Figure 1: Overall architecture of 2DSegFormer.

2DSegFormer: 2-D Transformer Model for Semantic Segmentation on Aerial Images.
Xinyu Li, Yu Cheng, Yi Fang, Hongmei Liang and Shaoqiu Xu.
IEEE Transactions on Geoscience and Remote Sensing.

This repository contains the official Pytorch implementation of training & evaluation code.

2DSegFormer is an efficient 2-D semantic transformer model for semantic segmentation on aerial images, as shown in Figure 1.

We use MMSegmentation v0.13.0 as the codebase.

For install and data preparation, please refer to the guidelines in MMSegmentation v0.13.0 and SegFormer.

Download the pretrained weights of Mix Transformers (MiTs) on ImageNet-1K in SegFormer, and put them in a folder pretrained/.

Training and evaluation according to the coding style of MMSegmentation v0.13.0.

License under an MIT license.

If you find the code or trained models useful, please consider citing:

@ARTICLE{9955997,
  author={Li, Xinyu and Cheng, Yu and Fang, Yi and Liang, Hongmei and Xu, Shaoqiu},
  journal={IEEE Transactions on Geoscience and Remote Sensing}, 
  title={2DSegFormer: 2-D Transformer Model for Semantic Segmentation on Aerial Images}, 
  year={2022},
  volume={60},
  number={},
  pages={1-13},
  doi={10.1109/TGRS.2022.3223416}}