gocv: Go language bindings for the OpenCV 4 computer vision library

What it solves

GoCV provides Go language bindings for OpenCV 4, allowing developers to use the powerful computer vision library within the Go ecosystem. It solves the problem of outdated or stalled bindings for newer OpenCV versions, enabling Go programmers to access modern computer vision algorithms and hardware acceleration.

How it works

The project uses a C-style wrapper around OpenCV 4 C++ classes. This approach avoids the complexity of using SWIG and ensures the mappings closely follow the original OpenCV project structure, making it the easier to locate and implement new features.

Who it’s for

Go developers who need to implement computer vision tasks such as face detection, object counting, and video streaming in their applications.

Highlights

• Hardware Acceleration: Supports CUDA for Nvidia GPUs and Intel OpenVINO for optimized performance.
• Cross-Platform: Compatible with Linux, Docker, macOS, and Windows.
• Memory Profiling: Includes a custom Mat profiler to help developers detect and repair memory leaks caused by C-based image allocations.
• Broad Functionality: Supports a wide range of utilities, including DNNs for face tracking and MJPEG streaming.