This repository hosts the codes for the implementation of the paper Weakly Supervised Image Segmentation Beyond Tight Bounding Box Annotations (under review).
Download Promise12 dataset, and put it on the "data/prostate" folder.
Download Atlas dataset, and put it on the "data/atlas" folder.
Run the following codes for preprocessing:
# trainig and valid subsets for promise12 dataset
python preprocess/slice_promise_train_val.py
python preprocess/slice_promise_augment_train_val.py
# trainig and valid subsets for atlas dataset
python preprocess/slice_atlas.py
# The following experiments include MIL baseline (exp_no=0),
# parallel transformation based MIL (exp_no=1)
# polar transformation based MIL (exp_no=2,3,4,5),
# the proposed approach with weighted alpha-softmax approaximation (exp_no=6,8),
# the proposed approach with weighted alpha-quasimax approaximation (exp_no=7,9),
CUDA_VISIBLE_DEVICES=0 python tools/train_atlas_beyond_tightbb.py --n_exp exp_no
CUDA_VISIBLE_DEVICES=0 python tools/train_promise_beyond_tightbb.py --n_exp exp_no
# Dice validation results for promise12 dataset, exp_no=0,1,2,3,4,5
CUDA_VISIBLE_DEVICES=0 python tools/valid_promise_beyond_tightbb.py --n_exp exp_no
# Dice validation results for atlas dataset, exp_no=0,1,2,3,4,5
CUDA_VISIBLE_DEVICES=0 python tools/valid_atlasbeyond_tightbbx.py --n_exp exp_no
python tools/report_promise_beyond_tightbb.py
python tools/report_atlas_beyond_tightbb.py
# exp_no = 1,2,3,4
python tools/plot_promise_polar_center.py --n_exp exp_no
python tools/plot_atlas_polar_center.py --n_exp exp_no
Please consider citing our paper in your publications if the project helps your research.
@inproceedings{wang2021bounding,
title={Bounding Box Tightness Prior for Weakly Supervised Image Segmentation},
author={Wang, Juan and Xia, Bin},
booktitle={International Conference on Medical Image Computing and Computer-Assisted Intervention},
pages={526--536},
year={2021},
organization={Springer}
}
@article{wang2022polar,
title={Polar Transformation Based Multiple Instance Learning Assisting Weakly Supervised Image Segmentation With Loose Bounding Box Annotations},
author={Wang, Juan and Xia, Bin},
journal={arXiv preprint arXiv:2203.06000},
year={2022}
}