Emotion recognition from face and speech with Python

Dataset

fer2013 is the dataset used for training the model.

How to install

Create and activate a virtual environment

It is recommended to work in a virtual environment, that can be created and activated with Anaconda using the following commands:

conda create -n emotion_detection python=3.4
source activate emotion_detection

Install dependencies

conda install scikit-learn
conda install -c menpo opencv3=3.1.0
pip install --upgrade keras
conda install pandas
conda install h5py
pip install SpeechRecognition
apt-get install portaudio19-dev
pip install pyaudio

Configure Keras

Create the file ~/.keras/keras.json with the following content:

{
    "image_dim_ordering": "th",
    "epsilon": 1e-07,
    "floatx": "float32",
    "backend": "theano"
}

Clone the repository

git clone https://github.com/SergioML9/emotion_recogniser.git

Usage

Emotion detection from face analysis

Bash option

Run face_analyser/run.py script, and the emotion detection process will start, showing the output in the terminal. Press q in order to stop the detection.

GUI option

Run face_analyser/gui.py script, and a very simple GUI will be shown. The GUI has five buttons:

1. Start emotion detection: starts the emotion recognition, printing the outputs in the terminal.
2. Train model: starts the model training with the data specified at configuration/data_settings.py.
3. Evaluate model: evaluates the model with the data specified at configuration/data_settings.py.
4. Get data from dataset: converts the data from the csv dataset to npy files.
5. Exit: closes the application.

Emotion detection from speech analysis

Bash option

Run speech_analyser/run.py script, and the emotion detection process will start, showing the output in the terminal.