Virtual Defect Image Blending to Minimize Visible Edge Artifacts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for generating virtual defect images result in visible edges when defects are cropped and synthesized into normal images, degrading the learning performance of artificial intelligence models.
Innovation Solution
A system and method for generating virtual defect images that segment defects on a pixel basis, apply transformation methods like rotation, scaling, and image blending to match defect edges with normal image edges, and adjust pixel values using differential equations to minimize edge artifacts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a defect portion of the defect image is cropped and synthesized into a normal image, then a virtual defect image is generated, but an edge of the defect becomes visible which degrades the learning performance of artificial intelligence models
Solution Approach 1:
The patent introduces an edge region as an intermediary zone between the defect region and the normal region. This edge region contains pixel values that are blended between the defect and normal image data, creating a transition zone that masks the sharp boundaries. By using this intermediary edge region, the visible edge artifacts are eliminated while preserving the defect information for AI model training.
Solution Approach 2:
The patent applies different processing strategies to different regions of the image: the defect region uses original defect image data, the edge region uses blended pixel values, and the normal region uses normal image data. This local differentiation allows each region to have optimized characteristics - the defect region preserves defect features, the edge region masks boundaries, and the normal region maintains background integrity.
2Reliability
If a defect is segmented and synthesized into a normal image, then a virtual defect image is generated for AI model training, but the edge of the defect becomes prominent and unnatural
Solution Approach 1:
The patent changes the pixel value parameters in the edge region by blending between defect region pixel values and normal region pixel values. This parameter transformation creates a gradual transition zone that eliminates sharp edges. The blending operation modifies the intensity, color, and other pixel parameters to create a natural-looking transition that masks the defect boundary while preserving the defect's presence for training purposes.
Data Source
AI summary
The present disclosure relates to a system and method for generating a virtual defect image. According to the present disclosure, a system for generating a virtual defect image may include an input unit configured to receive a normal image and a defect image comprising at least one defect, an extraction unit configured to segment and extract a defect, represented by a plurality of pixels. in the defect image, a preprocessing unit configured to apply a preset transformation method on the defect so as to transform the defect into a transformed defect, and a synthesis unit configured to synthesize the transformed defect into the normal image and adjust the transformed defect.