EDR-TransUnet: Integrating Enhanced Dual Relation-Attention With Transformer U-Net For Multiscale Rock Segmentation on Mars.

EDR-TransUnet

This is a network for Martian rock segmentation.

1. Prepare pre-trained ViT models

   Get models and training parameters.

2. Prepare data

   Please experiment with the data in the dataset, using ZhurongRock as an example.

3. Environment

   Please prepare an environment with python=3.7(conda create -n envir python=3.7.12), and then use the command "pip install -r requirements.txt" for the dependencies.

4. Train/Test

   Run the train script on ZhurongRock dataset.