Cognex Deep Learning uses color cameras to accurately verify that winding processes have been completed flawlessly. The defect detection tool can learn from a set of training images that includes images of zero-defect winding, as well as images that are labeled for a variety of other possible defects that can appear in different locations, including overlaps, misalignments, and crossovers.