interpolation networks for improved image reconstruction with limited scan specific training samples

Abstract

Purpose: To evaluate an iterative learning approach for enhanced performance of Robust Artificial-neural-networks for K-space Interpolation (RAKI), when only a limited amount of training data (auto-calibration signals, ACS) are available for accelerated standard 2D imaging. Methods: In a first step, the RAKI model was optimized for the case of strongly limited training data amount. In the iterative learning approach (termed iterative RAKI), the optimized RAKI model is initially trained using original and augmented ACS obtained from a linear parallel imaging reconstruction. Subsequently, the RAKI convolution filters are refined iteratively using original and augmented ACS extracted from the previous RAKI reconstruction. Evaluation was carried out on 200 retrospectively undersampled in-vivo datasets from the fastMRI neuro database with different contrast settings. Results: For limited training data (18 and 22 ACS lines for R=4 and R=5, respectively), iterative RAKI outperforms standard RAKI by reducing residual artefacts and yields strong noise suppression when compared to standard parallel imaging, underlined by quantitative reconstruction quality metrics. In combination with a phase constraint, further reconstruction improvements can be achieved. Additionally, iterative RAKI shows better performance than both GRAPPA and RAKI in case of pre-scan calibration with varying contrast between training- and undersampled data. Conclusion: The iterative learning approach with RAKI benefits from standard RAKIs well known noise suppression feature but requires less original training data for the accurate reconstruction of standard 2D images thereby improving net acceleration.