Skip to content

A deep learning model for skull reconstruction task.

Notifications You must be signed in to change notification settings

guitared/CraNeXt

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

16 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

CraNeXt

CraNeXt is a deep learning model for skull reconstruction tasks. The input is a binary voxel is a defective skull, and the output is a binary voxel representing a complete skull.

Usage Example

Open In Colab

For a detailed demonstration of how to use CraNeXt, please refer to our example.ipynb notebook. This notebook provides step-by-step instructions and code examples for:

  • Downloading sample SkullBreak dataset and preparing input data
  • Using the CraNeXt model with pretrained weights on SkullBreak
  • Running skull reconstruction and performing evaluation
  • Visualizing reconstruction result

You can run the notebook directly in Google Colab by clicking the Colab badge above or locally.

CraNeXt architecture

Requirements

Use pip to install the requirements as follows:

!pip install -r requirements.txt

Citation

Please refer to our full manuscript in IEEE Access. If you use the model, you can cite it with the following bibtex.

@article {CraNeXt,
  author    = { Kesornsri, Thathapatt and Asawalertsak, Napasara and Tantisereepatana, Natdanai and Manowongpichate, Pornnapas and Lohwongwatana, Boonrat and Puncreobutr, Chedtha and Achakulvisut, Titipat and Vateekul, Peerapon },
  journal   = { IEEE Access }, 
  title     = { CraNeXt: Automatic Reconstruction of Skull Implants With Skull Categorization Technique }, 
  year      = { 2024 },
  volume    = { 12} ,
  pages     = { 84907--84922 },
  keywords  = { Skull;Implants;Image reconstruction;Shape measurement;Three-dimensional displays;Computer architecture;Computational modeling;Skull reconstruction;deep learning;skull categorization;autoimplant;volumetric shape completion },
  doi       = { 10.1109/ACCESS.2024.3415173 },
  url       = { https://doi.org/10.1109/access.2024.3415173 }
}