Abstract
An average CT brain image is constructed to serve as a reference frame for intersubject comparison. The use of an average brain image prevents bias resulting from registration to a single patient image. The average CT brain image is constructed from a large number of images by means of rigid registration with scaling, and with normalized mutual information as the registration measure. The construction proceeds in two steps. First a temporary average image is constructed based on a small subset of images. Then this image is used as the reference image for the construction of the average CT brain image. It is shown that registration to the average CT brain image is consistent and effective for the registration of new CT images. Furthermore, registration of new CT images to the average CT brain image outperforms registration to a single patient image.
Original language | English |
---|---|
Title of host publication | CARS 2002 Computer Assisted Radiology and Surgery |
Subtitle of host publication | Proceedings of the 16th International Congress and Exhibition Paris, June 26–29,2002 |
Editors | H.U. Lemke, K. Inamura, K. Doi, M.W. Vannier, A.G. Farman, J.H.C. Reiber |
Place of Publication | Berlin |
Publisher | Springer |
Pages | 425-429 |
ISBN (Print) | 3-540-43655-3 |
DOIs | |
Publication status | Published - 2002 |
Externally published | Yes |
Keywords
- atlas
- average brain
- intersubject comparison