Emotion Tracer with Ellie

Angry face detection
Happy face detection
Chatbot
App homepage
Chatbot

Inspiration

A lot of students do not stay motivated or remain uninterested during the class. It is due to a lack of interest and also depends on the teaching style of the faculty. We wanted to resolve that issue by creating a platform that will resolve this issue.

What it does

The app detects the faces of the students of the class at certain intervals and finds the emotion associated with the face. It will act as a continuous mood tracker for each student throughout the class. Then, at the end of class, it'll find the average class emotion such as happiness, sadness, frustration, anger, and neutrality. If the maximum students of the class are unhappy then it'll inform the class instructor to look into the matter. There's a chatbot Ellie, which can help out students with their mental health issues and problems.

How we built it

We have used Android Studio to build the app, MLKit to detect faces, TensorFlow to train the emotion detection model, Chatterbot to create the chatbot. Flask is used to give an interface to the chatbot.

Challenges we ran into

Due to the unavailability of GCP credits in our region, we faced difficulties to do emotion recognition without Vision API. Also, we ran into issues while hosting the chatbot in Heroku.

Accomplishments that we're proud of

Our android app can detect emotions from faces, and the chatbot can solve real issues.

What we learned

We learned to use Chatterbot to create a chatbot. Also, we understood how to use MLKit in Android studio.

What's next for Emotion Tracer with Ellie

Furthermore, we have planned to introduce a report regarding each student's mood in certain classes, which will help parents to know the student's interest in a particular subject. It'll help students to select future career options.