
A novel method called ToNNO (Tomographic Reconstruction of a Neural Network's Output) that can perform dense prediction tasks on 3D medical images using only a 2D image encoder trained in a weakly supervised manner.


coremsg

weakly-supervised-segmentation-of-3d-medical-images-using-tomographic-reconstruction-of-neural-network-outputs


Weakly Supervised Segmentation of 3D Medical Images Using Tomographic Reconstruction of Neural Network Outputs
