View Code? Open in Web Editor
NEW
Facial Expression Recognition with TensorFlow
Jupyter Notebook 92.64%
Python 6.38%
CSS 0.98%
affective-computing-project's Introduction
Affective Computing - Final Project
- TASK: Facial expression recognition from images
- The entire model training is visible in the Notebook
- In the end the custom architecture was a way to go
- Transfer learning approach got significantly worse results
- Download the model HERE and put it into the MODELS folder
- You can test the functionality of the trained model via simple frontend interface
- To run the app:
- You can see them all in the TESTED_IMAGES folder
- Naming convention is as follows: predicted_actual.png
angry_angry
![](tested_images/angry_angry.png)
happy_happy
![](tested_images/happy_happy.png)
fear_fear
![](tested_images/fear_fear.png)
affective-computing-project's People
Contributors
Stargazers
Watchers