AR SDK API Architecture

Use the NVIDIA® AR SDK to enable an application to use the face tracking, facial landmark tracking, 3D Body Pose tracking, Eye Contact, and other features of the SDK.

• Working with Features
  • Creating an Instance of a Feature Type
  • Getting and Setting Properties for a Feature Type
  • Setting Up the CUDA Stream
  • Summary of AR SDK Accessor Functions
  • Key Values in the Properties of a Feature Type
  • Getting the Value of a Property of a Feature
  • Setting a Property for a Feature
  • Loading a Feature Instance
  • Running a Feature Instance
  • Resetting a Feature Instance
  • Destroying a Feature Instance
• Working with Image Frames on GPU or CPU Buffers
  • Converting Image Representations to NvCVImage Objects
  • Allocating an NvCVImage Object Buffer
  • Transferring Images Between CPU and GPU Buffers
• Properties for the AR SDK Features
  • Face Tracking Property Values
  • Landmark Tracking Property Values
  • Eye Contact Property Values
  • Body Detection Property Values
  • 3D Body Pose Keypoint Tracking Property Values
  • Facial Expression Estimation Property Values
  • LipSync Property Values
• Using the AR Features
  • Face Detection and Tracking
  • Landmark Detection and Tracking
  • Eye Contact
  • 3D Body Pose Tracking
  • Facial Expression Estimation
  • LipSync
• Using Multiple GPUs
  • Default Behavior in Multi-GPU Environments
  • Selecting the GPU for AR SDK Processing in a Multi-GPU Environment
  • Selecting Different GPUs for Different Tasks
  • Using Multi-Instance GPU (Linux only)