Cartoon Face Image Generation via Feature Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for generating cartoon face images often result in feature parts being integrated, making independent adjustments or replacements difficult, leading to reduced operability and similarity to real faces due to overall stylization.
Innovation Solution
A method and apparatus that obtain and adjust cartoon materials for face feature parts independently by using facial feature information, allowing for separate modification and combination to generate a cartoon face image, improving operability and similarity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If overall stylization conversion is performed on the entire face picture using image stylization algorithm, then the cartoon face image can be generated with bright colors and vivid appearance, but the feature parts become integrated and independent adjustments or replacements become difficult
Solution Approach 1:
The patent segments the face image into multiple independent feature part images (eyes, eyebrows, nose, mouth, cheeks, etc.) and processes each separately. This allows independent adjustment and replacement of individual features while maintaining the overall cartoon style, resolving the contradiction between generating cartoon images and enabling independent feature adjustments.
Solution Approach 2:
The patent extracts specific feature parts from the original face image using face detection and key point detection algorithms. By taking out individual features (eyes, nose, mouth) as separate images, the system enables independent processing and adjustment of each feature without affecting others, solving the operability issue.
2Productivity
If overall stylization conversion is performed on the entire face picture, then the cartoon face image can be generated efficiently, but the similarity to real faces is reduced due to integrated processing
Solution Approach 1:
By dividing the face into separate feature parts and processing each independently through the same stylization algorithm, the patent maintains consistent cartoon styling across all features while preserving the unique characteristics of each individual feature, thereby improving similarity to the original face.
Solution Approach 2:
The patent applies the stylization algorithm locally to each feature part rather than globally to the entire face. This allows each feature to maintain its specific characteristics and details while still receiving the cartoon style treatment, improving the overall similarity to the real face.
Data Source
AI summary
In a method for generating a cartoon face image that is performed by a computing device, a to-be-processed face image is obtained. Face feature information is recognized from the to-be-processed face image, and cartoon materials corresponding to feature parts of a face in the to-be-processed face image are obtained according to the face feature information. The obtained cartoon materials are cartoon face materials. The cartoon face materials include cartoon facial feature materials and another cartoon material other than the cartoon facial feature materials. Each of the cartoon facial feature materials are adjusted by using an adjustment parameter, and the other cartoon material are combined with the cartoon facial feature materials after adjustment to generate the cartoon face image.