Skip to content

Latest commit

 

History

History
136 lines (112 loc) · 4.21 KB

README.md

File metadata and controls

136 lines (112 loc) · 4.21 KB
Raw

visual-volume

demo

Introduction

Visual Volume is a project designed for visualizing volumetric data. This project leverages web technologies such as Three.js and WebGL to render and display three-dimensional volumetric data from sources like CT scans.

You can view the project live at https://visual-volume.vercel.app.

Purpose

The purpose of this project is to visualize the CT scan data of the ThaumatoAnakalyptor project into a point cloud. Through this tool, we aim to improve the accuracy of data conversion.

This project is an extension based on the GitHub repository: tomhsiao1260/pipeline-visualize.

Features

  • Load and parse NRRD files
  • Render and display volumetric data
  • Support for various visualization modes such as isosurfaces and volume rendering
  • Interactive features allowing users to adjust the view and visualization parameters

Tech stack

  • three.js
  • python

Setup

  1. Clone this repo 
    git clone https://github.com/pocper1/visual-volume.git
  2. Install the dependency 
    pip install numpy torch tifffile scipy open3d opencv-python pynrrd tqdm
  3. Init the web project 
    cd web
npm install

Process guideline

  1. Data collection, download the dataset PHercParis4.volpkg
  2. Transform CT-scan into pt
    • file format: tif->pt
    • location: code/surface_detection.py
  3. Transform into nrrd
    • file format: pt->nrrd
    • location: code/pt_nrrd.py
  4. Use WebGL to show the result
    • location: web/

Process detail

  1. Data collection Note: You need to sign up for the Vesuvius Challenge Data Agreement. Once registered, you will be granted a username and password to access the dataset.

    cd visual-volume
    wget --no-parent -r -nd --user=<userName> --password=<password> -P dataset https://dl.ash2txt.org/full-scrolls/Scroll1/PHercParis4.volpkg/volume_grids/20230205180739/cell_yxz_006_008_004.tif

  2. CT-scan transform into pt write data into /dataset/<tifName>/*

    python code/surface_detection.py dataset/cell_yxz_006_008_004.tif

  3. Transform into nrrd read data from /dataset/<tifName>/* and write data into /web/public/<tifName>/*

    python code/pt_nrrd.py

  4. Use web to show the result

    cd /web
npm run dev

File structure

.
├── code
│   ├── nrrd_tools.py
│   ├── pt_nrrd.py
│   └── surface_detection.py
├── dataset
├── README.md
├── umbilicus.txt
└── web
    ├── index.html
    ├── package.json
    ├── package-lock.json
    ├── public
    │   └── cell_yxz_006_008_004
    │       ├── adjusted_vectors_interp.nrrd
    │       ├── adjusted_vectors.nrrd
    │       ├── blurred_volume.nrrd
    │       ├── first_derivative.nrrd
    │       ├── mask_recto.nrrd
    │       ├── mask_verso.nrrd
    │       ├── origin.nrrd
    │       ├── second_derivative.nrrd
    │       ├── sobel_vectors.nrrd
    │       ├── sobel_vectors_subsampled.nrrd
    │       └── vector_conv.nrrd
    ├── src
    │   ├── css
    │   │   └── main.css
    │   ├── img
    │   │   └── favicon.ico
    │   └── js
    │       ├── config.js
    │       ├── core
    │       │   ├── shaders
    │       │   │   ├── fragmentShader.glsl
    │       │   │   └── vertexShader.glsl
    │       │   ├── textures
    │       │   │   └── cm_viridis.png
    │       │   ├── ViewerCore.js
    │       │   └── VolumeMaterial.js
    │       ├── main.js
    │       └── volume.js
    └── vite.config.js