Defect image sample generation method and device and panel defect detection method

Abstract

The invention provides a defect image sample processing method and device and a panel defect detection method. The method comprises the following steps: extracting a defect feature region on an original defect image sample, and performing texture suppression on the defect feature region to obtain a first simulated defect image; extracting an ROI (Region Of Interest) equal to the defect feature region in area from the background image sample; extracting background texture features and superposing the background texture features into the first simulated defect image to obtain a second simulateddefect image; and finally, fusing the second simulated defect image and the background image sample to generate a defect image sample. According to the effective method for synthesizing the defect image sample by simulating any scene provided by the invention, the defect area is enabled to have the texture of the existing background through a background texture feature extraction and coverage method, the pixel distribution is closer to the background area, and the image is more natural; therefore, the generated defect image sample can meet the training requirement of a deep learning model, andthe problem of insufficient training samples is solved.

Description

technical field [0001] The invention relates to the field of image data processing for automatic defect detection of display panels, in particular to a method and device for generating defect image samples. Background technique [0002] The deep learning model, also known as the neural network model, is a highly complex nonlinear dynamic learning system formed by a large number of simple processing units widely interconnected with each other. At present, it is already possible to obtain a deep learning data model with specific functions by using training data for model training. In particular, for the defect identification or detection requirements of a specific display panel automatic defect detection scenario, the corresponding algorithm can be used for deep learning model training based on the training samples to build a defect area recognition or detection model for a specific detection scenario. Therefore, sufficient quantity and quality of defect image samples are cru...

AI Technical Summary

Problems solved by technology

The generated defect image samples can meet the training requirements of the deep le

Images