Have you ever wanted to create a "rabbit detector" that alerts you when rabbits are in your garden eating your precious vegetables? Or maybe set up a camera program that can identify the make and model of every car going through an intersection? Object detection classifiers are a type of machine learning neural network that can be trained to detect and identify objects in images, videos, or camera feeds. They can be used to create object detectors that will detect and identify anything your heart desires.

I've been working on training a detector that can identify playing cards in a camera feed for my Blackjack Robot project. I've had some success:

I worked with Google's popular machine learning framework, TensorFlow. However, I noticed there was a lack of clear tutorials for how to set up TensorFlow's Object Detection API to train your own object detector, especially on Windows. (And I'm not about to install Linux on my gaming PC.) 

So, I wrote my own tutorial and created a video that walks through it! The written tutorial is located here at GitHub. It provides set up instructions and comes with a repository that has everything you need to train your own object detection classifier. The video (linked above) walks through the tutorial.

Put your high-powered PC to use and increase your machine learning savvy by training your own object detector using this tutorial. (Actually, any desktop or laptop PC will work, but having a powerful graphics card will significantly reduce the training time.)