A Comprehensive Approach To VR-Based Anti-Bullying Education

Objective:

Create a comprehensive virtual reality simulation software for the anti-bullying education of young children to train them to recognize and effectively resolve bullying incidents.

Features:

Realistic dialogue, tracking eye-contact, monitoring social distancing, and facial expression tracking.

Demo Video

Materials

Meta Quest Pro

Unity Engine

Language

Research Paper

Email Me!

SDKS and Libraries

ChatGPT. OpenAI-Unity.

The bully-victim dialogue is dynamically driven by the ChatGPT Turbo model using the OpenAI-Unity package.
The bully initiates the conversation, and following the user's interaction with the bully, I prompt ChatGPT to determine whether the user's sentence is positive, negative, or neutral.
I then instruct ChatGPT to generate a child-like response corresponding to that connotation to communicate with the user.

Microsoft Speech SDK. Text-to-Speech.

Utilizing Microsoft's Speech SDK enables seamless integration of Text-to-Speech functionality into the software.
ChatGPT's generated positive, negative, or neutral remark will be synthesized .

Meta Movement SDK. Face and Eye Tracking.

I created visual representations of facial expressions for seven universal emotions: happiness, sadness, fear, disgust, anger, surprise, and neutrality. I did this by altering an avatar's BlendShapes
The Meta's Movement SDK enables the script to monitor user facial movements corresponding to these expressions during gameplay.
When a user's facial expression matches or closely resembles any of our predefined expressions, an emoticon representing that expression is triggered during gameplay.