"You Only Look Once" (Tiny YOLOv2) Object Detection

This Open Neural Network Exchange, Tiny YOLOv2 object detector is trained on the Pascal VOC dataset and is made up of 15 layers (9 convolutional and 6 max-pooling layers) that can predict 20 different classes of objects. Keep in mind that Tiny YOLOv2 is a condensed version of the original YOLOv2 model, a tradeoff is made between speed and accuracy, but it demonstrates the ability to incorporate ML.NET technology into your web applications.

This particular model has been trained to detect the following 20 types of objects:

Person, Bird, Cat, Cow, Dog, Horse, Sheep, Aeroplane (Airplane), Bicycle, Boat, Bus, Car, Motorbike (Motorcycle), Train, Bottle, Chair, Dining Table, Potted Plant, Sofa and TV/Monitor

We've provided an example of each class in the preloaded image gallery that you can test by clicking the image.

Now, remember that this detector will misidentify, not identify or position the bounding box incorrectly depending on the image input for object detection - it's not meant to be super-accurate! It does a pretty good job for being "scaled-down" and to be working in a web application is just amazing considering that just a few years ago this wasn't even possible!

This demo is particularly entertaining if you are browsing with your mobile device because we all have lots of pictures on our phones that have objects which fall into the 20 categories above!

Have fun and upload some of your own images to see if it can detect any objects!