This is the Pytorch implementation of CDERNet proposed in paper 'DERNet: Driver Emotion Recognition Using On-board Camera' by Dingyu Wang, Shaocheng Jia, Xin Pei, Chunyang Han and Danya Yao.

python 3.8.12
pytorch 1.10.0
numpy 1.19.2
torchvision 0.8.1
matplotlib 3.3.2
opencv-python 4.4.0
dlib 19.8.1
face-recognition 1.3.0
scikit-image 0.17.2
scikit-learn 0.23.2

Download DEFE dataset, and make sure it have a structure like following (the first character of file name represents its label):

-datasets/DEFE/raw_data/
  train/
    0-xxx.jpg
    1-xxx.jpg
    2-xxx.jpg
    ...
  test/
    0-xxx.jpg
    1-xxx.jpg
    2-xxx.jpg
    ...

python facedetection.py

train.py --embedding_size 256 --alpha 0.7