Deploying ML model with Flask

The emotion classification model /flasksite/ml_model/models/mini_XCEPTION_AffectNet_stratified_with_loss_.60-0.70.hdf5 was developed in the spring 2020 as the requirement for the final project for CS5750 Machine Learning class at Saint Louis University. The model was trained on FER-2013 [5] and AffectNet [3] datasets.

The goal of this project was to deploy the model on the web using Flask framework, and also make it available via an API. The API documentation can be accessed here.

Later, a sudoku solver was added to this project, which is also available in the web application and via an API.

References:

1. Arriaga, O et al. (2017). Face detection and emotion classification. Link
2. Kumar, A. (2018). Demonstration of Facial Emotion Recognition on Real Time Video Using CNN : Python & Keras. Link
3. Mahoor, M. (2017). AffectNet. Link
4. Schafer, C. (2018). Flask Tutorial. Link
5. Wolfram Data Repository (2017). FER-2013. Link