Introducing DWORLD's accumulated know-how.
1 High-speed deep learning AI detection model
_ - Developed Deep learning defect image classification model
_ Developed Defect detection models such as aramid fiber yarn, paper plate images and mat tiles using Vision Deep Learning Models such as Resent and EfficientNet and also developed software
_ Developed Deep learning defect image detection model
_ Developed a model for detecting anomalies after learning a normal image using U-Net-based technology
2 Developed deep learning dataset augmented S/W
_ Deep learning AI models require a large number of data sets
_ We supply numerous DVS for product inspection, and have enough defect images acqruied from sample inspections for DVS production and supply.
_ Developed software that augments data sets to meet the need for technology that creates defects of various shapes and increases the quantity using the secured defect images through industry-university cooperation