OpenCV 5 brings new deep neural network engine, stronger ONNX support, and faster core

OpenCV 5 has been released as a major new version of the widely used open source computer vision library. The update delivers several foundational changes including a new deep neural network engine, stronger ONNX support, improvements to hardware acceleration, enhanced Python integration, support for new data types, expanded 3D vision capabilities, improved documentation, and a cleaner overall architecture.

Building on its long-standing role in computer vision, robotics, artificial intelligence, augmented and virtual reality, and embedded systems, OpenCV now sees more than 88,000 GitHub stars and over a million installs daily. This release marks one of the most substantial in its history, moving beyond routine updates to modernize the library for today’s requirements.

While OpenCV 5 advances technical capabilities, it also addresses the increasing demand to develop applications that combine classical vision, deep learning models, edge deployment, and hardware heterogeneity. The release aims for a faster, smaller core, better language support, updated APIs, enhanced DNN performance, broader hardware acceleration, better 3D tooling, and more accessible documentation. For a deep dive into what is new and how these changes can affect user code, refer to the official OpenCV announcement.