Face-Aware Sub Image Positioning for Video Calls

The patent applies local quality by differentiating the treatment of different face regions based on their importance. The sub-screen is positioned to avoid overlapping high-importance faces (such as primary subjects) while allowing overlap with low-importance faces. This creates a spatial differentiation where the same sub-screen serves different functions in different regions: it provides information without obscuring important faces while still utilizing screen space effectively.

The patent implements dynamics by making the sub-screen position adaptive rather than fixed. The position control unit dynamically adjusts the sub-screen position based on real-time detection of face regions and their importance degrees. When faces are detected, the sub-screen automatically repositions to avoid overlapping important faces, creating a dynamic response to changing image content that balances information display with face visibility.

The patent applies local quality by establishing a designated safe zone (specific region) where the sub-screen should be positioned. This region is defined as an area that typically does not contain important faces. By confining the sub-screen to this specific local region, the system achieves both goals: faces in other regions remain visible while the sub-screen maintains a consistent, predictable position within its designated zone, providing stability for users.

The patent implements preliminary action by pre-defining the safe zone region and establishing position constraints before the actual display occurs. The position control unit is programmed with advance knowledge of which regions are safe for sub-screen placement. This preliminary setup allows the system to automatically maintain stable positioning without requiring real-time complex calculations, ensuring both face visibility and position stability.