© Design Innovation All Rights Reserved
Meet the Team
Introduction: EmoBot:- Multilingual Emotion Detection Chatbot from Human Speech
To design a multilingual speech based emotion detection based
chatbot. Chatbot uses basic signal processing methods and psychological
study observations to attempt to improve Human Machine Interactions.
• Recognition of Speech in multiple languages viz: Hindi, Punjabi, English.
• Recognizing of Emotions separately from Audio and Video.
• Recognizing Action Points on human face to recognize six different types of Emotions such as ‘Anger, ‘Disgust’, ‘Happy’, ‘Fear’, ‘Neutral’, ‘Sad’.
• Proposed system accuracy is 29.74% higher in comparison to baseline human accuracy on CREMA-Dataset.
• Validated on Multiple Dataset and new dataset of Hindi, Punjabi and English is Created.
- Name of the Department: UIET, Panjab University
- Name of Project: Design Innovation Centre(DIC)
ChatBots need to be smarter and understand human emotions while communicating with humans. it can be used in any Machine / Robot that need to communicate with humans.