Cloud Optimized OpenCV For AWS Graviton4

STEP 1: LAUNCH THE AMI

After subscribing, launch an EC2 instance using this AMI.

Recommended instance types:

• m8g.4xlarge or larger for optimal performance on AWS Graviton4
• Compatible instance families: c8g, m8g, r8g

STEP 2: CONNECT TO THE INSTANCE

Connect using SSH:

STEP 3: ACTIVATE THE OPTIMIZED ENVIRONMENT (RECOMMENDED)

The optimized libraries are installed under:

COOL provides preconfigured Python virtual environments that automatically set all required paths.

Available Python environments:

• python_3.10
• python_3.11
• python_3.12

Example: Activate Python 3.12

Verify that the optimized OpenCV build is active:

OPTIONAL: USING SYSTEM PYTHON

If you prefer using the default system Python or an existing environment, export the required paths manually:

• export LD_LIBRARY_PATH="/opt/cool/cpp_sdk/lib:$LD_LIBRARY_PATH"
• export PYTHONPATH="/opt/cool/python_3.12/site-packages/"

Then run your script normally:

• python3.12 your_script.py

OPTIONAL: BUILDING CUSTOM C++ APPLICATIONS

To compile your own C++ applications against the optimized COOL OpenCV libraries:

1. Create a build directory in your project workspace.

2. Create a CMakeLists.txt file with the following content:

   cmake_minimum_required(VERSION 3.10)

   project(MyApp)

   set(CMAKE_CXX_STANDARD 11)

   set(OpenCV_DIR "/opt/cool/cpp_sdk/lib/cmake/opencv4")

   find_package(OpenCV REQUIRED)

   add_executable(my_app main.cpp)

   target_link_libraries(my_app ${OpenCV_LIBS})

3. Build and run:

   cd build

   cmake ..

   make

   ./my_app

NOTE: Ensure LD_LIBRARY_PATH includes /opt/cool/cpp_sdk/lib before running C++ binaries.