Hi guys! Today I want to show you how to create a custom object detector with Custom Vision and execute it in a browser. Through this walkthrough we will see how we could use advanced machine learning models without being a specialist, that’s idea of AutoML. You can read more about it in my previous article (link).

Before to start this tutorial, I’ll explain some tools to be used:

• CustomVision: it’s part of Microsoft Cognitive Services, with it you can generate custom computer vision models.
• TensorFlow.js: it’s a library to build machine learning models in javascript and execute them in a browser or node.js.
• Customvision-tfjs: it’s a NPM package to use TensorFlow.js models exported from Custom Vision service

So, let’s start to work! In this example we will train a AI to identify two popular anime characters: Goku (DragonBall) and Seiya (Saint Seiya).

1. Create…