Facial Key Point Fusion for Natural Face Morphing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image processing technologies lack an effective method to fuse different faces in two images into a single face image, resulting in suboptimal fusion results.
Innovation Solution
An image processing method that detects facial key points in both images, determines a conversion relationship based on these key points, and fuses the faces using morphing processing to create a new image with a unified face, incorporating techniques like weighted stacking and Gaussian blurring to enhance the fusion effect.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional image fusion methods are used, then the fusion process is simple, but the fusion result quality is poor and cannot effectively combine different faces into a single face
Solution Approach 1:
The patent segments the face image into multiple key point regions (eyes, nose, mouth, contours) and processes each region separately based on its characteristics. This segmentation enables precise control over different facial features during fusion while maintaining overall face structure integrity.
Solution Approach 2:
The patent applies different processing strategies to different local regions of the face. For example, it uses key point matching for landmark alignment, morphing for intermediate regions, and selective blending for specific features like eyes and mouth, thereby achieving high-quality fusion with natural transitions.
2Measurement precision
If key point detection is performed on both images to establish conversion relationship, then the positioning accuracy is improved, but the processing time increases
Solution Approach 1:
The patent performs key point detection and conversion relationship establishment as preliminary steps before the actual fusion process. By pre-aligning the images using detected key points and conversion relationships, the subsequent fusion operations can proceed efficiently without repeated positioning calculations.
Solution Approach 2:
The patent introduces a conversion relationship as an intermediary transformation that maps key points from one image to another. This intermediary step enables accurate alignment and facilitates efficient fusion by establishing a clear correspondence between features in different images.
Data Source
AI summary
Provided are an image processing method and device and an electronic device. The method may include: a first image and a second image are obtained; a facial key point of a target object in the first image is detected to obtain information of a first key point; a facial key point of a target object in the second image is detected to obtain information of a second key point; a conversion relationship is determined based on the information of the first key point and the information of the second key point; and faces of the target objects in the first image and the second image are fused based on the conversion relationship to obtain a third image.