Virtual Makeup Replication via Feature Template Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional virtual makeup applications have limited predefined makeup styles, making it difficult for individuals to replicate desired makeup appearances found in arbitrary images.
Innovation Solution
A computing device performs facial alignment and defines source regions in a source image, extracts attributes, identifies matching feature templates, and applies these templates to corresponding regions in a user's digital image to replicate makeup effects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional virtual makeup applications use predefined makeup styles, then the application is simple to operate, but the adaptability to replicate arbitrary makeup appearances is limited
Solution Approach 1:
The system pre-extracts makeup attributes from source images and stores them in feature templates before actual application. This preliminary processing allows the system to quickly match and apply predefined makeup styles without real-time complexity, resolving the contradiction between versatility and system complexity.
Solution Approach 2:
The system creates digital feature templates that copy and store the essential attributes of makeup appearances from source images. These templates can be repeatedly applied to different user images, enabling arbitrary makeup replication without requiring the full source image processing each time, thus reducing operational complexity while maintaining adaptability.
2Manufacturing precision
If the system performs detailed facial alignment and attribute extraction, then the manufacturing precision of makeup application is improved, but the processing time increases
Solution Approach 1:
Facial alignment and attribute extraction are performed in advance on source images to create feature templates. This preliminary action separates the time-consuming precision work from the actual application phase, allowing quick template matching and application while maintaining high precision through the pre-processed templates.
Solution Approach 2:
The system extracts only the essential makeup attributes from source images and stores them in feature templates, separating the critical precision information from the full image data. This extraction allows precise makeup application using only the necessary attribute data, reducing processing time while maintaining manufacturing precision.
3Adaptability or versatility
If the system uses arbitrary source images as templates, then the adaptability to user preferences is improved, but the difficulty of detecting and measuring makeup attributes increases
Solution Approach 1:
The system creates standardized feature templates that copy the essential makeup attributes from arbitrary source images into a uniform format. This copying process transforms diverse source images into consistent template structures, making attribute detection and measurement more manageable while maintaining the ability to replicate any makeup style.
Solution Approach 2:
The system transforms various makeup attributes from different source images into a standardized parameter set for feature templates. By changing and normalizing the parameters into a consistent format, the system simplifies the detection and measurement process while maintaining adaptability to replicate diverse makeup appearances from arbitrary sources.
Data Source
AI summary
A computing device obtains a source image depicting a facial region having one or more makeup effects. The computing device performs facial alignment and defines a plurality of source regions having the one or more makeup effects, the source regions corresponding to facial features in the source image. The computing device extracts attributes of the one or more makeup effects for each source region and identifies a closest matching feature template for each source region based on the attributes. The computing device obtains a digital image of a facial region of a user. The computing device performs facial alignment and identifies a plurality of target regions corresponding to the plurality of source regions. The computing device applies a matching feature template of a corresponding source region to each of the target regions.