Skin Segmentation for Blemish Removal with Detail Preservation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing facial blemish removal methods process faces globally, leading to the absence of details in facial features and skin texture, resulting in an obvious fake face appearance.
Innovation Solution
A method and apparatus for image processing that involves skin segmentation, fusion with a standard facial-feature mask, determination of blemish regions, and pixel adjustment to remove blemishes while preserving facial feature details.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If global face processing is used for blemish removal, then blemish removal efficiency is improved, but facial feature details and skin texture are lost
Solution Approach 1:
The face image is divided into multiple regions including skin regions and non-skin regions using skin segmentation and facial feature mask fusion. This segmentation allows different processing strategies to be applied to different regions, enabling efficient blemish removal in skin areas while preserving details in facial feature areas.
Solution Approach 2:
Different processing qualities are applied to different regions: aggressive blemish removal is applied to skin regions where it is safe, while conservative processing that preserves details is applied to facial feature regions. This local differentiation resolves the contradiction between removal efficiency and detail preservation.
2Reliability
If aggressive filtering is applied to remove blemishes, then blemish removal effect is improved, but skin texture and facial features become blurry
Solution Approach 1:
The patent applies different filtering intensities to different regions: strong filtering is applied to skin regions to effectively remove blemishes, while weak or no filtering is applied to facial feature regions to maintain texture clarity. This localized quality adjustment resolves the contradiction between removal effectiveness and texture preservation.
Solution Approach 2:
The patent uses mask images (skin segmentation mask and facial feature mask) as intermediaries to control the application of filtering operations. These masks act as mediators that guide the filtering strength to appropriate regions, enabling selective application that resolves the contradiction between removal effect and texture preservation.
3Device complexity
If uniform processing is applied across the entire face, then processing simplicity is improved, but realism of facial features is lost
Solution Approach 1:
The face is segmented into skin regions and non-skin regions using multiple mask images. This segmentation adds complexity to the processing pipeline but enables realistic facial feature preservation by allowing different processing parameters to be applied to different regions, thus resolving the contradiction between simplicity and realism.
Solution Approach 2:
Different processing parameters and strategies are applied to different facial regions based on their semantic meaning. Skin regions receive blemish-focused processing while facial feature regions receive detail-preserving processing, maintaining realism despite increased processing complexity.
Data Source
AI summary
Disclosed in embodiments of the disclosure are a method and apparatus for image processing, a device, and a storage medium. The method includes: performing skin segmentation on an original image, thereby obtaining a segmentation result image; fusing the segmentation result image and a standard facial-feature mask image, thereby obtaining a first skin region image; determining a blemish region according to the first skin region image and the original image; and adjusting a pixel of the blemish region in the original image, thereby obtaining a target image.