Real-Time-Emotion-Recognition

Web Application
Desktop Application
Angry
Disgust
Fear
Happy
Neutral
Sad
Surprise

Inspiration

The identification system has undergone great development since its invention. Related techniques are widely utilized and have many applications, such as the robot obstacle avoidance, the robotic arm system learning from human demonstrations, the recognition of handwritten characters, the classification of cats and dogs, the classification of flowers, etc.
Emotion recognition plays a significant role in recognizing one’s effect and in turn, helps in building meaningful and responsive Human-Computer Interface (HCI), ATM Security, Lie Detection, Face Detection in Interviews, etc.

What it does

This Project takes Real-Time images (i.e. from user web-camera) and predicts its emotion simultaneously. It gives seven basic emotion (Happy, Angry, Sad, Fear, Disgust, Surprise and Neutral).

How we built it

Built using the Python programming language.
Algorithm for the model is Convolutional Neural Networks (CNNs).
For Web Application: Flask is used and Desktop Application: PyQt5 is used, both are python based frameworks.

Challenges we ran into

The Challenge to build-up this project was to have better accuracy of the CNN model.

Accomplishments that we're proud of

Got Accuracy up to 91%. Accuracy is increased due to Google Colab Platform that provides inbuilt GPU and CUDA features.

What we learned

Convolutional Neural Networks (CNNs)
Flask
PyQt5.

What's next for Real-Time-Emotion-Recognition

The future scope of this project:

Can Predicts Depression Level using Emotion Recognition.
In the future, input for the model will be 3D format so, this project can be used for 3D input images.