Abstract: A method for machine learning includes learning, during a training stage, network parameter values of a neural network to obtain a trained neural network configured to perform reconstruction of medical images; refining, during a subsequent refinement stage, the learned network parameter values to generate refined network parameter values defining a refined neural network; and applying input medical image data to the refined neural network to generate a reconstructed medical image. The method retains benefits of machine learning image reconstruction to obtain a desired reconstructed image.

Abstract: A method for visualization of dynamic objects using a generative manifold includes steps of: acquiring a set of measurements associated with the dynamic objects using sensors; estimating parameters of a generator using the set of measurements and estimating latent variables using the set of measurements; modeling using a computing device the dynamic objects as a smooth non-linear function of the latent variables using the generator such that points in a latent subspace are mapped to a manifold in a generative manifold model; and generating a visualization of the dynamic objects using the generative manifold model. The set of measurements may include multi-slice data. The generative manifold model may provide for modeling deformations.

Abstract: A method for machine learning includes learning, during a training stage, network parameter values of a neural network to obtain a trained neural network configured to perform reconstruction of medical images; refining, during a subsequent refinement stage, the learned network parameter values to generate refined network parameter values defining a refined neural network; and applying input medical image data to the refined neural network to generate a reconstructed medical image. The method retains benefits of machine learning image reconstruction to obtain a desired reconstructed image.

Abstract: An apparatus for reconstructing or filtering medical image data is provided. The apparatus includes processing circuitry to receive a first medical image data and meta-parameters related to the first medical image data; apply the received meta-parameters to inputs of a first trained machine-learning (ML) network, e.g., a multilayer perceptron, to obtain, from outputs of the first trained ML network, tuning parameters of a second ML network (e.g., a convolutional neural network) different from the first ML network; apply the received first medical image data to inputs of the second ML network, as tuned by the obtained tuning parameters output from the first ML network, to obtain, from outputs of the second ML network, second medical image data; and output the second medical image data. In one embodiment, the first medical image data is magnetic-resonance k-space data and the second medical data is a magnetic-resonance image.

Abstract: A method for visualization of dynamic objects using a generative manifold includes steps of: acquiring a set of measurements associated with the dynamic objects using sensors; estimating parameters of a generator using the set of measurements and estimating latent variables using the set of measurements; modeling using a computing device the dynamic objects as a smooth non-linear function of the latent variables using the generator such that points in a latent subspace are mapped to a manifold in a generative manifold model; and generating a visualization of the dynamic objects using the generative manifold model. The set of measurements may include multi-slice data. The generative manifold model may provide for modeling deformations.

Abstract: Methods and a system to unify reconstruction and motion estimation steps in first pass cardiac perfusion MRI include a global objective function that meets data consistency, spatial smoothness, motion and contrast dynamics constraints. The global objective decomposed into simpler sub-problems which include low pass filtering of a deformed object, TV shrinkage, analytical Fourier replacement and an l2 minimizing problem. A registration tool based on the local cross-correlation similarity measure and enabled to perform both rigid and flexile deformations, is applied. Registration parameters are tuned by rigid, semi rigid and flexible models at different stages of iterations. A system to perform the methods is also disclosed.

Abstract: Methods and a system to unify reconstruction and motion estimation steps in first pass cardiac perfusion MRI include a global objective function that meets data consistency, spatial smoothness, motion and contrast dynamics constraints. The global objective decomposed into simpler sub-problems which include low pass filtering of a deformed object, TV shrinkage, analytical Fourier replacement and an l2 minimizing problem. A registration tool based on the local cross-correlation similarity measure and enabled to perform both rigid and flexile deformations, is applied. Registration parameters are tuned by rigid, semi rigid and flexible models at different stages of iterations. A system to perform the methods is also disclosed.