Mood-Up: Emotion Based Recommendation System with Face and Speech Recognition

Abstract

Author(s): Thenmozhi T, Geerisha Jain

Mood-Up is an emotion detection system that identifies the current mood of the user and attempts to modify/alter the mood by giving the user some things to do based on the recognized emotion. The project aims to build a system that would try to give the users emotion-based recommendations that would help them light up their mood for good. HAAR feature-based cascade classifiers have been used along with OpenCV in Python to recognize faces and evaluate expressions and mood. The face detection is done using the HAAR Cascades. The core to the program is the recommendation system. It is a simple conditional system that would use the different emotions as keywords to present user with various options he/she can choose. The system recognizes if the user is sad, angry or happy and based on the classification, the recommendation system presents options. The idea behind Mood-Up is to help the user and change his mood for good and therefore the system is made very interactive. The use of Pyttsx module allows the program to talk to the user interactively. To make it interactive on both sides, Mood-Up utilizes the power of speech recognition and eliminates typing process completely. Everything the user needs to give in as input or the choices he/she makes are all voice oriented. The program identifies what the user is saying and proceeds with the recommendations.