Moving object classification using bayesian networks

Article is devoted to the moving object classifier. This classifier can be used to trace an object in a moving object detection system together with the other methods. The classifier is based on Bayesian network and uses an object's color distribution histogram. This feature allows to classify objects overlaid by other objects, rotated, reduced or partially distorted objects. The description and formalization of this approach is performed, its advantages and shortcomings identified during testing ate indicated as well.