Sampurna Biswas


     Research scientist


     Milpitas, CA
   PhD (2013-2018), University of Iowa  

     Research interest: Machine learning, Compressed sensing; linear inverse problems in dynamic image recovery

  • Integrating deep learned priors in model based optimization for free breathing, ungated, undersampled, dynamic cardiac image reconstruction  on TensorFlow.
  • Recovery of structured signals with missing data by devising supervised signal decomposition and deriving performance guarantees. Optimized MR acquisition & reconstruction techniques, specifically in CINE, myocardial perfusion & brain parametric mapping MRI,  using model based compressed sensing techniques
  • Image series (MRI to CT) synthesis using deep learning techniques on Lasagne/ Theano platform.

Integrating patient specific and population geenric priors for accelerated free breathing MR recovery

In my current research, I am working on integrating model based and learn-able priors in real time reconstruction of accelerated free breathing, un-gated undersampled, dynamic cardiac MR image reconstruction on TensorFlow platform, on the UIowa HPC system. 

Relevant papers and posters



Image series (MRI to CT) synthesis using deep learning techniques on Lasagne/ Theano platform: Siemens internship 2017

Results and codes on medical imaging data are owned by Siemens Medical solutions proprietary.

(a) Original CT (b-c) competing methods (d) proposed


Relevant papers and posters : 


Convex Recovery of Continuous Domain Piecewise Constant Images From Nonuniform Fourier Samples

In a joint work with Dr. Greg Ongie, we devised performance guarantees on convex low rank recovery of piece-wise constant images with non-uniform Fourier measurements, pertaining to the MRI setting.    

Relevant papers: 



Study the spark of samples of Fourier (DFT) matrices for sparse recovery

Suppose W_N represents NXN Discrete Fourier Transform (DFT) matrix. Then what are the coprime conditions that help us choose L rows of the DFT matrix s.t its spark equals the maximum possible value  i.e L+1 ?Here, spark is the smallest number of linearly dependent columns in the matrix. 

Relevant papers: 




  1. S. Biswas, H. Aggarwal, M. Jacob, "Dynamic MRI using model-based deep learning and STORM priors: MoDL_SToRM", Magnetic Resonance in Medicine, 2019
  2. G. Ongie, S. Biswas, M. Jacob, "Convex Recovery of Continuous Domain Piecewise Constant Images From Nonuniform Fourier Samples",  IEEE Transactions on Signal Processing 66 (1), 236 - 250, 2017.
  3. S. Biswas, S. Dasgupta, R. Mudumbai, M. Jacob, "Subspace Aware Recovery of Low Rank and Jointly Sparse Signals", IEEE Transactions on Computational Imaging 3 (1), 22-35, 2016
  4.  H. Achanta, S. Biswas, M. Jacob, S. Dasgupta, R. Mudumbai, "The spark of Fourier matrices: Connections to vanishing sums and coprimeness", Digital Singal Processing, 61, 76-85, 2017. 


  1. S. Biswas, H.K. Aggarwal, S. Poddar, and M. Jacob, "Model-based free-breathing cardiac MRI reconstruction using deep learned & STORM priors: MoDL-STORM", accepted in ICASSP 2018. 
  2. G. Ongie, S. Biswas, M. Jacob, "Structured matrix recovery of piecewise constant signals with performance guarantees", ICIP 2016, Phoenix.
  3. S. Biswas, S. Dasgupta, M. Jacob, R. Mudumbai, "Spark under 2 D Fourier Sampling", EUSIPCO, Nice, France, 2015.
  4. S. Biswas, S. Poddar, S. Dasgupta, R. Mudumbai, M. Jacob. "Two step recovery of jointly sparse and low-rank matrices: theoretical guarantees", ISBI 2015, New York City. 
  5. H. Achanta, S Biswas, S. Dasgupta, M. Jacob, B. Dasgupta, R. Mudumbai. ”Coprime conditions for Fourier sampling for sparse recovery” Sensor Array and Multichannel Signal Processing Workshop (SAM), IEEE, 2014.


  1. Recipient of Best Graduate poster Award for the engineering research open house 2018, under the Iowa Institute of Biomedical Imaging category.
  2. Recipient of the trainee stipend award for attending the ISMRM workshop on machine learning, 2018.
  3. Recipient of Graduate College Post-Comprehensive research Award for Fall of 2017.
  4. Recipient of NIH travel award for attending ISBI 2015.