bgslibrary: a comprehensive C++ framework for foreground-background separation in video streams

What it solves

BGSLibrary provides a unified framework for background subtraction in computer vision, allowing users to easily separate moving foreground objects from a static or evolving background in video streams. It eliminates the need to implement multiple background subtraction algorithms from scratch by providing a comprehensive collection of pre-implemented techniques.

How it works

Built as a C++ framework that leverages the OpenCV library, BGSLibrary implements a wide variety of background subtraction algorithms (over 40). It uses a factory pattern to allow developers to instantiate and apply different algorithms to video frames to generate foreground masks and background models. The library is cross-platform and provides wrappers for Python, Java, and MATLAB to make these tools accessible to different programming environments.

Who it’s for

This library is designed for researchers and developers working in computer vision, particularly those focused on video surveillance, motion detection, and foreground-background segmentation.

Highlights

• Extensive Algorithm Library: Includes over 40 different background subtraction algorithms, such as ViBe, KNN, and Mixture of Gaussians.
• Multi-Language Support: Offers wrappers for Python (pybgs), Java, and MATLAB.
• Harnesses OpenCV: Fully compatible with OpenCV versions 2.4.x, 3.x, and 4.x.
• Flexible Build System: Supports building via CMake or Pixi for simplified dependency management.