CMSC 838S Information Visualization- Spring, 2006

Using Time Series Analysis to Visualize and Evaluate Background Subtraction Results for Computer Vision Applications

Samah Ramadan

sramadan@cs.umd.edu

Abstract:

Information visualization has proved its significance and importance for exploring and analysis of complex data. This paper introduces a novel framework for applying time series analysis and visualization methods to evaluate and compare the results of background subtraction algorithms for computer vision applications. Instead of judging the results of background subtraction algorithms by just viewing the resulted binary sequences which might not give enough insights about the results, we apply time series analysis and visualization to provide more details and insights of the results. The objective is to evaluate the results and possibly help to identify the weaknesses and strengths of background subtraction algorithms which leads to further improvements for the performance of these algorithms.

 

Report: pdf

Matlab Code: code

Demo: demo

Screenshots:

 


 

Web Accessibility