DevMesh

Game Character AI Facial Detection Leveraging OpenVino and DeepLink in Unity

Peter Cross

Peter Cross

Gilbert, Arizona

0 0
  • 0 Collaborators

Unity project which showcases the use of OpenVino and Intel DeepLink to do multiple live facial inferences on in-game characters ...learn more

Project status: Under Development

Game Development, Artificial Intelligence, Graphics and Media

Intel Technologies
Intel Iris Xe, Intel Deep Link, Intel Integrated Graphics, OpenVINO

Overview / Usage

<< Return to GDC Showcase 2021 Group

This AI inferencing project in Unity which showcases OpenVino and Intel DeepLink technologies demonstrates the power of machine learning models applied to gaming characters as it relates to facial detection and characterization. Because OpenVino supports many ML models, I believe this technology can be leveraged for multiple use cases, including providing a method for vision impaired gamers to have access to gaming experiences they would not have otherwise. Additionally, it could be combined with VR to do life-like sensitivity training for police or military forces.

Methodology / Approach

For this project, I used the Unity engine as foundation, and built wrappers to several OpenVino inference models related to face detection (including detecting age/gender, emotional states, facial landmarks, etc).

I leveraged Intel DeepLink to provide ability for user to simultaneously target inferences on the CPU, integrated graphics (Xe), and discrete GPU (Xe Max) running on an Asus Vivobook Tigerlake enabled laptop.

Technologies Used

Unity, OpenVino, Model Zoo ML models, Intel DeepLink, i+i, Xe Max

Comments (0)