Spatial Affordance-Aware Affine Transformation Between Heterogeneous Spaces for Mixed Reality Remote Collaboration

Area of Science:

• Computer Vision
• Human-Computer Interaction
• Robotics

Background:

• Existing spatial mapping methods struggle with continuous correspondence between dissimilar physical environments in shared Mixed Reality (MR).
• A gap exists in precise object alignment and seamless locomotion across heterogeneous spaces.
• Current approaches often rely on physical objects and walkable areas, limiting adaptability.

Purpose of the Study:

• To propose a spatial affordance-aware affine transformation method for continuous multi-object matching in shared MR spaces.
• To enable seamless mapping between heterogeneous physical environments.
• To support precise object alignment and locomotion in shared MR.

Main Methods:

• Structurally segmenting heterogeneous spaces into interaction zones.
• Constructing affine patches based on object adjacency and facing configuration.
• Utilizing spatial affordances for transformation between dissimilar spaces.

Main Results:

• The proposed method achieves broader spatial alignment compared to conventional grid-based methods.
• Demonstrates richer multi-object matching capabilities.
• Evaluated using a dataset of paired dissimilar spaces.

Conclusions:

• The developed method serves as an effective mapping framework for shared MR environments.
• It ensures semantic continuity and structural coherence across diverse real-world spaces.
• Facilitates robust object correspondence and locomotion in heterogeneous MR settings.