TV - MRI Reconstruction Compressed Sensing
Download the dataset from: https://drive.google.com/file/d/0B4nLrDuviSiWajFDV1Frc3cxR0k/view?usp=sharing
Physical data acquisition can be represented as,
A general formulation that also allows for noise in the observed signal $y $ is given by the following optimization problem, least squares minimization along with regularization term,
We can redefine the physical data acquisition model for multi-coil MRI as,
In TV MRI reconstruction, we represent the matrix D as follow,
Iteration updates:
The operators Ds and Dt represent spatial and temporal difference operators, which are designed to induce sparsity within both the image and temporal domains.
Iteration updates:
Iteration updates:
his method involves reconstructing the first frame at a lower acceleration rate independently and utilizing its magnitude as a reference for the reconstruction of subsequent images. By adopting this strategy, we can improve the Peak Signal-to-Noise Ratio (PSNR). When applying this approach to all frames within a dataset, we consistently observe that the PSNR exceeds that of independent reconstructions throughout the sequence.
Parisima Abdali: pa2297@nyu.edu