Explore the Google Coral Dev Board Micro's capabilities in this comprehensive 33-minute tutorial on low-power object detection. Learn to train a custom MobileNetV2-SSD object detection model and deploy it to the Coral Micro, achieving 8-12 FPS inference speeds using 320x320 color images at under 1 watt. Follow along as the process of data collection, labeling with makesense.ai, model training using MediaPipe, and deployment with the coralmicro SDK is demonstrated. Gain insights into debugging techniques, including visualizing camera output and bounding box information through a web server. Perfect for developers and enthusiasts looking to harness the power of AI in edge computing applications.
Getting Started with Google Coral Dev Board Micro - Low-Power Object Detection
Overview
Syllabus
Intro
Flash image collection code
Collect images
Label bounding boxes in images
Train object detection model
Test object detection model
Deploy model to Google Coral Micro
Test live object detection inference
Taught by
Digi-Key
