Gain Object Detection on Mobile Devices Using TensorFlow in a Browser

The fine folks over at Adafruit have put together a handy tutorial on how to run the TensorFlow machine learning application within a browser, which is then able to take advantage of a smartphone, PC, or laptop camera to identify objects it sees. The secret to running the AI application is the use of Google’s Colab cloud-based service, which offloads all the processing necessary using their GPU servers rather than relying on internal hardware.

"We've hacked together a Colab notebook that can see things using your computer, laptop, or phone camera! It takes live pictures from your camera and feeds them through the Mobilenet v2 + SSDLite model to find and box the objects it sees. This way, you can see what Mobilenet v2 + SSDLite can do, instantly!”

The platform works by using Mobilenet v2 + SSDLite — a neural network-based object detection system, which offers three outputs utilizing the camera’s video feed, including object identification, confidence level in determining what they are, and a bounding box drawn around the objects in question. TensorFlow takes that data from object detection system and learns to identify the items within its view with increased accuracy.

This is an excellent project for those who would like to learn about AI without needing a Ph.D. in computer science, or for those with limited hardware that would like to see what all the fuss is about machine-learning.