Deep Learning video thumb
Key Features
Speed up development with graphical training
The graphical training interface in VisionPro Deep Learning makes it easy to collect images, train the neural network, and test on different image sets, enabling faster and more efficient deployment. VisionPro Deep Learning offers key features to speed up implementation, such as:
- Fast labeling: Accelerate setup with a quick labeling and review process that reduces the number of images and time required for application development.
- Flexible training: Easily train and re-train the system for seamless scalability across multiple production lines.
- Tool-chaining capabilities: Simplify complex problems with unique tool-chaining capabilities that let you break tasks into smaller steps.
Explore the full Cognex portfolio
Download the product guide or compare products to find the right solution for your application.
Automate demanding vision applications with a smarter toolset
VisionPro Deep Learning includes a set of powerful deep learning tools, each designed to solve specific vision challenges. Choose from Blue Locate, Red Analyze, Green Classify, or Blue Read, so you can apply the tool that best meets your needs.
Blue Locate for fixturing, counting, and assembly verification
Reliably find features on noisy backgrounds, in low-light environments, on low-contrast parts, and on parts with significant variation.
Red Analyze for defect detection and segmentation
Detect subtle defects on a variety of backgrounds and surface textures, while adapting to changes in lighting, positioning, and defect type.
Green Classify for object and scene classification
Categorize products based on visual characteristics like color, texture, material, packaging, or defect type, while accounting for natural variation.
Blue Read for text and character recognition
Reliably read distorted, skewed, or poorly etched text and codes, including application-specific characters.
Overcome deep learning limitations with specialized modes
Change between modes without re-labeling images to see which one best suits your requirements.
- Few Sample Mode: high accuracy with limited data
Few Sample Mode delivers powerful results with just a few images per class, minimizing development time while generating precise heatmaps for improved inspection accuracy. This mode is optimized for scenarios with limited datasets, such as rare defects or feasibility studies. - Robust Mode: reliable performance in challenging environments
Robust Mode handles optical challenges such as camera tilt, lighting fluctuations, and white balance shifts, ensuring consistent, high-quality inspections even in the most demanding manufacturing scenarios.
Maximize system efficiency with purpose-built features
Support
Best-in-class technology and tailored guidance to easily plan, deploy, and expand any vision project
Our devices can see even the smallest details with unrivaled accuracy, every time. Our industry-leading support works the way you want, from self-service resources to a global network of experts.
Related Products