Foreground Detection in Dynamic Scenes using Robust Principal Component Analysis in comparison with Gaussian Mixture Model to measure F-score

Abstract

Aim: This paper describes the poor performance of accuracy and F-Score in the detection of moving objects using a novel
Robust Principal Component Analysis (RPCA). Materials and Methods: ClinCalc is a tool to compute sizes and display the
results of sample analyses. Cdnet 2014 dataset demonstrates our foreground detection using Robust Principal Component
Analysis algorithm. In this study, the Robust Principal Component Analysis is compared with the Gaussian Mixture model.
Results: Poor detection of moving objects is improved by using Robust Principal Component Analysis algorithm with mean
F-score rate of 83% and accuracy of 87%. Conclusion: Robust Principal Component Analysis subtracts the background and
detects the moving objects by measuring the f-score and accuracy of the dataset.