sterotto / facial-expression-recognition-svm Goto Github PK

#Facial expression recognition using SVM

Extract face landmarks using Dlib and train a multi-class SVM classifier to recognize facial expressions (emotions).

##Motivation Fer2013 images are not aligned and it's difficult to classify facial expression from it.

The best accuracy for Fer2013 (as I know) is 67%, the author trained a Convolutional Neural Network during several hours in a powerful GPU to obtain this results. Let's try a much simpler (and faster) approach by extracting Face Landmarks and HOG features and feed them to a multi-class SVM classifier.

##Results:

/--------------------------------------------------------
| Features | 7 emotions | 5 emotions | |--------------------------------------------------------| | HoG features | 29.0% | 34.4% | | Face landmarks | 39.2% | 46.9% | | Face landmarks + HOG | 48.2% | 55.0% | |--------------------------------------------------------| | Max training time | 443 sec | 288 sec | --------------------------------------------------------/

While the training time is very short compared to CNN, we lost 19% in accuracy compared to the actual best result that uses CNN.

Note: It's possible to obtain better results by changing parameters. One may implement a hyperparameters search to find the best parameters.

##How to use

1. Extract "fer2013_landmarks+hog.zip" file

2. Install dependencies

pip install Numpy
pip install argparse
pip install sklearn

3. Train model:

python train.py --train=yes

4. Evaluate model

If you have already a pretrained model

python train.py --evaluate=yes

5. Train and evaluate [instead of step 3 and 4]

python train.py --train=yes --evaluate=yes