Models

In this folder you can find all the models that we have trained. The analysis of each model can also be found here.

Content

• Baseline model

  Simple sequential model that can recognize whether someone is happy or not.

• Classification model

  Conv:

  A sequential model that tries to recognize the following 7 emotions.

  1. Neutral

  2. Happy

  3. Surprise

  4. Sadness

  5. Anger

  6. Disgust

  7. Fear

     VGG16:

     A pre-trained model often used to recognize emotions. For more information about this model check this keras page. We use here both the pre-trained version and the model without weights. Which makes us train the model ourselves with our own photos. We are still trying to recognize the same 7 emotions.

• Validation model

  A pre-trained model to use as a validation model of our self-made model.

Installation

To run our models you need to get the data first. Therefore you can have a look at this page.

Sinds you have the data, place it into a folder called data in models. After that you will be abel to run the code of every model.