LUX: Smart Mirror with Sentiment Analysis for Mental Comfort

Area of Science:

• Artificial Intelligence
• Natural Language Processing
• Human-Computer Interaction

Background:

• Growing interest in mental health necessitates advanced sentiment analysis tools.
• In-home AI devices like smart mirrors offer scalable solutions for mental comfort.
• Existing Korean natural language processing (NLP) lacks comprehensive emotion datasets for everyday conversations.

Purpose of the Study:

• To develop a smart mirror system (LUX) capable of performing sentiment analysis on Korean text.
• To address the challenges in Korean NLP due to linguistic differences from English.
• To provide interactive emotional support through a smart mirror interface.

Main Methods:

• Implementation of a deep learning model analyzing GRU, LSTM, CNN, Bi-LSTM, and Bi-GRU networks for sentiment analysis.
• Development of four emotional labels: anger, sadness, neutral, and happiness.
• Integration of three interactive response types: reciting wise sayings, playing music, and sympathizing.

Main Results:

• The LUX smart mirror successfully performs Korean text sentiment analysis using advanced deep learning models.
• The system is equipped to identify four distinct emotional states.
• Personalized interactive responses are generated based on detected emotions.

Conclusions:

• The LUX smart mirror demonstrates a viable approach to in-home mental well-being support through Korean sentiment analysis.
• The system's deep learning models effectively process Korean emotional nuances.
• LUX integrates essential smart mirror functions with advanced AI-driven emotional support capabilities.