Surface matching of multimodality image volumes by a fuzzy elastic registration technique

Multimodality image registration is useful in diagnostic imaging and treatment planning for radiation therapy. In this paper, we present a technique which registers the surfaces of two volumes acquired by different medical imaging modalities. We represent the image volumes in terms of their surface elements known as tiles. We identify the fuzzy variables, assign fuzzy membership functions to them and generate a fuzzy rule database. The fuzzy algorithm reduces the discrepancy between the two set of tiles until the surfaces are matched. In order to study the efficacy of our approach, we severely warp a simulated image and register it with its original. We register CT and MR volumes of humanoid phantom images. Finally, we present the results at the end of the article. (C) 1998 Elsevier Science Ltd. All rights reserved.