How do we verify if the trained model is suitable for cartons?

How do we determine if the cartons we’re using can be recognized using Mech-Mind’s trained deep learning model? If the super model isn’t suitable for recognition, are there any alternative solutions?

  1. Test the camera on-site and collect the images for iteration.
  2. For items that are confirmed not usable, the cardboard supermodel iteration will be completed quickly (40~60 images).