A repo of a project in DS503 - Spring 2024 in KAIST.
conda env create -f environment.yml
conda activate Wavelet_MedSeg
# Run the Wavelet model for segmentation #
python train.py --num_epoch [# of epochs] --save_content --batch_size [# of batch size] --config [# name of the config file]
### example ###
python train.py --num_epoch 100 --save_content --batch_size 14 --config monuseg_wavelet_NoGAN.yml
# Run the Unet model for segmentation #
python train_Unet.py --num_epoch [# of epochs] --save_content --batch_size [# of batch size] --config monuseg.yml
# Run the Wavelet model for segmentation #
python infer.py --config [# name of the config file]
# Run the Unet model for segmentation #
python infer_Unet.py --config monuseg.yml
### The output of inferencing is located in the experment folder ###
# Change the root folder is the file that contain the inference output images in the file 'bf-score.py' then run #
python bf-score.py
Data setlink:
Please download then change the data path in the config file.
This wavelet idea comes from Wavelet Diffusion. If you use this wavelet idea, please cite below:
@inproceedings{phung2023wavelet,
title={Wavelet diffusion models are fast and scalable image generators},
author={Phung, Hao and Dao, Quan and Tran, Anh},
booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
pages={10199--10208},
year={2023}
}