Cognex, 'BD + SuaKit + VisionPro' integration
Deep learning-based machine vision software, Vision Pro Deep Learning
Launched and opened a deep learning research lab with 80 employees.


Cognex announced on the 18th that it has launched 'VisionPro Deep Learning', which integrates 'Cognex ViDi', a deep learning-based machine vision software, and 'SuaKit', a deep learning software from SuaLab, which it acquired in October 2019.

Cognex BD consists of four tools for character recognition (OCR): Blue-Locate, Red-Analyze, Green-Classify, and Blue-Read. Later, it was integrated with rule-based software, VisionPro, enabling the use of multiple inspection and development tools.

VisionPro Deep Learning's Green-Classify tool has been specifically enhanced. Users can now easily switch between "Green–High Detail" mode, which significantly improves classification accuracy, and "Green–Focused" mode, which provides faster learning and inference, without additional labeling. The Red-Analyze tool has also been enhanced, with "Red-High Detail" mode enabling 5-25% more accurate defect size and shape measurements. "Red-Focused" mode also enables faster learning and inference capabilities.

Additionally, it is integrated with rule-based software, 'Vision Pro 10', enabling analysis by loading the deep learning runtime workspace into Vision Pro 10's 'GUI Quick Build'. Furthermore, the newly added 'Edit Control' function in Vision Pro 10 enables editing of runtime parameters, results, and graphs.

Cognex also opened a dedicated deep learning research lab. The lab, staffed by approximately 80 deep learning engineers and a deep learning marketing team, serves as Cognex's deep learning research and development center. It is responsible for developing new deep learning technologies tailored to market demand and developing VisionPro deep learning products. It also provides Q&A on deep learning and technical support for real-world challenges.