Training

As with traditional rule-based vision tools, how the VisionPro Deep Learning deep learning tools are trained is the largest determinant of their runtime performance.

Unlike rule-based vision tools, successful deep learning training is largely a function of the number and representativeness of the images that you collect for your training image set, and the consistency with which you label them.

The key steps in the training phase are the following:

• Image set collection and annotation
• Image set labeling
• Model definition (for the Blue Locate Tool and Blue Read Tool only)
• Defining a Training Set
• Defining a Validation Set (for the Green Classify Tool in High Detail mode only)
• Neural network training