Learnbot is a small low-cost robot designed to develop computational thinking in kids of age 10 and above. Presently, Learnbot can be programmed to show emotions via a display. However, it does not have the ability to recognize human emotions. Having this ability would help improve the Human-Robot Interaction(HRI) functions in Learnbot.

This project aims at applying state-of-the-art techniques to create a fast, accurate and robust emotion recognizer. It would be able to recognize 5 basic emotions: Happiness, Sadness, Anger, Surprise, Neutral, based on facial expressions. This involves training a Convolutional Neural Network based classifier.

The ability to recognize emotions would be added as a new component in Learnbot. Other components would be able to access the data generated by this component through an interface.



Sayali Deshpande


  • Ibarbech
  • Pilar Bachiller