| 摘要 |
The invention concerns a method for segmenting an image or a sequence of initial three-dimensional images to obtain a three-dimensional image of segmentation comprising partitioning into regions of interest, said image or sequence of images including measurements, for each voxel and during n time intervals (n1), of the real evolution of a signal representing at least one variable of said image or sequence, which essentially includes: a) modelling (10) the signal including definition of a parametric model of spatio-temporal evolution of said signal, said model comprising sets of homogeneous parameters respectively pertaining to structures corresponding to said regions of interest; b) extracting (30) voxel samples respectively included in said structures; then c) merging (50b) the samples by assembling those whereof the model of evolution pertains to the same structure, said merging following, preceding or including classifying all the voxels of said image or sequence of images or a region of interest thereof by being aggregated to a group of samples. |
