Facial Recognition Preprocessing via Edge Discontinuity Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Face recognition systems face challenges in classification accuracy due to illumination, pose, background, and color leakage, which significantly reduce performance.
Innovation Solution
A novel preprocessing scheme that identifies edges and discontinuities in images, generates a contiguous outline using linear interpolation, and fills the background with a specific color to isolate the face portion, improving classification accuracy across varying conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If color leakage is present in the input image, then the face region becomes indistinguishable from the background, but classification accuracy significantly reduces
Solution Approach 1:
The patent segments the image into distinct regions (face region and background) by detecting edges and identifying discontinuities. This segmentation allows the system to separate the face from the background even when color leakage is present, thereby maintaining classification accuracy despite the harmful color leakage effect.
Solution Approach 2:
The patent extracts the face region from the image by identifying edges and discontinuities, then isolates it by filling the background with a uniform color. This extraction removes the harmful color leakage effect from the face region, allowing accurate classification to proceed without the interfering background colors.
2Measurement precision
If background clutter is present in the image, then the face detection becomes more difficult, but classification accuracy deteriorates
Solution Approach 1:
The patent extracts the face region by detecting edges and discontinuities that separate the face from the background clutter. By isolating the face region and filling the background with a uniform color, the system removes the harmful background clutter effect, enabling accurate classification despite the presence of clutter in the original image.
Solution Approach 2:
The patent applies different processing to different regions of the image: the face region is preserved with its original characteristics while the background is filled with a uniform color. This local quality approach maintains the important face features while eliminating the harmful background clutter, thereby improving classification accuracy.
3Measurement precision
If edge discontinuities are present due to color leakage, then the face boundary becomes inaccurate, but classification performance degrades
Solution Approach 1:
The patent performs preliminary edge detection and discontinuity identification before final face region extraction. By detecting and marking edge discontinuities in advance, the system can compensate for inaccurate boundaries caused by color leakage, ensuring that the final face region extraction maintains high accuracy despite the presence of discontinuities.
Data Source
AI summary
A method preprocessing scheme that improves classification accuracy of a face recognition system comprises identifying a plurality of edges of an image. The method further comprises step of identifying a plurality of discontinuities in the plurality of edges of the image. Furthermore, the method also comprises generating contiguous outline for the image. The last step of the method comprises filling a background of the image with at least one color.