Eye Beautification Using Localized Glint and Smoothing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing techniques for portrait images, particularly in digital photography, are inefficient and resource-intensive, especially in embedded systems like digital cameras, as they often require global analysis and memory-intensive operations for face detection and beautification, failing to effectively address localized issues such as eye defects and lighting variances.
Innovation Solution
The development of a method that applies selective smoothing or blurring to specific regions of a face, such as the eyes, using localized color smoothing kernels and noise reduction techniques, which identifies borders between the iris and sclera, and enhances the appearance of eyes by adding artificial glints, while being computationally efficient and suitable for embedded devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If global image analysis techniques are used for face detection and beautification, then comprehensive face feature detection is achieved, but computational complexity and memory usage increase significantly
Solution Approach 1:
The patent segments the face image into distinct regions (eyes, mouth, forehead, cheeks, chin) and applies different processing operations to each region. This allows localized beautification operations without requiring global image analysis, reducing computational complexity while maintaining detection accuracy for each specific region.
Solution Approach 2:
The patent applies different processing operations to different regions of the face based on local characteristics. For example, smoothing is applied to skin regions while preserving eye and mouth details, and selective blurring is applied to specific areas like the forehead and cheeks. This localized approach reduces overall computational complexity by avoiding uniform global processing.
2Manufacturing precision
If global image manipulation techniques are used for lighting correction and beautification, then comprehensive image quality improvement is achieved, but memory usage and processing time increase
Solution Approach 1:
The patent divides the image processing into separate stages: face detection, region segmentation, and localized beautification operations. Each stage processes only the relevant portions of the image, avoiding redundant computations on the entire image and reducing overall processing time.
Solution Approach 2:
The patent applies beautification operations selectively to specific regions that require improvement (such as smoothing skin in certain areas) rather than applying uniform processing to the entire image. This partial action approach maintains image quality in critical regions while reducing processing time by avoiding unnecessary operations elsewhere.
3Measurement precision
If resource-intensive global analysis is used for face detection, then accurate face feature identification is achieved, but embedded system performance deteriorates
Solution Approach 1:
The patent implements a segmented processing approach where face detection is performed first to identify the face region, then the face is divided into sub-regions (eyes, mouth, forehead, cheeks, chin) for targeted beautification. This segmentation allows the embedded system to process only relevant regions with appropriate algorithms, maintaining accuracy while improving performance by avoiding global image processing.
Solution Approach 2:
The patent performs preliminary face detection and region identification before applying beautification operations. By pre-identifying the face boundaries and internal regions, the system avoids redundant computations during the beautification stage, improving embedded system performance while maintaining detection accuracy.
Data Source
AI summary
Sub-regions within one or more face images are identified within a digital image, and enhanced by applying an artificial glint symmetrically and/or synchronously to image data corresponding to sub-regions of eyes within the face image. An enhanced face image is generated including an enhanced version of the face that includes certain original pixels in combination with pixels corresponding to the one or more eye regions of the face with the artificial glint.