| 摘要 |
A method is disclosed for segmenting an organ from volume data sets in medical imaging. A likelihood data set is generated. A binary data set is generated with the aid of a threshold value method, in which the organ region is eroded step by step. When the organ region is separated into a number of subregions, only one of the subregions, which is most likely to contain the organ, is selected for further erosion in each instance. This subregion is identified from geometrical features of the organ, which are known from the training data sets. In the likelihood data set the likelihood values of voxels of the subregions separated in the binary data set are reduced, so that a correspondingly corrected likelihood data set results. On the basis of this corrected likelihood data set the organ from the likelihood data set is segmented. |
