Joint Reconstruction of Image and Motion for PET Using Linear Diffusion Regularization

We present a modification of our recently proposed joint reconstruction algorithm for PET. A problem related to the original presentation is the need of a regularization parameter which defines the trade-off between data term and regularization term. This regularization parameter is hard to find and usually there is no other way than trying different values and finally choosing the one which leads to the most satisfactory reconstruction. In our modified version, we eliminate the regularization parameter. Both quantitative and visual comparison show that the modified version of our joint reconstruction algorithm provides the same reconstruction quality and is therefore an acceptable substitute.     «

We present a modification of our recently proposed joint reconstruction algorithm for PET. A problem related to the original presentation is the need of a regularization parameter which defines the trade-off between data term and regularization term. This regularization parameter is hard to find and usually there is no other way than trying different values and finally choosing the one which leads to the most satisfactory reconstruction. In our modified version, we eliminate the regularizatio...     »