Camera Exposure Control via Region Extension

The patent extends the AE target region from a two-dimensional face detection box to a three-dimensional spatial volume by projecting the face position onto the optical axis and creating a depth-based exposure region. This allows the exposure calculation to incorporate not only the face area but also the surrounding spatial context in the depth direction, thereby increasing the effective area for exposure calculation while maintaining focus on the subject.

The patent divides the exposure calculation into multiple regions: the face detection box, the extended AE target region in the optical axis direction, and the subject region. By segmenting the exposure calculation into these distinct zones with different weighting, the system can maintain precise exposure on the face while incorporating additional spatial information from the extended region to improve overall exposure stability.

The patent introduces specific parameters for the extended region including the extension distance from the face detection box along the optical axis, and weighting coefficients for different regions. By parameterizing the extension distance and weighting factors, the system can adjust the exposure region size and influence dynamically without fundamentally changing the control architecture, thus managing complexity through configurable parameters rather than structural complexity.

The patent uses the face detection box as an intermediary to define the starting point for extending the AE target region. Rather than directly calculating complex three-dimensional exposure volumes from scratch, the system uses the existing face detection results as a basis and extends outward from there, simplifying the overall calculation process by building upon an intermediate detection result.