Temporal registration for MRI time series

Abstract

Time-course analysis in medical image series often suffers from serious motion. Registration provides voxel correspondences among images, and is commonly employed for correcting motion in medical images. Yet, the registration procedure fails when aligning volumes that are substantially different from template. We present a robust method to correct for motion and deformations in MRI time series. We make a Markov assumption on the nature of deformations to take advantage of the temporal smoothness in the image data. Forward message passing in the corresponding hidden Markov model (HMM) yields an estimation algorithm that only has to account for relatively small motion between consecutive frames. We demonstrate the utility of the temporal model by showing that its use for in-utero MRI time series alignment improves the accuracy of the segmentation propagation through temporal registration. Our results suggest that the proposed model captures accurately the temporal dynamics of deformations present in in-utero MRI time series. We also demonstrate that our method can be used for cardiac cine MRI. By propagating segmentation labels of one volume to the other frames in the cine MRI through deformation estimated by our method, 4D (3D+time) cardiac MRI series can be segmented.