COVA

COVA is an on-premise desktop app that automatically anonymizes faces in videos and images while preserving as much visual quality as possible. It analyzes each face's pose and brightness, then picks the minimum blur or mosaic strength that still prevents re-identification, even against AI restoration tools like Real-ESRGAN and GFPGAN. Built with Python, PyQt5 and OpenCV, COVA runs fully on your local machine, making it ideal for CCTV, medical recordings, research datasets and broadcast content. Features include live Original/Anonymized preview, support for images and video, adjustable risk levels and attack models, and timeline graphs showing angle, brightness and anonymization strength.