Enhancing Emotional Analysis

Project Roadmap

Current Limitations

Limited Emotional Analysis: Relies on raw data of heart rate & eye gaze dwell time
Difficulty in Inferring Emotions: Difficult to infer the drug abuser’s emotions from raw sensor signals
Linear and Static VR: Lacking in realism & responsiveness

Problem Statement

How can AI and sensor technologies enhance emotional analysis in VR-based CBT in real-time?

Design Objectives

AI Model: Predict the drug abuser’s emotional state in real-time
Physiological Sensor(s): Capture physiological and image data for accuracy and comfort
Unreal Engine 5 VR Simulations: More realistic and interactive scenarios for enhanced realism

Solution: Integrated System

❤️‍🩹Our solution combines AI, ECG sensor & VR into an integrated system. This system enables real-time emotional monitoring of the drug abuser, reduces cognitive workload, and personalises rehabilitation, making VR-based CBT for drug addiction recovery more effective.

Key Innovations:
- High accuracy emotion prediction powered by AI
- Immersive VR environments simulate life-like scenarios for deeper emotional engagement
- AI-driven emotional insights ease therapists’ workload

Impact & Applications:
- Real-time emotion detection
- Intuitive wearable sensors
- Enhances therapy effectiveness through realistic, engaging environments

Key Features of the System:

AI: AI-Driven Emotion Analysis – Predicts emotional states every 5-30 secs

Sensors: Physiological Sensors – ECG tracks stress & valence (99% accuracy)

VR: Immersive VR – MOCAP & LiveLink for realistic animations

In partnership with :

Supported by :

Acknowledgements

We would like to express our heartfelt gratitude to Assistant Professor Sun Zhu and Oka Kurniawan, Capstone Mentors, for their unwavering encouragement, dedication, and guidance throughout the project.

We sincerely thank our supervisors at the Home Team Science and Technology Agency, Derek Chong, Adrian Oh and Shawn, for their unwavering support, continuous guidance, and vital feedback throughout this project.

We extend our heartfelt thanks to our partners from Singapore Prisons, Eng Hao, Renee Li and Quentin Lee for their invaluable insights and their crucial feedback throughout the development of our systems.

We extend our heartfelt thanks to our Centre for Writing and Rhetoric mentor, Pek Har Belinda Seet, for her clear instructions, constructive critique, and ongoing support during this project.