Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Gemma's Counting Magic: Bringing Object Detecti...

Gemma's Counting Magic: Bringing Object Detection to Your Android App

In this workshop, we will explore the integration of machine learning into Android applications, focusing on object detection and counting objects using Gemma in combination with TensorFlow Lite. This session is ideal for Android developers looking to integrate powerful AI and machine learning features into their projects, with little or no experience in machine learning required.

Hannah Olukoye

October 16, 2024
Tweet

More Decks by Hannah Olukoye

Other Decks in Technology

Transcript

  1. Gemma's Counting Magic: Bringing Object Detection to Your Android App

    Engineering Manager | Google Developer Expert Hannah Olukoye
  2. "“It is hard to show your skill as a sailor

    when there is no wind.” Good Strategy, Bad Strategy - Richard P. Rumelt
  3. Importance of AI in Android Applications 2 1 3 Enhanced

    Experience Advanced Functionality Automation of Tasks
  4. Gemma? • Gemma 2 – Scalable power for all platforms,

    • CodeGemma – Your coding co-pilot, optimized for AI-assisted development • PaliGemma – Handles both text and images seamlessly across devices. • RecurrentGemma – Master of sequential data tasks, in any size. • ShieldGemma – Fortified for security, perfect for sensitive data handling.
  5. TensorFlow Lite is a ML framework optimized to run pre-trained

    models on mobile devices like Android smartphones. TensorFlow Lite? • Mobile-Optimized • Pre-built Models • Model Conversion • Core components: • Interpreter: Runs models on-device. • Converter: Converts models to TFLite format. • Hardware acceleration
  6. Setup On Android Studio 1. Add TensorFlow Lite Dependencies 2.

    Add your TensorFlow Lite Model 3. Add your Labels File