摘要 |
Potential lesions in a plurality of medical images 302 are identified by lesion segmentation 304, and parameter values for each of the potential lesions are acquired 306. The potential lesions are then classified manually 308. The sets of parameter values and their corresponding classes are balanced 310. Subtractive clustering 312 is then applied to the balanced parameter value sets of each class, to derive cluster centres. Fuzzy inference rules are derived between the cluster centres and the classes, to create a fuzzy inference model 316. |