Summary of Utilizing Generative Adversarial Networks For Image Data Augmentation and Classification Of Semiconductor Wafer Dicing Induced Defects, by Zhining Hu et al.
Utilizing Generative Adversarial Networks for Image Data Augmentation and Classification of Semiconductor Wafer Dicing Induced Defects
by Zhining Hu, Tobias Schlosser, Michael Friedrich, André Luiz Vieira e Silva, Frederik Beuth, Danny Kowerko
First submitted to arxiv on: 24 Jul 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Machine Learning (cs.LG); Image and Video Processing (eess.IV)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The paper explores the application of generative adversarial networks (GAN) for image data augmentation and classification of semiconductor wafer dicing induced defects. The goal is to enhance the variety and balance of training data for visual inspection systems. By generating synthetic yet realistic images that mimic real-world dicing defects, the authors aim to improve classification accuracies. The study employs three GAN variants: DCGAN, CycleGAN, and StyleGAN3. Preliminary results show an average improvement of up to 23.1% in balanced accuracy, which could enable yield optimization in production. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper looks at how to use computer learning to help make chip-making better. Right now, computers are really good at looking at pictures and finding problems. But they need a lot of practice data to get good. To help with this, the authors created new fake images that look like real problems that might happen during chip-making. They used special computer programs called GANs (Generative Adversarial Networks) to make these images. The results show that using these fake images can help computers find problems better, which could make chip-making more efficient and less wasteful. |
Keywords
» Artificial intelligence » Classification » Data augmentation » Gan » Optimization
