On the application of Cellular Logic Array Processing and Background Subtraction Methods for the Improved Object Detection in Video Surveillance using Local Neighborhood Difference

The research article proposes an innovative method of object motion detection techniques using “Local Neighborhood Differencing” in statistical background subtraction (BGS) procedures of video surveillance. Existing methods such as Adaptive Mean (AM) and Adaptive Median (AMD) BGS methods are fast but they fail in challenging environments of high noise and dynamic background. This paper uses an innovative procedure of differencing based on the local neighborhood in which instead of finding the difference between a pixel of the current frame and the corresponding pixel of background frame, the average difference of whole of the neighborhood is found and then thresholded for object motion detection in the current frame. A very effective improvement over basic method is verified in the experiments.