Relative-Intensity Pattern Registration 0.4.0
Copyright (c) 2026 Jamie Malcolm
Developed at the Brancaccio Lab, UK Dementia Research Institute, Imperial College London.

RIPR source and frozen RIPR scientific engines are provided under BSD-3-Clause.
The frozen scientific resources preserve previously accepted bytecode, not later
experimental recompilations. Their checksums are in
src/main/resources/ripr/longitudinal/runtime.properties.

The native longitudinal bundle also redistributes the following upstream
components. Their upstream licence texts are distributed alongside the frozen
JARs in META-INF/licenses in the plugin and ripr/licenses in the Python package:

- ImageJ 1.54p: https://imagej.net/ij/
- JavaCPP 1.5.11: https://github.com/bytedeco/javacpp
- JavaCPP presets for OpenCV 4.10.0 and OpenBLAS 0.3.28:
  https://github.com/bytedeco/javacpp-presets
- OpenCV: https://github.com/opencv/opencv (Apache-2.0)
- OpenBLAS: https://github.com/OpenMathLib/OpenBLAS (BSD-3-Clause)

JavaCPP and its presets are used under their Apache-2.0 licensing option.
ImageJ is public-domain United States Government work; its disclaimer is at
https://imagej.net/ij/disclaimer.html.

The Windows native JARs retain the registration-only native binaries used in the
accepted benchmark. The JavaCPP native JAR has been repackaged to omit Microsoft
runtime DLLs (and unused LLVM OpenMP); only its original jnijavacpp.dll is retained.
Microsoft's runtime is a separately installed Windows prerequisite, not a component
redistributed under RIPR's licence. The remaining JARs are not general-purpose
OpenCV/OpenBLAS packages. OpenCV third-party licence texts accompany this bundle;
their inclusion does not claim that every listed component is present in it.
OpenCV's Enhanced Correlation Coefficient optimizer and phase correlation are
upstream components, not methods invented by RIPR.
