Defect segmentation

The defect segmentation feature uses deep learning to find and size defects in individual products or a continuous flow of production. If you can see the area of interest in an image, Vision can be trained to identify it as well. The sizes and counts of the defects can further be used to classify production for quality control.

  • Use cases:
      • Find defects in individual products or a continuous flow of production
      • Extract the size and frequency of defects in production
      • Classify your production based on the size and type of defects
      • Increase performance over time with new training examples
      • Add new defects or classes later as needed
  • Requirements:
      • Tens or hundreds of examples for each defect type
      • The defects should be visible to the human eye from an image