Camera ROI Tracking Using Multi-Detector Occlusion Recovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Electronic devices fail to maintain the region of interest (ROI) at the center of the display screen when the subject, such as a face, is partially covered, leading to deactivation of zooming/panning functions during image capturing.
Innovation Solution
The electronic device employs multiple detectors to detect and estimate ROIs, updating the position of the preview region based on matching and non-matching ROI detections across images, ensuring the ROI remains centered on the screen.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a single detector is used to detect ROI, then the device complexity is low, but the reliability of ROI detection deteriorates when the subject is partially covered
Solution Approach 1:
The detection system is segmented into multiple specialized detectors: a first detector for detecting faces and a second detector for detecting bodies. This segmentation allows each detector to specialize in specific features, improving overall detection reliability when the subject is partially covered while maintaining manageable complexity through functional division.
Solution Approach 2:
The system implements multi-functionality by using both face detection and body detection capabilities within the same ROI tracking system. This allows the system to adaptively switch between detection modes depending on which features are visible, enhancing reliability without requiring entirely separate systems.
2Reliability
If multiple detectors are used to detect ROI, then the reliability of ROI detection improves, but the device complexity increases
Solution Approach 1:
The detection system is segmented into multiple specialized detectors: a first detector for detecting faces and a second detector for detecting bodies. This segmentation allows each detector to specialize in specific features, improving overall detection reliability when the subject is partially covered while maintaining manageable complexity through functional division.
Solution Approach 2:
The system dynamically adjusts which detector results to use based on detection success. When the first detector fails to detect a face, the system transitions to using the second detector's body detection results, creating a dynamic adaptation mechanism that improves reliability without requiring all detectors to operate simultaneously at full complexity.
3Ease of operation
If ROI detection is discontinued when detection fails, then the device complexity is low, but the ease of operation deteriorates due to loss of zooming/panning functions
Solution Approach 1:
The system performs preliminary actions by detecting bodies in advance when face detection fails. By proactively using body detection results to estimate and maintain ROI positioning, the system preserves zooming and panning functions without waiting for face detection to succeed, thereby maintaining ease of operation.
Solution Approach 2:
The system uses body detection as an intermediary when direct face detection fails. The body detector serves as a mediator that provides alternative information to maintain ROI tracking, allowing the zooming/panning functions to continue operating through this intermediate detection method rather than discontinuing entirely.
4Reliability
If the system switches to body detection when face detection fails, then the reliability of ROI positioning improves, but the measurement precision may deteriorate
Solution Approach 1:
The system dynamically adjusts which detector results to use based on detection success. When the first detector fails to detect a face, the system transitions to using the second detector's body detection results, creating a dynamic adaptation mechanism that improves reliability without requiring all detectors to operate simultaneously at full complexity.
Solution Approach 2:
The system creates a copied or alternative representation of the ROI based on body detection when face detection fails. By generating ROI information from body detection data as a copy or substitute, the system maintains positioning reliability while accepting that the precision may be lower than direct face detection, thereby resolving the contradiction between reliability and precision.
Data Source
AI summary
An electronic device is provided. The electronic device includes a camera module, a display, a plurality of detectors, and at least one processor, and the at least one processor may be configured to detect at least one first region of interest (ROI) in a first image received through the camera module by using a first detector for first ROI detection among the plurality of detectors, detect at least one second ROI in a second image received through the camera module by using a second detector for second ROI detection among the plurality of detectors when failing in detecting at least one first ROI matching the at least one first ROI detected in the first image in the second image by using the first detector, estimate at least one first ROI based on the at least one second ROI, update the estimated at least one first ROI to at least one first ROI when the at least one first ROI detected in the first image matches the estimated at least one first ROI, and change a position of a preview region including the updated at least one first ROI based on a position of the estimated at least one first ROI matching the at least one first ROI detected in the first image.