Approximating images using minimum bounding rectangles

Abstract

In surveillance systems, video cameras record specific scenes for long times. However, by the end of the recording period, such tapes may hold many useless scenes which need to be eliminated. In order to reduce the time in reviewing these worthless scenes while seeking for a specific object, an approximation technique must be considered. In this article, we propose a new technique for finding the minimum bounding rectangle of objects which appear in a specific bitmap image. The minimum bounding rectangle of an image object is the Rectangle containing the object such that the sides of the rectangle touch the object boundaries. The dimensions and locations of the minimum bounding rectangle of an object can be utilized as features to identify the corresponding object.