Multi-Category Decomposition Editing Network for the Accurate Visual Inspection of Texture Defects

Spotting blemished areas automatically on a tex- tured surface is a particular challenge, as both nominal and de- fective surface samples are inconsistent in large-scale industrial manufacturing. The most efficient solution uses the memory bank extracted from the nominal samples to detect outliers. We approach our strategy, the multi-category decomposition editing network (MCDEN), from a similar viewpoint. Notably, we do not use defect-free samples. Instead, we use virtual results to construct a defect library. MCDEN decomposes abnormalities to basic elements from the library while editing outlier features to reconstruct the texture normality, offering a rational segmen- tation map through decomposition and reconstruction. Based on the strategy, MCDEN is more interpretable than most neural network methods since interpretability is particularly important in industry to ensure stability. Experiments on texture surface samples from the MVTAD dataset confirm the efficacy of MCDEN with a pixel-level AUC score of 96.6%. In other experiments collected from semi-manufactured inkjet printing OLED panels, MCDEN demonstrates competitive results with a 99.2% detection rate and rapid real-time detection capability.