Hipposeg

sagital

coronal

axial

Hippocampal segmentation using the UNet network for each axis, inspired by https://github.com/MICLab-Unicamp/e2dhipseg

  • Red: False Positive
  • Green: True Positive
  • Blue: False Negative

Install

Clone the repository

git clone https://github.com/aguirrejuan/hipposeg.git

Install requeriments

pip install -r requirements.txt

Inference

Download Pretrained Weights

Download the Weigths trained with ADNI HarP Dataset and put in the model directory.

Run the code

python inference.py  --path_vol "example.nii"

The result will be in the current directory

Train

Download the VGG11 model

Download the VGG11 model and put in the model directory.

Run the code

python train.py  --train_path "directory_mris/"  \
                  --train_paht_label "directory_mris_labels/" \
                  --epochs number_of_epochs 

Evaluate the Model

python  evaluate.py --path_data "directory_mris/" \
                    --path_data_label "directory_mris_labels/"

GitHub

https://github.com/aguirrejuan/hipposeg