ICE-GAN

[our code for model design is released]

ICE-GAN: ICE-GAN: Identity-aware and Capsule-Enhanced GAN with Graph-based Reasoning for Micro-Expression Recognition and Synthesis. (Arxiv version)

Author: Jianhui Yu, Chaoyi Zhang, Yang Song, Weidong Cai

Model Architecture

Requirements

• Python >=3.6
• Pytorch >= 1.4
• Packages: tqdm, sklearn

Dataset

• The links for the data we use are provided below (Please contact the corresponding authors to obtain the data):
  1. SMIC dataset
  2. CASME II dataset
  3. SAMM dataset

Code

This repo contains Pytorch implementation of the following modules:

• training log file for all splits;
• model/discriminator.py: model architecture for capsule-based discriminator;
• model/generator.py: model architecture for Unet-based generator with graph reasoning module (GRM);
• ...

Performance

Citation

If you find our data or project useful in your research, please cite:

@InProceedings{yu2020ice,
    title     = {ICE-GAN: Identity-aware and Capsule-Enhanced GAN with Graph-based Reasoning for Micro-Expression Recognition and Synthesis},
    author    = {Yu, Jianhui and Zhang, Chaoyi and Song, Yang and Cai, Weidong},
    booktitle = {International Joint Conference on Neural Networks (IJCNN)},
    year      = {2021}

Acknowledgement

Our code borrows a lot from:

• pix2pix
• Pytorch-CapsuleNet