In this app you can teach computer recognize 2 or more objects using Tensorflow.js

IMPORTANT: This application does not send user data to the backend, everything is done in the web browser, training and predictions are made locally.

Application is loading ...

Model was NOT trained yet.

• Press Camera ON it enable webcam - allow access - this app not sending any data from your computer.
• Now find an object, point cam at it, click and hold button for first object - it will take 30 photos of the object
• Repeat the previous step for other objects.
• Click train and predict. Wait a moment - remember No data is sent to server.
• After a few seconds, the computer will display information that the object has been found.