Multi-level scene understanding via hierarchical classification

Abstract

In applications where the use of video surveillance is necessary and/or beneficial, it is a common goal to identify the contents of the video automatically. Of particular interest in such applications is the ability to recognize locations in the environment, where events occur, and describe the events common to those locations. This is one of the goals of scene understanding.Scene understanding is traditionally addressed from one of two separate points-of-view: the description of the underlying environment or the action taking-place throughout the scene. Each of these facets is required to address the overarching goal but, is insufficient independently to address the problem entirely. These facets are, in fact, dependent and by considering both, a more complete description becomes available. In this paper, we describe a novel, data-driven scene understanding and classification technique that captures and utilizes information about both the environment and activity within a scene.

Extraction and classification of moving targets in multi-sensory MAMI-1 data collection

“…1. The output of this optimization, as described in [5], is two components: the foreground, or movers, and the smoothly changing background. Figure 4 provides an illustrative example of the candidate detections produced by this portion of our work.…”

“…The mover extraction is performed as described in [5] by a similar low-rank/sparse decomposition to separate the movers (sparse part) from the backgroud (low-rank component). Here we show that by appropriately parameterizing the objective function of the decomposition, we can use the same tool for both tasks.…”

“…We apply techniques based on the sparse and low-rank matrix decomposition [4], [5] to extract moving objects from a sequence of images. This process creates a set of candidate detections for our classifier.…”