Graphical image correlation and processing system for the evaluation of multidimentional Positon Emission Tomographic images. Pseudo-3D representation and correlation of structural/neurofunctionnal data sets

Multimodality 2D and 3D display are likely to be particularly valuable for interpreting Positon Emission Tomographic studies. Even under ideal conditions differences in scanner characteristics prevent correlation of features observed in separate scans. Methods for selecting and calculating arbitrary image sections for displaying are investigated. MRI data were used to construct a 3D computer model of the brain surface, PET derived measured were encoded as densities and mapped into the 3D models. The object re-sampling along the zz axis is based on the triangulation approach between successive contours. Arbitrary oblique image section is calculated and displayed with coordinate pixel transformation. The PET and MR images were then aligned to obtain the best match of brain contours in order to use elastic deformations.