Steel surface inspection with better labeling
QualityLandingLens
- Setup
- Use consensus labeling to clean training data before training defect models.
- Time saved / result
- Accuracy ~70-80% to 93%
- Source
- Credit: Landing AI case study
https://landing.ai/wp-content/uploads/2021/07/LandingAI_CaseStudy_Steel_Inspection.pdf
Every new AI feature in factory software, once a week. Free, weekly. Or follow via RSS.