Awesome

T-Mamba

Jing Hao, Lei He, Kuo Feng Hung.

This repository is the official implementation of the T-Mamba: A unified framework with Long-Range Dependency in dual-domain for 2D & 3D Tooth Segmentation .

More experiments are running 🔥

We are conducting more experiments and analysis on 3D CBCT and 2D X-ray images, and will update the whole manuscript. 🏃‍♂️

The code, pre-trained weights, and datasets will be fully available.

Currently, our T-Mamba supports 2D & 3D vision tasks. Welcome to try it for improving your model's performance.

The proposed TED3 dataset is available at: Hugging Face.

If u have any quesitons, pls feel free to drop me via isjinghao@gmail.com.

Install

conda create -n tmamba python=3.9
conda activate tmamba
pip install -r requirements.txt

cd ../causal-conv1d
python setup.py install
cd Vim-main/mamba
python setup.py install

=============================
Requirement specific version:
mamba_ssm==1.0.1
causal_conv1d==1.0.0
=============================

Training

sh train_3d.sh # for 3D
sh train_2d.sh # for 2D

Testing (for evaluations)

sh test_3d.sh # for 3D
sh test_2d.sh # for 2D

Inference

sh infer_3d.sh # for 3D
sh infer_2d.sh # for 2D