Method for determining at least one enhanced object feature (19) of an object (16, 17, 18, 32) of interest that is at least partially depicted by an image (14) in a region of an anatomy of a patient (43), said image (14) having been provided from an imaging device (48), wherein a respective object feature (35) for each object (16, 17, 18, 32) and at least one semantic feature (20) are determined from the image data of the image (14), wherein said object feature (35) depends on information comprised in image data of a first region (34) of interest, said first region (34) of interest containing the object region (33), and the semantic feature (20) depends on information comprised in image data of a second region (37), said second region (37) of interest comprising a region excluding the object region (33), and wherein said semantic feature (20) comprises an anatomic context information, wherein the enhanced object feature (19) is determined from the object feature (35) and the semantic feature (20) .