Defect Transfer GAN: Diverse Defect Synthesis for Data Augmentation
This addresses data-hunger and imbalance issues in manufacturing defect detection, offering a domain-specific solution that is incremental in its approach.
The paper tackles the problem of data scarcity and imbalance in defect detection by introducing Defect Transfer GAN (DT-GAN), which synthesizes diverse and realistic defective images, resulting in up to a 51% reduction in error rates for defect classification compared to other augmentation methods.
Data-hunger and data-imbalance are two major pitfalls in many deep learning approaches. For example, on highly optimized production lines, defective samples are hardly acquired while non-defective samples come almost for free. The defects however often seem to resemble each other, e.g., scratches on different products may only differ in a few characteristics. In this work, we introduce a framework, Defect Transfer GAN (DT-GAN), which learns to represent defect types independent of and across various background products and yet can apply defect-specific styles to generate realistic defective images. An empirical study on the MVTec AD and two additional datasets showcase DT-GAN outperforms state-of-the-art image synthesis methods w.r.t. sample fidelity and diversity in defect generation. We further demonstrate benefits for a critical downstream task in manufacturing -- defect classification. Results show that the augmented data from DT-GAN provides consistent gains even in the few samples regime and reduces the error rate up to 51% compared to both traditional and advanced data augmentation methods.