Live Classroom Camera Control via Region-Based Tracking

The patent introduces an intermediary region (first region and second region) between the target object and the coordinate point. Instead of directly tracking the coordinate point, the system tracks the target object's position within the intermediary region and adjusts the camera based on the object's distribution in this region. This intermediary mechanism smooths out position fluctuations and reduces direct coordinate point tracking that causes jitter.

The patent performs preliminary determination of the first region and second region before coordinate point tracking. By pre-establishing these regions based on the target object's initial position and size, the system creates a buffer zone that allows for smoother camera adjustments. The camera angle is adjusted based on the target's distribution within these pre-established regions rather than making continuous adjustments to maintain exact coordinate point alignment.

The patent applies partial action by adjusting the camera angle based on the target object's distribution within the first region rather than making full adjustments to maintain exact coordinate point alignment. When the target is within the first region, the camera makes smaller adjustments compared to traditional coordinate point tracking, reducing the frequency and magnitude of camera movements while still keeping the target in view.

By pre-establishing the first and second regions before tracking begins, the system creates a buffer zone that allows for smoother, less frequent camera adjustments. The regions are determined based on the target's initial position and size, allowing the camera to adapt to position changes within these regions without making continuous adjustments, thereby reducing visual jitter while maintaining target visibility.

The patent segments the detection process by dividing the image into multiple detection points along the target object's contour. Instead of treating the entire target as a single entity, the system identifies multiple discrete detection points (e.g., leftmost, rightmost, topmost, bottommost points) along the contour. This segmentation allows for more accurate position and size determination while keeping each individual detection point calculation relatively simple.

The patent transitions from two-dimensional contour detection to three-dimensional spatial reasoning by establishing the first and second regions in 3D space and determining the target's distribution within these regions. This dimensional transformation allows the system to calculate camera adjustment angles more accurately by considering the target's position, size, and orientation in three-dimensional space, while the underlying contour detection remains a relatively simple 2D image processing task.