MSSDF: Modality-Shared Self-supervised Distillation for High-Resolution Multi-modal Remote Sensing Image Learning

This is the offical code of the paper: MSSDF: Modality-Shared Self-supervised Distillation for High-Resolution Multi-modal Remote Sensing Image Learning

The dataset HR-Pairs in the paper can be download by baidu driver

The pretrain checkpoints can be download by vit-base, vit-large

Release TODO:

• Train code
• Pretraining checkpoints
• Downstream Task
  • Classification
  • Segmentation
  • Depth Estimation
  • +DSM Segmentation
  • Change Detection
  • HBB Object Detection
  • RBB Object Detection

Acknowledgments

This project includes code from scale-mae