Regions of interest tracking in temporal scans based on statistical analysis of gray scale and edge properties and registration of images

Precise measurement of T1 is needed for its use in temperature monitoring in vivo. Movement of tissue relative to fixed regions of interest can result in large variations in apparent T1, with consequent substantial errors in the measured temperature. This paper evaluates methods of tracking regions of interest as tissue moves during a study in an effort to minimize errors from this cause. Tracking techniques evaluated are based on maintaining constant gray scale levels, locating nearby edges and maintaining position relative to them, and global image registration.