METHOD AND SYSTEM FOR AUTOMATIC RECOGNITION OF PRENEOPLASTIC ANOMALIES IN ANATOMIC STRUCTURES BASED ON AN IMPROVED REGION-GROWING SEGMENTATION, AND COMPUTER PROGRAM THEREFOR

摘要

A process for the automatic recognition of anomalies in anatomical structures, as well as a processing system and a computer program for implementing the process are described, the process comprising the steps of : - acquisition (10) of a plurality of two-dimensional images of at least a portion of a patient's body, which are suitable for forming a three-dimensional representation of at least one anatomical structure under observation, - segmentation (20) of a region of interest of the three-dimensional representation, potentially bearing the anomalies, selection (30, 40) of the volume image elements (voxels) of the segmented region that are candidates for belonging to anomalies of the anatomical structure represented, on the basis of predetermined morphological parameters, and classification (70) of the voxels as elements suspected of belonging to anomalies or elements not belonging to anomalies, in which the segmentation (20) takes place by progressive growth of the region of interest by aggregation of voxels in accordance with a predetermined criterion of similarity, starting with seed voxels situated within the region of interest, the region growth step comprising a first volume growth stage in which the growth is carried out on the basis of a predetermined global criterion of similarity between voxels in each direction of growth, and a second fine growth stage for the definition of a boundary of the region of interest, in which the growth is carried out on the basis of local similarity criteria for each respective direction of growth.