Process

The Process menu offers a denoising and a reconstruction option, depending on the datatype of the loaded data.

Denoising

If the loaded file contains image data denoising can be conducted. In Data Selection the user can choose wether the whole scan or just the current slice should be denoised. With the Model Type option it is possible to choose between different regularizations and dataterm types for the assumed inverse problem model. The selection of the model determines which type of noise can be removed from the image data. The Parameters option holds the value for the regularizations parameter Lambda, whereby a lower value enforces higher regularization. Furthermore, the amount of iterations for the algorithms can be chosen as well as the values for Alpha (only Huber-L2) and Alpha0 and Alpha1 (both only TGV-L2). The last option is the definition of the Inverted Spacing which defines the inverted spacing for the isotropic x and y dimension between the pixels as well as inverted spacing between the slices of a dataset in the z dimension. After the button OK is pressed, the calculation of the denoised image starts in a separate thread, so that the user interface does not freeze. Once the calculation terminates, the denoised dataset is presented in a new window and can be loaded in the main window by pressing OK in this new window. From there the new dataset can be saved in a new file with the Save dataset option.

Reconstruction

If the loaded file contains kspace data reconstruction can be conducted. In Data Selection the user can choose wether the the current slice, the whole dataset applying 2D-FFT or the whole dataset applying 3D-FFT should be reconstructed. With the Model Type option it is possible to choose between TV and TGV reconstruction, alway with a L2 dataterm. TV favors piecewise linear solutions whereas TGV favors piecewise continous solutions. The Parameters option holds the value for the regularizations parameter Lambda, whereby a lower value enforces higher regularization. Furthermore, the amount of iterations for the algorithms can be chosen as well as the values Alpha0 and Alpha1 (both only TGV-L2). The last option is the definition of the Inverted Spacing which defines the inverted spacing for the isotropic x and y dimension between the pixels as well as inverted spacing between the slices of a dataset in the z dimension which is important for the gradient calculation. After the button OK is pressed, the calculation of the reconstructed image starts in a separate thread, so that the user interface does not freeze. Once the calculation terminates, the reconstructed dataset is presented in a new window and can be loaded in the main window by pressing OK in this new window. From there the new dataset can be saved in a new file with the Save dataset option.