Imaging Device Scene Determination Using Face Detection Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing imaging devices inaccurately determine photographic scenes due to unreliable distance information from autofocus mechanisms, particularly when a deep focus depth is achieved with a short focal length and low aperture, leading to erroneous classification of scenes.
Innovation Solution
An imaging device that evaluates the reliability of distance information using a combination of autofocus mechanism data, focal length, aperture value, and secondary distance information from face detection, enabling accurate determination of photographic scenes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Stability of the object's composition
If hill-climbing autofocus is used to achieve deep focus depth at close range, then the autofocus operation can maintain focus within one meter, but the distance measurement accuracy becomes unreliable and may erroneously indicate close-up scenes
Solution Approach 1:
The system uses the detected face position and face size information as feedback to evaluate the reliability of autofocus distance measurements. When the face detection indicates the subject is closer than the autofocus measurement suggests, the system adjusts the scene determination accordingly, using the feedback from face analysis to correct potentially inaccurate distance-based scene classification.
Solution Approach 2:
Face detection acts as an intermediary mechanism between the autofocus system and scene determination. Instead of directly using autofocus distance measurements for scene classification, the system introduces face detection as an intermediate step to verify and supplement the distance information, particularly for subjects within one meter where autofocus becomes unreliable.
2Productivity
If autofocus distance information is used directly for scene determination, then the process is simple and fast, but erroneous scene classification occurs when focus depth extends beyond the subject distance
Solution Approach 1:
The system performs preliminary face detection and face position/size measurement before final scene determination. This preliminary action using face analysis allows the system to pre-evaluate the reliability of autofocus distance measurements, enabling faster and more accurate scene classification by having the verification step ready in advance.
Solution Approach 2:
The system changes the parameters used for scene determination based on the reliability evaluation. When face detection indicates low reliability of autofocus measurements (such as when the subject is within one meter), the system switches to using face-based distance estimates or manual input instead of autofocus data, thereby maintaining accuracy while preserving processing efficiency.
Data Source
AI summary
In an imaging device, a photographic scene is automatically determined with high accuracy. Distance information is acquired, through a TTL-AF, from an image signal produced by a CCD. Further, a face recognition circuit detects the face of a person in a subject. An AE-AF-AWB value computing section evaluates the reliability of the distance information acquired through the TTL-AF operation, on the basis of a focal length of a lens, an aperture, and a distance to the subject computed from a distance between eyes included in the detected face. A mode-recognition-and-determination-and mode-selection section automatically determines the photographic scene from the distance information acquired through the TTL-AF operation, the reliability of the distance, and an estimated white balance value.