Face emotions recognize
Solving Face emotions recognise problen with keras-built nn.
Dataset : FER 2013 (https://www.kaggle.com/datasets/msambare/fer2013)
Referencies: Haar cascade classifiers (https://github.com/opencv/opencv/tree/master/data/haarcascades)
Stack: Keras, TF, numpy, REST-API keras, flask
I used 3 different Haar classifiers, because they really depends on brightness, contrast and gamma of original picture. So you can choose your favorite one, or just use it as is.
How to use:
- Copy repository from github using
git clone https://github.com/FitzTata/face-emotion-recognize
and install some requirements using
pip install -r requirements.txt
- Run file run_server.py
- Put photos in ./test_pics
- Edit client.py path to photo like that
ln9: IMAGE_PATH = "../test_pics/test1.jpg" to "../test_pics/your_pic.jpg"
- Run file client.py
- You and your photo are awesome! You will see result in about 3 seconds (depends on your pc resources).
Example:
