Image Segmentation in Video Sequences: A Probabilistic Approach

TL;DR

This paper introduces a probabilistic, mixture-of-Gaussians model for pixel classification in video segmentation, improving accuracy in detecting slow-moving objects and shadows over traditional background subtraction methods.

Contribution

It presents an unsupervised, incremental EM algorithm for learning pixel models, enhancing video segmentation by better handling slow objects and shadows.

Findings

Improved detection of slow-moving objects.

More effective shadow elimination.

Enhanced vehicle tracking in traffic surveillance.

Abstract

"Background subtraction" is an old technique for finding moving objects in a video sequence for example, cars driving on a freeway. The idea is that subtracting the current image from a timeaveraged background image will leave only nonstationary objects. It is, however, a crude approximation to the task of classifying each pixel of the current image; it fails with slow-moving objects and does not distinguish shadows from moving objects. The basic idea of this paper is that we can classify each pixel using a model of how that pixel looks when it is part of different classes. We learn a mixture-of-Gaussians classification model for each pixel using an unsupervised technique- an efficient, incremental version of EM. Unlike the standard image-averaging approach, this automatically updates the mixture component for each class according to likelihood of membership; hence slow-moving objects are handled perfectly. Our approach also identifies and eliminates shadows much more effectively than other techniques such as thresholding. Application of this method as part of the Roadwatch traffic surveillance project is expected to result in significant improvements in vehicle identification and tracking.