Beauty Effect Processing Apparatus for Face Image Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing digital image processing devices apply beauty effects uniformly to all faces in an image, leading to insufficient enhancement of the intended subject, as they prioritize small details and color differences, resulting in unsatisfactory complexion results.
Innovation Solution
A beauty effect processing apparatus and method that selectively applies beauty effects by identifying and excluding certain subjects, determining a protagonist face, and adjusting the beauty level based on distance and face size relative to the protagonist, allowing for customized sharpness application to each face in an image frame.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If a beauty effect is applied uniformly to all faces in an image, then processing simplicity is maintained, but the beauty effect on the intended subject becomes insufficient
Solution Approach 1:
The patent segments the image processing by dividing faces into different categories (excluded subjects, protagonists, and other subjects) and applying different beauty effect levels to each segment. This allows the intended subject to receive enhanced beauty processing while other faces receive reduced or no processing, resolving the contradiction between processing precision and complexity.
Solution Approach 2:
The patent implements local quality by applying different beauty effect intensities to different regions of the image based on face importance. The protagonist face receives maximum beauty effect, excluded subjects receive no effect, and other subjects receive intermediate effects, thereby achieving precise beauty enhancement on the intended subject without uniformly processing the entire image.
2Reliability
If beauty effect is applied to all faces, then processing consistency is maintained, but user satisfaction with the intended subject's complexion deteriorates
Solution Approach 1:
The patent applies local quality by differentiating beauty effect application based on face identification results. The protagonist face receives full beauty processing for optimal complexion enhancement, while excluded subjects are deliberately omitted from processing. This selective approach maintains reliability through systematic processing rules while dramatically improving complexion enhancement quality on the intended subject.
3Productivity
If uniform beauty processing is applied, then processing time is reduced, but the beauty effect on the protagonist becomes insufficient
Solution Approach 1:
The patent segments faces into priority levels and processes them accordingly. By identifying the protagonist and excluded subjects first, the system can apply intensive beauty processing only to the protagonist face while applying minimal or no processing to other faces. This segmented approach maintains processing efficiency while achieving high beauty effect intensity on the intended subject.
Solution Approach 2:
The patent implements partial action by applying beauty effects selectively rather than uniformly. The protagonist receives excessive beauty processing (maximum effect), excluded subjects receive zero processing, and other subjects receive partial processing. This approach achieves sufficient beauty effect on the protagonist without the time cost of processing all faces at maximum intensity.
Data Source
AI summary
Disclosed is a beauty effect processing apparatus that increases the effectiveness of a beauty effect applied to a face included in an image frame by selecting a protagonist, to which the beauty effect having a maximum sharpness is to be applied, from at least one received face image, and determining a beauty level corresponding to sharpness of the beauty effect to be applied to the received at least one face image based on at least one of a spaced distance from the protagonist and a face size of the received at least one face image relative to a face size of the protagonist.


