Feature Extractor with ml5.js

You can use ml5’s feature extractor to train a machine learning image classifier with your own images of two different objects to be recognized.

  1. First, allow this website to use the camera.
  2. Click on Object 1 button several times with an object or, even better, different objects of the same type, in front of the camera in order to record all images.
  3. Repeat the same process with Object 2 button.
  4. Click on Train button, in order to execute the feature extactor function included in ml5.js.
  5. The result below contain the name or label of the objects and the probability that it is correct.