Facial Detection Using Edge-Based Window Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object detection systems, such as the Viola-Jones framework, face challenges in efficiently processing large digital images in real-time due to high computational demands, particularly when analyzing every possible window at multiple scales, which can be impractical for embedded processors.
Innovation Solution
The method involves dividing a digital image into windows, performing edge detection using the Sobel operator to classify each window as a foreground or background candidate based on edge thresholds, and only applying additional object detection to foreground candidates, thereby reducing processing load and clock cycles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If Viola-Jones object detection framework processes every possible window at multiple scales, then detection reliability is improved, but computational demand and processing time increase significantly
Solution Approach 1:
The patent divides the image into multiple windows and applies edge detection to each window independently. By segmenting the image this way, the system can process only relevant regions rather than every possible window at multiple scales, reducing computational demand while maintaining detection reliability through targeted analysis of edge-rich regions.
Solution Approach 2:
The patent performs edge detection as a preliminary action before full object detection. By first identifying windows with sufficient edges using the Sobel operator, the system pre-filters the image to eliminate background regions, thereby reducing the number of windows that require intensive Viola-Jones analysis and improving overall processing speed.
2Measurement precision
If edge detection threshold is set high to reduce false positives, then detection precision improves, but false negatives increase
Solution Approach 1:
The patent applies different threshold criteria for different windows based on their edge density. By evaluating each window's edge count individually and applying adaptive thresholds, the system maintains high precision in edge-rich regions while ensuring sufficient sensitivity in edge-sparse regions, thereby balancing false positive and false negative rates.
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
This approach allows for real-time or near-real-time object detection, such as facial recognition, with a high reliability and minimal false negatives, while tolerating a higher rate of false positives, thus reducing processor-intensive analysis to manageable levels, suitable for security applications at VGA resolution.
Implementation Method 1
performing edge detection on each window
Data Source
AI summary
In one aspect, there is disclosed a digital signal processor and method performed by the same for performing object detection, including facial detection, in a reduced number of clock cycles. The method comprises using Sobel edge detection to identify regions with many edges, and classifying those regions as foreground candidates. Foreground candidates are further checked for vertical or horizontal symmetry, and symmetrical windows are classified as face candidates. Viola-Jones type facial detection is then performed only on those windows identified as face candidates.