DIRECT VOLUME RENDERING OF 4D DEFORMABLE VOLUME IMAGES

摘要

A method and system are presented for direct volume rendering of a deformable a volume dataset that represent an object in motion. An original 3D image dataset of the object is acquired at an initial stage of the motion, and segmented. Deformed 3D image datasets of the object are acquired at subsequent stages of the motion. Deformation matrices are computed between the segmented original 3D image dataset and the deformed 3D image datasets. A plurality of deformed mesh patches are generated based on the deformation matrices. 2D textures are generated by dynamically sampling the segmented 3D image dataset, and applied to the deformed mesh patches. The textured deformed mesh patches are evaluated, fended, and composited to generate one or more volume-­rendered images of the object in motion.