Sticky Projections-A Model-Based Approach to Interactive Shader Lamps Tracking

Area of Science:

• Computer Vision
• Augmented Reality
• Robotics

Background:

• Augmenting physical objects with virtual replications requires precise registration and tracking.
• Previous methods for object registration and tracking were often cumbersome and intrusive.

Purpose of the Study:

• To present a new, mobile method for interactively tracking complex-shaped physical objects.
• To enable "sticking" virtual projections to physical objects using solely the projection itself.

Main Methods:

• A two-stage approach: fast pose initialization using structured light patterns.
• Non-intrusive frame-by-frame tracking based on features detected in the projection.
• Rendering a radiometrically corrected virtual camera view for comparison with captured images and feature triangulation for pose prediction.

Main Results:

• The proposed method accurately and robustly tracks complex objects at interactive rates.
• Detailed experiments evaluated overlay accuracy, demonstrating the system's effectiveness.
• The system is mobile and relies solely on projection for tracking.

Conclusions:

• This novel approach simplifies and enhances the process of tracking physical objects for augmented reality applications.
• The system offers a robust and efficient solution for real-time object tracking without intrusive hardware.
• Future work may involve further optimization for even more complex scenarios and object types.