Imaging Apparatus Face Detection Shallow Depth of Field

The patent segments the detection field into multiple focus detection points and divides objects into multiple groups based on their distance from the camera. By processing each group separately and determining focus for each group independently, the system can reliably focus on the intended main object even when multiple persons are at different distances, resolving the contradiction between shallow depth of field and focus reliability.

The patent implements feedback by displaying focus detection results to the user and allowing selection of the desired focus target. The system provides visual feedback showing which object will be focused, enabling the user to confirm or adjust the focus selection. This resolves the contradiction by maintaining ease of operation through automatic detection while ensuring focus accuracy through user feedback and selection capability.

The patent applies local quality by performing contrast detection specifically within detected face regions rather than across the entire image. By restricting the focus detection to local areas where faces are identified, the system ensures that the main object (face) is focused even if other areas with higher contrast (background, clothes) exist. This resolves the contradiction between automation and accuracy by localizing the automated detection process to relevant regions.

The patent extends detection into another dimension by performing face detection across multiple focal positions, not just within the current depth of field. By sweeping through different focus distances and detecting faces at each position, the system can identify the main object even when it initially lies outside the depth of field. This resolves the contradiction by expanding the effective detection range without sacrificing detection accuracy.