[IROS24] SwinMTL: Multi-Task Learning with Swin Transformer

A multi-task learning framework designed for simultaneous depth estimation and semantic segmentation using the Swin Transformer architecture.

News

• [30th June] Paper Accepted at the IROS 2024 Conference 🔥🔥🔥

Installation

To get started, follow these steps:

0. Only for ROS installation (otherwise skip this part)

   cd catkin_ws/src
   catkin_create_pkg SwinMTL_ROS std_msgs rospy
   cd ..
   catkin_make
   source devel/setup.bash
   cd src/SwinMTL_ROS/src
   git clone https://github.com/PardisTaghavi/SwinMTL.git
   chmod +x inference_ros.py
   mv ./launch/ ./..  

1. Clone the repository:

   git clone https://github.com/PardisTaghavi/SwinMTL.git
   cd SwinMTL

2. Create a conda environment and activate it:

   conda env create --file environment.yml
   conda activate prc

Testing

To run the testing for the project, follow the below steps:

1. Download Pretrained Models:

   • here access the pretrained models.
   • Download the pretrained models you need.

2. Move Pretrained Models:

   • Create a new folder named model_zoo and move the pretrained models into the model_zoo folder you created in the project directory.
   • Refer to testLive.ipynb for testing.

ROS Launch

roslaunch SwinMTL_ROS swinmtl_launch.launch

3D Mapping

Zero-shot Results on the Kitti Dataset

Contributions

• Introduction of a multi-task learning approach for joint depth estimation and semantic segmentation.
• Achievement of state-of-the-art performance on Cityscapes and NYUv2 datasets.
• Utilization of an efficient shared encoder-decoder architecture coupled with novel techniques to enhance accuracy.

We welcome feedback and contributions to the SwinMTL project. Feel free to contact [email protected].

---

Acknowledgments

Special thanks to the authors of the following projects for laying the foundation of this work. Our code relies on:

• GLPDepth
• MIM-Depth-Estimation
• SwinTransformer