WFOV Image Correction for Face Detection Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Face detection technologies are not effectively applied to wide field of view (WFOV) imaging systems due to geometrical distortions, which conventional methods like rectangular classifiers or integral image techniques cannot conveniently handle.
Innovation Solution
A WFOV correction engine processes raw image data using rectilinear and cylindrical projections to correct distortions, allowing face detection by applying different projections to center and outer pixels, enabling the use of cascades of regular or modified object classifiers for detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If conventional face detection methods (rectangular classifiers or integral image techniques) are applied to WFOV imaging systems, then the detection process is simple, but the detection accuracy deteriorates due to geometrical distortions at different positions in the image
Solution Approach 1:
The image is divided into multiple regions (center region and peripheral regions) based on their distortion characteristics. Different classification approaches are applied to different regions: conventional rectangular classifiers for the center region and distorted-region classifiers for peripheral regions. This segmentation allows each region to be processed with the most appropriate method, resolving the contradiction between simplicity and accuracy.
Solution Approach 2:
Different classification algorithms are applied to different spatial locations within the image. The center region uses standard rectangular classifiers while peripheral regions use specialized distorted-region classifiers that account for geometric transformations. This local differentiation ensures high detection accuracy across the entire image while maintaining operational simplicity through automated region-based processing.
2Area of stationary object
If WFOV imaging systems are used to capture wider scenes, then the field of view increases, but geometrical distortions increase causing face detection to fail
Solution Approach 1:
The classification system dynamically adapts to the distortion characteristics at different positions in the WFOV image. By using position-dependent classifiers that are selected or adjusted based on the spatial location of detected faces, the system maintains reliable detection across the entire wide field of view, resolving the contradiction between capturing wider scenes and maintaining detection reliability.
3Adaptability or versatility
If faces near the edge of WFOV images are detected, then the coverage of face detection is improved, but the geometrical distortions become more severe making detection difficult
Solution Approach 1:
The system applies specialized distorted-region classifiers to peripheral areas of the image where geometric distortions are most severe. These classifiers are specifically designed to account for the distortion patterns at image edges, enabling reliable face detection in regions that would otherwise be undetectable with conventional methods, thus expanding coverage without proportionally increasing difficulty.
Data Source
AI summary
An image acquisition device having a wide field of view includes a lens and image sensor configured to capture an original wide field of view (WFoV) image with a field of view of more than 90°. The device has an object detection engine that includes one or more cascades of object classifiers, e.g., face classifiers. A WFoV correction engine may apply rectilinear and/or cylindrical projections to pixels of the WFoV image, and/or non-linear, rectilinear and/or cylindrical lens elements or lens portions serve to prevent and/or correct distortion within the original WFoV image. One or more objects located within the original and/or distortion-corrected WFoV image is/are detectable by the object detection engine upon application of the one or more cascades of object classifiers.