Defect Image Augmentation and Closed-Loop Model Refinement System

For AOI equipment, GAI generates rare defect images to solve data imbalance and sample shortages, reducing model bias.
With RAG industry knowledge and closed-loop refinement, feedback, retraining, and data regeneration enhance unexpected defect recognition.
It cuts data collection and annotation by over 50%, reaches >90% detection with 1,000 real rare defect images, and improves accuracy to ≥90%.

Industrial Applications:

This technology reduces data collection and labeling without on-site visits, shortens PoC preparation, and supports deployment. It can expand to smart inspection in metal processing, automotive parts, and PCB. By integrating data import, pseudo-labeling, cluster updates, retraining, and inference, it achieves 89% detection with 50% labeled images, halves labeling time, and cuts PoC time by 30–40%.

Industrial Technology Research Institute

ITRI is a world-leading applied technology research institute with more than 6,500 outstanding employees. Its mission is to drive industrial development, create economic value, and enhance social well-being through technology R&D. Founded in 1973, it pioneered in IC development and started to nurture new tech ventures and deliver its R&D results to industries. ITRI has set up and incubated companies such as TSMC, UMC, Taiwan Mask Corp., Epistar Corp., Mirle Automation Corp., and Taiwan Biomaterial Co.

Contact Information

Name：Wei-Chen Li

Tel：03-591-6670

Address：No. 195, Sec. 4, Chung Hsing Rd., Chutung, Hsinchu County 310401, Taiwan (R.O.C.)

Additional Information

Pavilion：Innovation Economy Pavilion Smart Manufacturing IE06

Exhibiting Entity：Department of Industrial Technology,MOEA

Application Category：Machinery and System

Exhibition