Comprehensive Analysis and Evaluation of Background Subtraction Algorithms for Surveillance Video

Yan Feng, Shengmei Luo, Yumin Tian, Shuo Deng, Haihong Zheng
2014 Sensors & Transducers  
Subtraction techniques are the basis for moving target detection and tracking in the domain of video surveillance, while the robust and reliable detection and tracking algorithms in complex environment is a challenging subject, so evaluations of various background subtraction algorithms are of great significance. Nine state of the art methods ranging from simple to sophisticated ones are discussed. Then the algorithms were implemented and tested using different videos with ground truth, such as
more » ... baseline, dynamic background, camera jitter, and intermittent object motion and shadow scenarios. The best suited background modeling methods for each scenario are given by comprehensive analysis of three parameters: recall, precision and F-Measure, which facilitates more accurate target detection and tracking.
doaj:b8def3d98b794f9f835e1c96ad226912 fatcat:djuxsqujsze27kwsqqfil735ly