An Iowa Institute for Biomedical Imaging (IIBI) research team won best machine-learning paper at the International Symposium on Biomedical Imaging (ISBI 2019) recently in Venice, Italy. ISBI is a scientific conference dedicated to mathematical, algorithmic, and computational aspects of biological and biomedical imaging, across all scales of observation. ISBI is a joint initiative from the Institute for Electrical and Electronics Engineers (IEEE) Signal Processing Society (SPS) and the IEEE Engineering in Medicine and Biology Society (EMBS). 

Aniket Pramanik, Hemant Aggarwal, and Mathews Jacob

The paper, titled Off-the-Grid Model Based Deep Learning (O-MoDL), was authored by Aniket Pramanik, Hemant Aggarwal, and Mathews Jacob, who are researchers at the Computational Biomedical Imaging Group (CBIG) within the Electrical and Computer Engineering and the Iowa Institute of Biomedical Imaging at the University of Iowa.

The paper introduces a deep learning algorithm for the reconstruction of MRI data from highly under-sampled measurements, which helps significantly reduces the scan time of MRI scans. The model based deep learning strategy (MoDL) significantly reduces the training data demand compared to other deep learning image reconstruction schemes. Moreover, the off-the-grid strategy builds upon super-resolution image reconstruction algorithm, also developed at CBIG and which won the best paper award at ISBI 2015.