Randomized Low-Rank Dynamic Mode Decomposition for Motion Detection

TL;DR

This paper presents a fast, randomized low-rank Dynamic Mode Decomposition algorithm for motion detection in videos, demonstrating its efficiency and effectiveness for real-time background subtraction compared to existing methods.

Contribution

It introduces a novel randomized low-rank DMD method tailored for motion detection and background subtraction, with comprehensive evaluation and real-time processing capabilities.

Findings

Efficient real-time background subtraction in high-resolution videos.

Competitive performance against robust principal component analysis algorithms.

Effective for motion detection in static video sources.

Abstract

This paper introduces a fast algorithm for randomized computation of a low-rank Dynamic Mode Decomposition (DMD) of a matrix. Here we consider this matrix to represent the development of a spatial grid through time e.g. data from a static video source. DMD was originally introduced in the fluid mechanics community, but is also suitable for motion detection in video streams and its use for background subtraction has received little previous investigation. In this study we present a comprehensive evaluation of background subtraction, using the randomized DMD and compare the results with leading robust principal component analysis algorithms. The results are convincing and show the random DMD is an efficient and powerful approach for background modeling, allowing processing of high resolution videos in real-time. Supplementary materials include implementations of the algorithms in Python.