Skip to content

Latest commit

 

History

History

AG-SonoNet

README.md

Attention Gated Networks
(Image Classification & Segmentation)

Pytorch implementation of attention gates used in U-Net and VGG-16 models. The framework can be utilised in both medical image classification and segmentation tasks.


The schematics of the proposed Attention-Gated Sononet


The schematics of the proposed additive attention gate

References

  1. "Attention-Gated Networks for Improving Ultrasound Scan Plane Detection", MIDL'18, Amsterdam
    Conference Paper
    Conference Poster

  2. "Attention U-Net: Learning Where to Look for the Pancreas", MIDL'18, Amsterdam
    Conference Paper
    Conference Poster

Dependent library installation

cd code/AG_SonoNet
pip install -e git+https://github.com/ozan-oktay/torchsample.git#egg=torchsample

Please note that the original repository does not specify the environment in which its code is run, such as the versions of the various Python libraries, which may cause errors in the run, such as when using VisDOM for visualization.

Train FETAL_PLANES_DB

python ./train_FPD_Classification.py