South KoreaManufacturing2024 – ongoingKorea Electric Terminal
Self-learning AI vision for defect inspection and injection-mold monitoring
Korea Electric Terminal applied self-learning AI vision to factory defect inspection
Problem
Korea Electric Terminal manufactured many different components requiring inspection. Collecting and labeling sufficient training examples for each product made conventional AI inspection difficult.
Approach
The company worked with Chungbuk National University's Industrial AI Research Center on a system using very little training data. It generated and inserted defect patterns for self-learning anomaly detection, with reported factory application in October 2024.
Results
No published figures
Takeaway
The project addressed the shortage of labeled defect examples by generating training patterns.
Categories
- Industry
- Manufacturing
- Business function
- Production and quality
- Innovation type
- Process innovation
- AI technology
- Computer vision
- Adoption
- Partner co-development
- Evidence
- Vendor case study
Sources
Source summaries
Chungcheong Biz reported the university development team's account of successful factory application with Korea Electric Terminal. It described synthetic defect patterns and reduced labeling requirements, without publishing an inspection-accuracy figure.
Full details are in the original sources.