The code uses SegFormer for Semantic Segmentation on Drone Dataset.
The details for the SegFormer can be obtained from the following cited paper and the drone dataset can be downloaded from the link below.
Alternatively, you can also download the dataset from Kaggle, the link is mentioned below.
Clone the repository and install all the packages mentioned in the requirement.txt file.
If you just want to infer the semantic segmentation, open the, change the image file name you want to test and run the code.
Make sure the trained model is in the model folder. You can download the model at
Alternatively, you can train the model and save it, locally, by running

If you want to train the SegFormer on the drone dataset. Make sure that the directory structure is as follows:
| drone_dataset

Demo Inference
Alt text

Citations and References

  title={SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers},
  author={Xie, Enze and Wang, Wenhai and Yu, Zhiding and Anandkumar, Anima and Alvarez, Jose M and Luo, Ping},
  journal={arXiv preprint arXiv:2105.15203},

Drone Dataset

Aerial Semantic Segmentation Drone Dataset


GitHub - sander-ali/Sematic_Segmentation_With_SegFormer at
The code uses SegFormer for Semantic Segmentation on Drone Dataset. - GitHub - sander-ali/Sematic_Segmentation_With_SegFormer at