• Code for the paper:"Contrastive learning-based knowledge distillation for RGB-thermal urban scene semantic segmentation", KBS, 2024
• More details refer to the paper "Contrastive learning-based knowledge distillation for RGB-thermal urban scene semantic segmentation", which has been published in the Knowledge-Based Systems.
python>=3.6, Pytorch>=1.7, cudatoolkit>=10.1
Experiments results on the MFNet dataset
| Model | mIoU | weights |
|---|---|---|
| CLNet-T | 58.2% | weight |
| CLNet-S | 53.8% | weight |
| CLNet-S* | 57.3% | weight |
{Xiaodong Guo, Wujie Zhou, Tong Liu,
Contrastive learning-based knowledge distillation for RGB-thermal urban scene semantic segmentation,
Knowledge-Based Systems,
Volume 292,
2024,
111588,
ISSN 0950-7051,
https://doi.org/10.1016/j.knosys.2024.111588.}
If you have any questions, please feel free to contact 3120245534@bit.edu.cn.