Multi-Angle Video Communication for Consistent Eye-to-Eye Display

The patent divides the camera system into multiple independent cameras positioned at different horizontal angles. Each camera captures video from its specific viewpoint, allowing the system to provide multiple viewing angles simultaneously. This segmentation enables participants to see each other in proper orientation regardless of their viewing angle, achieving the eye-to-eye effect without requiring a single complex adjustable camera.

The patent implements dynamic selection and switching of video feeds based on participant position and viewing angle. The system dynamically determines which camera feed to display to each participant, or synthesizes multiple feeds in real-time. This dynamic adaptation allows the system to maintain consistent eye-to-eye communication效果 as participants move or change viewing angles, without requiring physical camera repositioning.

The patent makes each camera serve multiple functions: capturing video for its specific viewpoint, providing feed for participants at certain angles, and contributing to synthesized views for other participants. This multi-functionality reduces the need for dedicated cameras for each viewing angle, optimizing the camera system while maintaining consistent video display quality across all participants.

The patent creates virtual copies of video feeds through image synthesis techniques. Instead of requiring physical cameras at every possible viewing position, the system synthesizes additional video perspectives from the captured feeds using computer vision and image processing. This copying approach provides consistent eye-to-eye effect for multiple viewing angles without proportionally increasing the physical camera count.

The patent transitions from a single fixed-camera-dimension approach to a multi-dimensional camera array. By positioning cameras at different horizontal angles (adding the angular dimension), the system can serve participants across a range of viewing angles. This dimensional expansion allows the system to adapt to different participant positions without requiring complex adjustable mechanisms on each camera.

The patent introduces image synthesis and selection algorithms as intermediaries between the fixed cameras and the participants. These software intermediaries process the fixed camera feeds, select appropriate views, or synthesize new perspectives, thereby adapting the fixed-camera output to match each participant's viewing angle and position. This intermediary layer provides versatility without requiring physical camera repositioning.