MULTI-CENTER EFFECT COMPENSATION METHOD BASED ON PET/CT INTELLIGENT DIAGNOSIS SYSTEM

摘要

Disclosed is a multi-center effect compensation method based on a PET/CT intelligent diagnosis system, belonging to the field of medical images. The method comprises: on the basis of a location scale model regarding additive and multiplicative multi-center effect parameters, using a non-parametric mathematical method for data in a training center A and data in a test center B, estimating multi-center effect parameters of the test center B relative to the training center A, and using the parameters to compensate for the data in the test center B so as to eliminate a multi-center effect between the test center B and the training center A. By means of the present invention, a multi-center effect between the training center A and the test center B can be compensated for, such that the data in the test center B, after compensation, can be used in a model trained in the training center A, thereby indirectly improving the generalization ability of the model.