Wide Field of View Face Detection via Multi-Camera Array
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current facial detection and tracking systems using pan-tilt-zoom cameras struggle to simultaneously detect and track individuals within a wide field of view while providing sufficient resolution for facial recognition, especially when individuals are moving quickly or in vehicles, limiting their ability to cover the entire image field and detect multiple individuals.
Innovation Solution
A system comprising a wide field of view camera for initial detection and one or more narrower field of view cameras arranged in an array, which can obtain higher-resolution images of detected individuals without the need for pan-tilt-zoom mechanisms, utilizing tri-band imaging cameras that combine low band near-IR, high band near-IR, and visible light for feature analysis, and near-IR illuminators for enhanced image capture.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If a single PTZ camera is used to cover a wide field of view, then the coverage area is improved, but the resolution and tracking accuracy deteriorate when individuals are moving quickly
Solution Approach 1:
The system divides the wide field of view into multiple sub-regions, each monitored by a dedicated fixed camera. This segmentation allows each camera to maintain high resolution for its specific zone while collectively covering the entire wide area, resolving the contradiction between coverage and precision.
Solution Approach 2:
The patent transitions from a single camera with mechanical movement (PTZ) to a multi-camera array in a fixed configuration. This dimensional change from one movable camera to multiple stationary cameras eliminates the need for positioning mechanisms while maintaining both wide coverage and high resolution through spatial distribution.
2Measurement precision
If a PTZ camera zooms in to achieve higher resolution, then the facial recognition capability is improved, but the ability to track moving individuals deteriorates
Solution Approach 1:
The system assigns specific high-resolution zones to individual cameras in the array. When an individual moves, the system identifies which camera's zone the individual is in and uses that camera's pre-positioned high-resolution view, eliminating the delay associated with mechanical zooming while maintaining both resolution and tracking speed.
Solution Approach 2:
The patent replaces the mechanical zoom mechanism with a computational approach using multiple fixed cameras. Instead of mechanically adjusting a single camera's focal length, the system selects from multiple pre-positioned cameras, each optimized for specific zones, thereby eliminating mechanical limitations and improving tracking response time.
3Measurement precision
If multiple PTZ cameras are used to cover the entire field of view with sufficient resolution, then the detection capability is improved, but the device complexity increases
Solution Approach 1:
The system segments the monitoring area into discrete zones, each covered by a fixed camera. This segmentation simplifies the overall system architecture compared to multiple PTZ cameras, as each camera remains stationary and requires no complex positioning mechanisms, reducing device complexity while maintaining detection accuracy.
Solution Approach 2:
Instead of using multiple PTZ cameras that can move and adjust, the patent inverts the approach by using multiple fixed cameras with predetermined fields of view. This inversion eliminates the need for complex positioning systems on each camera, reducing overall device complexity while achieving the same detection coverage through spatial distribution.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables continuous tracking of individuals within a wide field of view without the limitations of positioning mechanisms, allowing for accurate facial recognition and detection of multiple individuals, even when moving, by providing higher-resolution images through fixed, overlapping narrower field of view cameras.
Implementation Method 1
A near-IR illuminator can be provided to generate near-IR light on the individual, which can then be sensed by the one or more TBI cameras to determine the presence of skin and/or to detect various facial features
Data Source
AI summary
Facial detection and tracking systems and methods within a wide field of view are disclosed. A facial detection and tracking system in accordance with an illustrative embodiment of the present invention can include a wide field of view camera for detecting and tracking one or more objects within a wider field of view, and at least one narrower field of view camera for obtaining a higher-resolution image of each object located within a subset space of the wider field of view. The wide field of view camera can be operatively coupled to a computer or other such device that determines the subset space location of the individual within the wider field of view, and then tasks one or more of the narrower field of view cameras covering the subset space location to obtain a high-resolution image of the object.