Skip to content

README.md

Latest commit

 

History

History
51 lines (43 loc) · 2.18 KB

README.md

File metadata and controls

51 lines (43 loc) · 2.18 KB

Virtual Full-stack Scanning of Brain MRI via Imputing Any Quantised Code

by Yicheng Wu*+, Tao Song+, Zhonghua Wu, Jin Ye, Zongyuan Ge, Wenjia Bai, Zhaolin Chen, and Jianfei Cai.

News

<20.03.2026> We released the codes;

<21.02.2026> The paper is accepted by CVPR 2026;

Introduction

This repository is for our paper: "Virtual Full-stack Scanning of Brain MRI via Imputing Any Quantised Code". This work aims to generate pseudo full-stack brain MRI scans from any incomplete ones.

Requirements

All experiments in our paper were conducted on eight NVIDIA GeForce 4090 GPUs. This repository is based on PyTorch 2.9.1+cu128 and Python 3.12.12. We further validate this repository via a single NVIDIA GeForce 5090 GPU. The performance is slightly higher than the reported ones. Check ./CodeBrain/models/ for more details.

Usage

  1. Clone this repo.;
git clone https://github.com/ycwu1997/CodeBrain.git
  1. Put the 2D slices into "./MRI/";
  2. Training;
cd ./CodeBrain
# for the two-stage training
sh train.sh
  1. Testing;
cd ./code
# for the imputation inference
python evaluate_grad.py

Citation

If our CodeBrain model is useful for your research, please consider citing:

@InProceedings{wu2026codebrain,
    author    = {Wu, Yicheng and Song, Tao and Wu, Zhonghua and Ye, Jin and Ge, Zongyuan and Bai, Wenjia and Chen, Zhaolin and Cai, Jianfei},
    title     = {Virtual Full-stack Scanning of Brain MRI via Imputing Any Quantised Code},
    booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
    year      = {2026}
}

Acknowledgments:

Our code is adapted from NAFNet, vector-quantize-pytorch, CodeFormer, MMGAN, and PyTorch_GAN. Thanks to these authors for their valuable works, and hope our model can promote the relevant research as well.

Questions

If any questions, feel free to contact me at 'ycwueli@gmail.com'