Change Detection of Medical Images for Three Dimensional Volumetric Data

Change detection is a fundamental problem in various fields, such as image surveillance, remote sensing, medical imaging, etc. This paper extends our recent work in the area of automated change detection for MR images, by considering three dimensional (3D) volumetric data. We aim to detect changes within the 3D volumes of the same anatomical objects taken at two different times. We define the change detection as an optimization problem and propose two new 3D volumetric algorithms, the 3D AEDL-2 and the 3D EigenBlockCD-2 for detecting the changes automatically. We show the performance of the 3D EigenBlockCD-2 algorithm using real MR brain images.