SpatialML XR Utils

Reusable NativeActivity/OpenXR base sources for SpatialML XR Android apps.

This repository packages the SecureMR-Samples-Private/base C++ app base as a small Android Studio library project plus a CMake project declaration that SpatialML XR samples can use.

CMake Usage

Add this repository as a submodule in an Android app project, then include the CMake integration file:

include(path/to/SpatialML-XR-Utils/cmake/SpatialMLXRUtils.cmake)

spatialml_xr_project(
  TARGET your_native_target
  SOURCES src/main/cpp/your_runner.cpp
)

The consuming app still owns its Android manifest and its app-specific CreateSecureMrProgram(...) implementation. The shared project declaration owns the native entrypoint, OpenXR setup, platform/graphics code, shaders, OpenXR loader target, and securemr_utils sources.

Package Loading

The utility API includes:

• SecureMrUtils::LoadModelPackagePipelinesFromAssets(...)
• SecureMrUtils::LoadModelPackagePipelinesFromFiles(...)

The file loader supports pySpatialML schema v2 packages:

• top-level schema_version: "2"
• top-level pipelines: [{id, path}]
• inline LiteRT/TFLite model metadata on run_algorithm operators under operator.model
• package-relative model paths such as model/face_detector.tflite
• package-relative glTF assets such as gltf/frame.gltf

When running with supplied image tensors instead of device VST camera input, pass ModelPackageLoadOptions{.stripRectifiedVstAccess = true} to remove RECTIFIED_VST_ACCESS operators before deserialization.