Breast Lesion segmentation

3D Slicer module to deploy deep learning models for segmentation of ultrasound images.

Prerequisites

• Install SlicerOpenCV extension from the 3D Slicer Extension Manager. If you are using Slicer 5.0.2 or later open the Python interactor in Slicer and type the following command:
  

slicer.util.pip_install("opencv-python")

• Install PyTorch in 3D Slicer's Python console: install PyTorch. For that, open the Python interactor in Slicer and type the following command:
  

pip_install('torch torchvision torchaudio')

• To load the deep learning model used in the module "Breast Lesion Segmentation", install the library segmentation-models-pytorch in 3D Slicer. For that, open the Python interactor in Slicer and type the following command:
  

pip_install('segmentation-models-pytorch')

Information about the Dataset

The models used for classification and segmentation have been trained using Breast Ultrasound Images Dataset (Dataset BUSI)

Al-Dhabyani W, Gomaa M, Khaled H, Fahmy A. Dataset of breast ultrasound images. Data in Brief. 2020 Feb;28:104863. DOI: 10.1016/j.dib.2019.104863