The official implementation of ShareCMP: Polarization-Aware RGB-P Semantic Segmentation.
This project is based on MMSegmentation.
- Python 3.8
- PyTorch 1.13.1+cu116
Step 1. Create a conda virtual environment and activate it.
conda create -n sharecmp python=3.8 -y
conda activate sharecmpStep 2. Install PyTorch following official instructions.
Linux and Windows
# Wheel CUDA 11.6
pip install torch==1.13.1+cu116 torchvision==0.14.1+cu116 torchaudio==0.13.1 --extra-index-url https://download.pytorch.org/whl/cu116# Conda CUDA 11.6
conda install pytorch==1.13.1 torchvision==0.14.1 torchaudio==0.13.1 pytorch-cuda=11.6 -c pytorch -c nvidiaStep 3. Install MMSegmentation and dependent packages.
pip install -U openmim
mim install mmengine==0.8.5
mim install mmcv==2.0.1
mim install mmsegmentation==1.1.2
pip install -r requirements.txtThe data structure UPLight ( download here) looks like below:
# UPLight
data
├── UPLight
│ ├── images_rgb
│ ├── images_rgb_0
│ ├── images_rgb_45
│ ├── images_rgb_90
│ ├── images_rgb_135
│ ├── labels
│ ├── train.txt
│ ├── val.txt
bash tools/dist_train.sh configs/sharecmp/sharecmp_mit-b2_2xb4-200e_uplight-512x612.py 2The weight .pth of ShareCMP is available here.
bash tools/dist_test.sh configs/sharecmp/sharecmp_mit-b2_2xb4-200e_uplight-512x612.py sharecmp_mit-b2_2xb4-200e_uplight-512x612.pth 2@article{liu2025sharecmp,
title={ShareCMP: Polarization-Aware RGB-P Semantic Segmentation},
author={Liu, Zhuoyan and Wang, Bo and Wang, Lizhi and Mao, Chenyu and Li, Ye},
journal={IEEE Transactions on Circuits and Systems for Video Technology},
volume={35},
number={10},
pages={10316-10329},
year={2025},
doi={10.1109/TCSVT.2025.3570764},
}