Improved Fitting Calculation For Diffusion Tensor Imaging

Diffusion tensor imaging is a new MRI modality that is becoming an important technique to assess various types of brain diseases. However, this technique has one weakness; its extreme sensitivity to patient motions. The image is often corrupted by patient motions as small as cardiac pulsation of the brain. Even though this problem is ameliorated by employment of rapid imaging, we often find corrupted images are included in the scans, which leads to substantial errors in tensor calculation. Currently, one of the best methods is to visually inspect all the raw images and remove corrupted images from the subsequent tensor calculation. However, the number of raw images is typically 1,000 - 4,000 and the process could be very time consuming. In this invention, techniques have been developed to automatically detect and remove the corrupted images.

Category(s):