werywjw/SEP-CVDL
GiMeFive: Towards Interpretable Facial Emotion Classification ππ²ππ‘π€’π¨ (PyTorch Implementation)
GiMeFiveβ: Towards Interpretable Facial Emotion Classification
Abstract
Deep convolutional neural networks have been shown to successfully recognize facial emotions for the past years in the realm of computer vision. However, the existing detection approaches are not always reliable or explainable, we here propose our model GiMeFive with interpretations, i.e., via layer activations and gradient-weighted class activation mapping. We compare against the state-of-the-art methods to classify the six facial emotions. Empirical results show that our model outperforms the previous methods in terms of accuracy on two Facial Emotion Recognition (FER) benchmarks and our aggregated FER GiMeFive. Furthermore, we explain our work in real-world image and video examples, as well as real-time live streams.
Overview
Results
Requirements
A detailed list of requirements can be found in requirements.txt.
pip install -r requirements.txt
python3 -m pip install -r requirements.txt
How to run the script to get the CSV file of classification scores?
Please change the filepath to the image folder that you would like to try. π
- Option 1οΈβ£: Run the following command:
python3 script_csv.py 'data/valid'
- Option 2οΈβ£: Use shell:
./script_csv.sh
How to run the script to generate an XAI video from a given video path or a camera stream? π₯Ή πΉ
Run one of the following commands:
python3 script_video.py -h
python3 script_video.py -s camera
python3 script_video.py -s video -i 'video/video.mp4'
How do our videos look like?
How to test the live demo with camera (old version)?
Run the following command:
python3 eval_livecam.py
Exit via control + C.
Please kindly cite if you find this repository helpful πΈ
@article{wang2024gimefive,
title={GiMeFive: Towards Interpretable Facial Emotion Classification},
author={Wang, Jiawen and Kawka, Leah},
journal={arXiv preprint arXiv:2402.15662},
year={2024}
}
Q&A
Where is our paper (final report) π?
The PDF Version is here.
Where are our presentation slides π₯?
On goolge docs.
The shared link is available here.
The PDF Version is here.
Where is our proposal (preliminary report) π?
The PDF Version is here.
What are the emotion labels?
π happiness (0)
π² surprise (1)
π sadness (2)
π‘ anger (3)
π€’ disgust (4)
π¨ fear (5)
How to generate a requirements.txt?
pipreqs /Users/wery/Desktop/SEP-CVDL
pip3 freeze > requirements.txt
How to resolve the conflicts?
git status
See what is asked to be modified, then:
git add .
git pull
git config pull.rebase false
git pull
git commit
git push
How to coauthor?
git add .
git commit -m " <bababa>
Co-authored-by: leahkawka <leahkawka@gmail.com>
Co-authored-by: werywjw <werywjw@gmail.com>"
git push origin main
How to convert .ipynb to .py?
colab-convert livecam.ipynb livecam.py -nc -rm -o
How to convert .mp4 to .gif?
Go to ezgif webpage.
How to create submodules?
git submodule add https://github.com/werywjw/data.git
How to create the virtual environment?
source /Users/wery/venv/bin/activate



