Differential Elastic Image Registration
Senthil Periaswamy, Hany Farid · 2001
Introduction There are a variety of methods for medical image registration (see [9, 10, 26, 19, 17] for general surveys) . Differential registration techniques, however, are often cited as being ineffective, and as such have received little attention (exceptions include [4, 16, 20, 21]). This is unfortunate as differential motion techniques have been quite effective in the Computer Vision community (e.g., [15, 18, 25, 1, 14, 2, 13, 5, 12, 6, 7, 24]). Here we present an effective technique for elastic image registration built upon a differential framework. This technique models the mapping between images as a locally affine but globally smooth warp, and explicitly accounts for variations in image intensities. The resulting registration is simple and automatic. Results from several synthetic and clinical images are shown. 2. Methods We formulate the problem of image registration between a source and target image within a differential (non-feature based) framework. This formulation