Segmentation of Moving Images

Abstract

The segmentation and detection of objects in a scene is one of the most frequent tasks in digital image processing, being also one of the most important. Normally it involves separating the object from a statistically distinct background. When the background and object in the scene have the same statistical properties, problems arise. The method presented uses temporal information about the scene to aid in the detection of moving objects. This work tries to bring an object with the exact same statistical characteristics of the background, from its camouflaged position, by the use of motion analysis instead of static picture analysis.