Vascular-Tree-Reconstruction

KO | EN

gitlite — search

#python #react #ai #typescript #android #machine-learning #linux #nodejs #java #windows #vue #security #mcp #openai

Vascular-Tree-Reconstruction

(★ 31)

Source code for CVPR 19 (Divergence Prior) & 21 (Confluent Vessel Tree) papers.

Vascular-Tree-Reconstruction 최신버젼 다운로드

최종 버전 다운로드 (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# Vascular-Tree-Reconstruction
This repository holds the source code for the papers [Divergence Prior and Vessel-tree Reconstruction](https://openaccess.thecvf.com/content_CVPR_2019/papers/Zhang_Divergence_Prior_and_Vessel-Tree_Reconstruction_CVPR_2019_paper.pdf) and [Confluent Vessel Trees with Accurate Bifurcations](https://openaccess.thecvf.com/content/CVPR2021/papers/Zhang_Confluent_Vessel_Trees_With_Accurate_Bifurcations_CVPR_2021_paper.pdf) appeared in CVPR19 and CVPR21 respectively.

## How to install
The following installation is tested using cmake-3.10.2 under Ubuntu 18.04.3 LTS.
1. c/c++ compiler: gcc-4.8/g++-4.8
2. CUDA: 8.0
3. Dependencies: boost-1.65.0, hdf5-1.10.1, eigen-3.3.4, ITK-4.13.3, flann-1.8.4, cusp-0.5.1
4. If these dependencies are installed in a custom folder, you may need to help cmake find them by explicitly setting some variables in [CMakeLists.txt](CMakeLists.txt), [CMake/FindEigen.cmake](CMake/FindEigen.cmake), [CMake/FindFlann.cmake](CMake/FindFlann.cmake) and [CMake/FindITK.cmake](CMake/FindITK.cmake).
5. Create a new folder *Build* inside the folder where this repository is cloned.
```
mkdir Build
cd Build
```
6. To build the project
```
cmake -DCMAKE_INSTALL_PREFIX=[path to install the project] ..
make -j 8
make install
```

## How to use 
The example script is contained in the [Experiments](Experiments). To obtain the synthetic vessel volume data, use [fetchData.sh](Experiments/fetchData.sh). After downloading the data, simply run [ExprOutter.sh](Experiments/ExprOutter.sh).

Five stages for vessel tree/centerline extraction:

1. Do ObjectenessMasureImageFilter (Frangi filtering)
 
 2. Do NonmaximumSuppressionVolume
  
    Important 
    ```
    --thresholdValue --- this is threshold for plots (ROC, angle error)
    ```
	
 3. Do GenerateNeighborhoodGraph
  
 4. Do LevenbergMarquardtMinimizer (use GPU)
  
    Important 
    ```
    --lambda  --- curvature coefficient
    --beta  --- divergence coefficient
    --tau  --- the hyperparameter in the oriented curvature
    --voxelPhysicalSize  --- defines allowed error in ||l_p - p||_+
    ```
 
 5. Do GenerateTreeTopology
  
    Important 
    ```
    --directedLabel   --- a flag used to decide whether to use minimum arborescence or MST
    --root   --- root index needs to be specified if using the minimum arborescence
    ```
	
	Code for the minimum arborescence: https://github.com/atofigh/edmonds-alg.git

Vascular-Tree-Reconstruction

Vascular-Tree-Reconstruction 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Vascular-Tree-Reconstruction

Vascular-Tree-Reconstruction 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Download README