Worked on enhancing the OpenCV image processing stack by introducing a new parameter, itersPerLevel, to the ECCParameters class within the opencv/opencv repository. This addition improved the flexibility and accuracy of the ECC multi-scale alignment algorithm, allowing for more precise control over image registration tasks. The work involved C++ development for the core feature and Python development to update the wrapper, ensuring API parity between languages. A dedicated integration test was added to verify the new parameter’s behavior, and compatibility with existing workflows was maintained. The changes addressed developer experience and closed gaps between the C++ and Python APIs.