Case Study: a Saudi-based Electronics Manufacturer achieves 97.4% defect detection accuracy with instinctools’ computer vision solution

Saudi-based Electronics Manufacturer reaches 97.4% defect detection accuracy with instinctools

The customer, a Saudi-based electronics manufacturer, faced significant limitations with its rigid and costly automated optical inspection (AOI) system for defect detection. Their rule-based machine vision was inflexible, slow to adapt, and overly dependent on perfect imaging conditions. They turned to *instinctools to implement a more advanced and scalable computer vision solution.

*instinctools developed a custom system using a pre-trained YOLOv8 model, enhanced with sophisticated data augmentation and transfer learning techniques. This new solution achieved 97.4% accuracy in identifying production line flaws. The results included a $2.4 million annual savings from reduced warranty claims, a 67% drop in defect-related claims, a 24% increase in production throughput, and a reduction in falsely discarded items to below 0.5%.