Cartoonization Pixel Adjustment for Diverse Skin Tones
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current cartoonization processing models are limited to processing images with light skin colors, as they are primarily trained on light skin data, failing to accurately transform faces with darker skin tones.
Innovation Solution
A method that performs skin color recognition on facial images to determine the target skin color, processes the image using an existing cartoonizing model to generate a reference cartoonized image, and adjusts pixel values based on a pixel adjustment parameter to achieve the target skin color, expanding the model's applicability to various skin tones.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a cartoonizing model trained on light skin data is used, then the cartoonization processing is accurate for light skin colors, but it fails to accurately transform faces with darker skin tones
Solution Approach 1:
The patent adjusts pixel values of the cartoonized image based on skin color parameters. Specifically, it determines a pixel adjustment parameter based on the target skin color and reference skin color, then adjusts pixel values to transform the skin tone from light to dark while maintaining cartoonization quality. This resolves the contradiction by keeping the original light-skin-trained model (maintaining accuracy) while adding parameter-based pixel adjustment (achieving skin color adaptability).
Solution Approach 2:
The patent introduces a pixel adjustment mechanism as an intermediary between the cartoonizing model and the final output. The pixel adjustment parameter acts as a mediator that bridges the gap between the model's fixed light-skin output and the desired variable skin color output, enabling the model to serve multiple skin tones without retraining.
2Adaptability or versatility
If multiple cartoonizing models for different skin colors are trained, then the adaptability to various skin tones is improved, but the device complexity and training data requirements increase
Solution Approach 1:
The patent makes a single cartoonizing model universal by adding a pixel adjustment layer that can handle multiple skin colors. Instead of creating separate specialized models for each skin tone, the system uses one model with an adaptable output adjustment mechanism, thereby achieving multi-functionality without increasing model complexity.
Solution Approach 2:
The system determines the pixel adjustment parameter automatically based on the target skin color and reference skin color without requiring manual intervention or multiple pre-configured models. The single model serves all skin tones by self-adjusting through the pixel value modification process.
Data Source
AI summary
The disclosure discloses a cartoonlization processing method for an image, and relates to a field of computational vision, image processing, face recognition, deep learning technologies. The method includes: performing skin color recognition on a facial image to be processed to determine a target skin color of a face in the facial image; processing the facial image by utilizing any cartoonizing model in a cartoonizing model set to obtain a reference cartoonized image corresponding to the facial image in a case that the cartoonizing model set does not contain a cartoonizing model corresponding to the target skin color; determining a pixel adjustment parameter based on the target skin color and a reference skin color corresponding to the any cartoonizing model; and adjusting a pixel value of each pixel point in the reference cartoonized image based on the pixel adjustment parameter, to obtain a target cartoonized image corresponding to the facial image.