A Comparison of Robust Background Modeling Methods for Enhancing Event Detection in Video

• Activity Number: 655
• Type: Contributed
• Date/Time: Thursday, August 7, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Imaging
• Author(s): Richard Wood*+ and John Irvine
• Companies: Draper Laboratory and Charles Stark Draper Laboratory
• Keywords: Activity Recognition ; Background Modeling ; Video Image Quailty ; Object Detection ; Object Tracking

Abstract

Automated event recognition in video data has numerous practical applications. The ability to recognize events in practice depends on accurate tracking of objects in the video data. Scene complexity has a large effect on tracker performance. Background models can address this problem by providing a good estimate of the image region surrounding the object of interest. However, the utility of the background model depends on accurately representing current imaging conditions. Changing imaging conditions, such as lighting and weather, render the background model inaccurate, degrading the tracker performance. We have investigated a range of background modeling techniques, which can substantially improve system performance. We present a comparison of several approaches, including methods that incorporate historical lighting and weather conditions, robust statistical models, and fast approximations to robust background estimates. We describe the formulation of these models and discuss model selection in the context of real-time processing. Using results from a recent experiment, we demonstrate empirically the strengths and weaknesses of competing approaches.