Live Coding: Adding Machine Learning to Android Things in 10 Mins

on YouTube ↗21m

" summary

Live-codes Firebase ML Kit into an Android Things camera pipeline: capture bitmap, wrap as FirebaseVisionImage, call VisionLabelDetector with a 50% confidence threshold and listeners. Demonstrates Firebase ML Kit on-device (~400 labels, low latency) versus cloud (>10,000 labels, higher accuracy) with example confidences (83%, 68%, 94%). Builds a custom TensorFlow “hoy or not” model from ~60 personal photos, gets spurious 99% hits due to dataset bias, and extracts practical fixes: control input, use quantization, and match float vs uint data types.

AI-written summary · about

" chapters (7)

1. 0:00Introduction and agenda
2. 0:43ML in Wear OS products
3. 3:58ML Kit overview and demo
4. 5:44Live coding ML Kit integration
5. 13:23Training custom TensorFlow models
6. 17:45Lessons learned from custom models
7. 20:12Best practices for ML adoption

" related talks

• 24mML Kit: Machine learning for mobile developersSachin Kotwani, Isabella Chen · Firebase Summit 2018
• 14mDo you want an easy way to add Machine Learning into your app?Sandra Dupré · Android Makers 2019
• 47mExploring Firebase MLKit on AndroidHiren Dave · GDG DevFest Ahmedabad 2018
• 30mMachine Learning in Android Apps with ML KitFelipe Pedroso · AndroidFest Campinas 2018
• 34mIt’s an ML full world! MLKit for Android DevsBritt Barak · DevFest Ukraine 2018
• 36mMachine Learning for Mobile DevelopersLaura Morinigo · Mobilization 8
• 04mTop 5 reasons to add ML Kit to your Android or iOS appSachin Kotwani · Firebase Summit 2018
• 19mMaking Your Android Things Devices SmarterPaul Trebilcox-Ruiz · DevFest Florida 2019