chenhbo / FUNSR

MIT License
7 stars 2 forks source link

FUNSR: Neural Implicit Surface Reconstruction of Freehand 3D Ultrasound Volume with Geometric Constraint


Paper | Arxiv

The official implementation code for Medical Image Analysis paper: Neural Implicit Surface Reconstruction of Freehand 3D Ultrasound Volume with Geometric Constraint

FUNSR is a Freehand 3D Ultrasound Neural Surface Reconstruction method built on neural implicit signed distance functions for freehand 3D ultrasound imaging and surface rendering.

 Method Overview

Demo

Framework

Usage

Our code is implemented in Python 3.8, PyTorch 1.12.1 and CUDA 11.6

Install Dependencies

conda create -n funsr python=3.8
conda activate funsr
conda install pytorch==1.12.1 torchvision==0.13.1 torchaudio==0.12.1 cudatoolkit=11.6 -c pytorch -c conda-forge
pip install tqdm pyhocon==0.3.57 trimesh PyMCubes scipy
pip install matplotlib

Data Preparation

Run FUNSR

python run_normalizedSpace.py --gpu 0 --conf confs/conf.conf --dataname case000070.nii_ds  --dir case000070.nii_ds

Contact

For any queries, please contact chenhb[at]shanghaitech.edu.cn.


References

The reference codes are from the following links. We appreciate all the contributors.