search

JingyinLin / PKEcho-Net

Super-efficient Echocardiography Video Segmentation via Proxy- and Kernel-Based Semi-supervised Learning (AAAI 2023)
2 stars 0 forks source link

Model Outputs #3

Open OrientKing opened 1 month ago

OrientKing commented 1 month ago

This is an excellent work, when reading the paper and the code, we observe that the model has 4 outputs of the same shape, how to deal with these outputs respectively during training and testing? Looking forward to your reply, thank you.

JingyinLin commented 1 month ago

During training, all outputs are used to calculate the loss function, supervising the predictor and each kernel refinement in the KSS module. Only the last output is used for evaluation when testing.