High-Performance Image Analysis and Change Detection Engine
Aquamarine is a highly optimized, precision-engineered tool designed for advanced image recognition, detection, and analysis. Its core capability lies in identifying delta-changes (differences) between images, making it exceptionally suited for movement detection, tracking object appearances and disappearances, and filtering out environmental noise such as lighting changes or micro-movements (e.g., wind affecting foliage).
The primary objective of Aquamarine is to provide an optimal, computationally efficient solution for recognizing image characteristics within strictly defined hardware constraints.
The engine outputs analytical bounding regions—rectangular coordinates identifying areas where significant color or pixel-cluster shifts have occurred. This allows downstream systems to process only the relevant telemetry rather than scanning the entire image payload.
Note: Formal compliance and stress testing stages are currently in progress.
Aquamarine includes a standalone, cross-platform graphical user interface powered by Qt. This GUI provides real-time telemetry and visualization of the detection algorithm, allowing engineers to interactively adjust detection parameters (e.g., affinity thresholds, calculation timeouts) and instantly view the resultant bounding regions overlaying the base imagery.
To build the standalone Qt application, Qt6 is required. The build system will automatically detect the presence of Qt6 and compile the graphical interface alongside the core library. If Qt6 is absent, the build system gracefully falls back to compiling only the headless backend service.
macOS:
brew install qt
# Note: Ensure Qt is linked or its path is accessible by CMake.
Linux (Debian/Ubuntu):
sudo apt update
sudo apt install qt6-base-dev qt6-base-dev-tools libqt6svg6-dev
Windows:
PATH, or set CMAKE_PREFIX_PATH to your Qt installation directory during the CMake configuration step.For detailed dependency requirements and build logic, please refer to the Build Documentation.
You can build all targets natively using the provided build script:
./build.sh
Compliance and Usage Policies:
For inquiries, licensing, or suggestions, please contact the author directly via private message or email.