for my study i do a practical project with augmented reality. The goal of the project is to recognize defects on industrial objects. My first method i want to use is image tracking and use reference images for the defects to recognize them. Now i want to test it with 3d scanning with ArKit. And now my question. Is it possible to get a percentage number of how much the object got recognized? Like Tensorflow does it. And if its possible i wanted to use this percentage like “if the percentage the scanned object is recognized is higher than 80%, its not defective and if its under this percentage it could be defect.”
