Motion correction of PET images using realignment for intra-frame movement

A method is presented for the motion correction of PET images using realignment for intra-frame movement. A newly introduced aspect of the method is that it corrects not only inter-frame but also intra-frame movement, using currently used PET images (which are not corrected for intra-frame movement) and motion tracking data. Although our method requires motion tracking data, it does not require very short time image acquisition nor list mode data acquisition. So, it is applicable to currently used PET images. In our method the following hypothesis is assumed. That is If no movement happens, there exists a linear model such that counts of each voxel per unit time follows the model. Model parameters may depend on the voxels. In a simple example, our method successfully corrected motion artifact and estimates the parameters accurately. It may give a simple and practical solution to the motion correction problems.