Fig. 1.
A schematic diagram of the proposed method with the feature extractor and the multi-view semantic embedding (MVSE) network. In this diagram, ΨX (·), ΨCT (·) and ΨT (·) are the x-ray, CT and trait signatures corresponding to different diseases and conditions. The MVSE network has a common input layer (CIL) and two parallel branches (x-ray branch and CT branch) with guiding networks.