BFSeg-Efficient-Building-Footprint-Segmentation-Framework
The manuscript is in the peer review process and thus only the evaluation code is avalible~
The training code will be released once it is accepted~ Coming Soon~
Dataset
1.WHU building
2.DeepGlobe
3.Typical City Building Dataset
The code
Requirements
- torch
- torchvision
- pillow
- cv2
Usage
Clone the repository:git clone https://github.com/HaonanGuo/BFSeg-Efficient-Building-Footprint-Segmentation-Framework.git
->Run s2_Eval_BFSeg.py to evaluate the performance of BFSeg
Dont forget to download the model weights before evaluation:
Dataset | WHU | DeepGlobe | Typical City |
---|---|---|---|
Backbone | ConvNext | ConvNext | ConvNext |
Backbone | SwinTrans | SwinTrans | SwinTrans |
Help
Any question? Please contact us with: haonan.guo@whu.edu.cn