CustomObjectClassifier

A little demo of a object classification model which you can train. It can be trained to detect 3 different objects. 

To use it, select 3 objects. For each object:

1. Put it in front of the camera
2. Select an object number to correspond to that object (1-3)
3. Click and hold the Object # button. A count next to the button will say how many images have been taken of the object
4. Make sure to move the object around slightly to give the model variable data

After you have finished taking pictures for all 3 objects. train the model by clicking Train. The project may freeze for a little bit before updating the status from untrained to training. Do not click again.

Wait until the status is Trained, then you can press the detect button to see which object the model thinks is in front of the camera.

Press Reset Model to reset (duh).