PRICE

Download Matlab Code: Code

Framework: We introduce a fast and an efficient patch smoothness regularization scheme, which implicitly compensates for interframe motion, to recover dynamic MRI data from highly undersampled measurements. The regularization prior is a sum of distances between each rectangular patch in the dataset with other patches in the dataset using a saturating distance metric. Unlike current motion estimation and motion compensation (ME-MC) methods, the proposed scheme does not require explicit prior information such as reference frames or motion model. The proposed algorithm, which alternates between inter-patch shrinkage step and conjugate gradient algorithm, is considerably more computationally efficient than ME-MC methods. The reconstructions obtained using the proposed algorithm is compared against stateof-the-art methods using free-breathing cardiac cine and myocardial perfusion datasets, using retrospective and prospective multi-channel experiments

llustration:

                                           

(a) Illustration of the proposed PRICE scheme. The regularization term penalizes the differences between each patch and other patches in its cube shaped neighborhood. The green squares indicate the location of the patch in the current frame and the ones with the highest similarity in the neighboring frames. The dashed red box represents the neighborhood where the patches move within. The distance metric used for the comparison is shown by the dotted black curve in (b). The metric heavily penalizes the distances between similar patches, while it eliminates dissimilar patches from the comparison; this minimizes the spatiotemporal blurring induced by the averaging of dissimilar patches. The colored curves correspond to the different approximations of the distance metric, which enables fast algorithms. (c) The shrinkage rule for the inter-patch differences t v(|t|) using lp. We rely on continuation schemes as shown in (b) and (c) starting with low values of and gradually increase it to high values, when the approximation is more accurate. (d) The algorithm alternates between a simple shrinkage step to denoise inter-patch differences and image update step, which involves a computationally efficient conjugate gradients algorithm.

Results (see [1] for experimental details)

Heavily-breathing stress myocardial perfusion datataset, acquired using Cartesian acquisition. 

Free-breathing stress myocardial perfusion datataset, acquired using golden angle radial acquisition. 

Breath-held ECG-gated cardiac CINE dataset acquired using Cartesian acquisiton.

     

  1. Free-breathing ECG-gated cardiac CINE.

References:

  1. Y.Mohsin, S.G Lingala, E. DiBella, M.Jacob, Accelerated dynamic MRI Using Patch Regularization for Implicit motion CompEnsation (PRICE),  Magnetic Resonance in Medicine, in press.
  2. Y. Mohsin, S.G. Lingala,E Dibella, M Jacob, Motion compensated free breathing myo-cardial perfusion MRI using iterative non-local shrinkage, ISMRM, Toronto, Canada, 2015.
  3. Y. Mohsin, Z. Yang, S.G.Lingala, M.Jacob, "Motion compensated dynamic imaging without explicit motion estimation", ISMRM, Milan, Italy, 2014. 
  4. Z. Yang, M. Jacob Robust non-local regularization framework for motion compensated dynamic imaging without explicit motion estmation, IEEE ISBI, Barcelona, May 2012.