⚡ This repository includes the official implementation of the paper:
👋 RoSD-SAR: Robust Ship Detection in SAR Images with Noisy Box Labels
👨💻 Code: GitHub
- Set up environment
conda create -n RoSD_SAR python=3.7
conda activate RoSD_SAR
- install pytorch
conda install pytorch==1.10.0 torchvision==0.11.0 -c pytorch -c conda-forge
- Install
git clone https://github.com/douling843/RoSD_SAR.git
cd RoSD_SAR
pip install -r requirements/build.txt
cd mmcv
MMCV_WITH_OPS=1 pip install -e .
cd ..
pip install -e .
python ./utils/gen_noisy_ssdd.py --box_noise_level 0.4
python tools/train.py configs/faster_rcnn/faster_rcnn_r50_fpn_1x_ssdd_RoSD_SAR.py --work-dir='work_dir/ssdd/faster_rcnn_r50_fpn_1x_ssdd_RoSD_SAR'
python tools/train.py configs/faster_rcnn/faster_rcnn_r50_fpn_1x_ssdd_RoSD_SAR.py work_dirs/ssdd/faster_rcnn_r50_fpn_1x_ssdd_RoSD_SAR/epoch_12.pth --show-dir work_dirs/vis/ssdd/RoSD_SAR
This repository is based on mmdetection 🤝 and OA-MIL 👯.