Fig. 1.
Overview of (a) our proposed Incremental Transfer Learning framework, and (b) the multi-site expert model. Note that in this study, we only use one multi-site expert model and one source decoder network, which will not introduce additional parameter.