Who is this for?

This Learning Path is for Android and machine learning developers who want to run an optimized image-segmentation model locally on an Arm-based Android phone.

What will you learn?

Upon completion of this Learning Path, you will be able to:

  • Prepare the Android command-line tools and connect an Arm-based Android phone
  • Download and run the MobileSAM ExecuTorch model from the Arm AI Portal
  • Understand how the Android adapter validates, prepares, and renders MobileSAM results
  • Optionally inspect an unsupported image model before implementing a model-specific adapter

Prerequisites

Before starting, you will need the following:

  • A macOS, Linux, or Windows development computer with Git, Python 3, and Java 17 or later
  • An Arm-based Android phone running Android 9 or later
  • A Hugging Face account
  • A data-capable USB cable
  • Basic familiarity with terminal commands and Android applications
  • Network access for the first Gradle build and model download

Summary

AI-assisted

This summary was drafted with an approved AI-assisted workflow and reviewed by Arm contributors before publication. Human technical review remains part of the process so the final page reflects engineering rigor, accuracy, and Arm editorial standards.

Close
?
You’ll build Image Analysis and use its validated ExecuTorch adapter to run MobileSAM locally on an Arm-based Android phone. You’ll prepare the Android SDK, connect the phone, and let Gradle provision the JDK 17 build toolchain. You’ll then download and stage the model, generate and validate masks from two images, and inspect the adapter’s preprocessing and output checks. An optional final section helps you plan and validate another model-specific adapter.

Frequently asked questions

AI-assisted

These FAQs were drafted with an approved AI-assisted workflow and reviewed by Arm contributors before publication. Human technical review remains part of the process so the final page reflects engineering rigor, accuracy, and Arm editorial standards.

Close
?
How do I verify my Android phone is connected and authorized?
Enable Developer options and USB debugging, then run adb devices -l. The output should list the device with its serial and show it as authorized. If it reports unauthorized, unlock the phone and accept the debugging prompt. On Windows, you might also need the phone manufacturer’s USB driver.
How does the project obtain the JDK 17 build toolchain?
Your installed java and javac commands must report JDK 17 or later so that Gradle can start. The project configures Gradle to use a compatible JDK 17 build toolchain and download one automatically when it isn’t already available.
How do I run MobileSAM on the Android phone?
Download mobile_sam_raspberry_executorch_optimized.pte. Copy it into the directory named by the MobileSAM catalog entry under application-private storage, then start Image Analysis. Select Load model, choose an image, and select Run segmentation to generate the mask.
What confirms that MobileSAM produced a valid result?
Check that the application displays a translucent cyan mask over a plausible object boundary. Confirm that Image Analysis reports finite intersection over union (IoU), coverage, and logit values. Run a second image and confirm that the mask changes with the input.
Can Image Analysis run another model without code changes?
Image Analysis already includes validated ExecuTorch adapters for MobileSAM and Depth Anything V2. Another model needs a compatible catalog entry and a model-specific adapter. The included LiteRT and ONNX adapter files are stubs that need implementation and device validation before use.
Next