Virtual Eyewear Fit Simulation Using Collision Detection

The patent introduces an intermediary collision detection algorithm between the eyewear model and the face model. This intermediary system uses virtual colliders attached to key points on the eyewear (temple tips, lens edges, bridge) to detect and prevent unrealistic intersections with the face model, thereby achieving accurate fit simulation without requiring complex physical engine calculations.

The patent replaces the traditional mechanical physics-based simulation with a computational geometry approach. Instead of using force-based physical engines to simulate eyewear resting on the face, the system uses mathematical collision detection algorithms that calculate intersections between virtual colliders and the face mesh, providing accurate results more efficiently.

The patent segments the collision detection system into discrete, manageable components: virtual colliders are attached to specific key points on the eyewear model (temple tips, lens edges, bridge), and the face model is represented as a mesh with vertices. This segmentation allows the complex problem to be solved through simple point-to-surface distance calculations rather than full volumetric collision detection.

The patent applies partial collision detection by placing colliders only at the critical contact points of the eyewear (temple tips, lens edges, bridge) rather than modeling the entire eyewear volume. This partial approach provides sufficient accuracy for fit simulation while significantly reducing computational complexity compared to full 3D collision detection.

The patent uses a minimal set of colliders placed only at the critical contact points where the eyewear actually touches the face (temple tips, lens edges, bridge). This partial coverage provides realistic fit simulation while minimizing the number of collision calculations required, thereby reducing computational energy consumption.

The patent creates simplified virtual copies (colliders) of the eyewear's critical contact surfaces rather than using the full complex 3D geometry. These simplified collider representations maintain the essential fit characteristics while requiring far fewer computational resources for collision detection algorithms.