DCNN-GAN

Code repository for MVA 2019 paper "DCNN-GAN: Reconstructing Realistic Image from fMRI"

Prerequisites

• Linux / macOS
• NVIDIA GPU with CUDA CuDNN
• Python 3

Getting Started

Installation

• Clone this repo

git clone https://github.com/CreeperLin/DCNN-GAN.git
cd DCNN-GAN
git submodule update --init

• Install requirements (using Anaconda is also recommended)

pip3 install -r requirements.txt

fMRI decoder train/test

• Download fMRI on Imagenet datasets

./datasets/download_fmri.sh

• Generate image features for training

python3 ./decode/train_dataloader.py --img_data ./datasets/image_fmri --output ./tmp/feat_data

• Train fMRI decoder and decode

python3 ./decode/decode.py --fmri_data ./datasets/fmri_data --feat_data ./tmp/feat_data --output ./tmp/decoded_feat

DCNN-GAN train/test

• Data Preparation

python3 ./reconstruction/train_dataloader.py --dataset ./datasets/train_dcnn_img --output ./tmp/dcnn_train

• Train DCNN-GAN

python3 ./reconstruction/train.py --DCNN_dataset ./tmp/dcnn_train --pix2pix_dataset ./datasets/train_gan_img

• Test DCNN-GAN

python3 ./reconstruction/test.py --decoded_feat ./tmp/decoded_feat --output ./reconstruction/results

Run the full pipeline (training & reconstruction)

./run_all.sh

Results

The example reconstructed images are listed below:

Citation

@article{Lin2018DCNN-GAN
    author = {Yunfeng, Lin and Jiangbei, Li and Hanjing, Wang",
    title = {DCNN-GAN: Reconstructing Realistic Image from fMRI},
    year = {2018},
    howpublished={\url{https://github.com/CreeperLin/DCNN-GAN}}
}

Acknowledgements

The GAN model is based on the pytorch implementation of pix2pix.

The fMRI data is obtained using the datasets from Generic Object Decoding.