Introduction

Facial expression recognition based on Resnet18 and convolutional Autoencoder. A real time facial expression detector based on YOLOv5.

Architecture

Requirement

• Python 3.x
• Pytorch
• Linux
• A CPU or a GPU with CUDA

Download dataset and Pretrained model

Download dataset

How to use it?

• Train FER model: cd ./FER/; python train.py or bash train.sh
• Predict FER model: python predict.pyor bash predict.sh
• Test the detector : cd ..; python detect.pyor bash detect.sh

Files and folders in this repo

• ./FER/ -> Train and evaluate FER models.
• ./models_fer/resnet_autoencoder.py -> The architecture of our model.
• ./detect.py-> The FER detector.
• ./video/-> Put video here which you want to test.

References

https://github.com/WuJie1010/Facial-Expression-Recognition.Pytorch

https://github.com/deepcam-cn/yolov5-face