Background Subtraction Using Background Sets With Image- and Color-Space Reduction

Background subtraction is a basic step for a variety of multimedia applications such as live video, traffic monitoring, communication system, interactive learning space, etc. Many approaches have been proposed for this problem, but the need for lower cost approaches still exists. In this paper, a relatively low-cost background-subtraction method is proposed, using background sets with image- and color-space reduction. The background sets are used to detect objects from dynamic backgrounds, which contain waves, trees, and fountains. The image space is reduced to deal with jittered and unsteady frames, e.g., the input from handheld mobile devices. The color space is reduced to compensate for color noise, e.g., the scattered RGB values of a digital camera. To reduce the cost, a combination of color-space reduction and hash-table look-up operations are used. The results compared with other methods show the feasibility of our method; moreover, it can be useful in mobile or embedded environments.