A machine-learning-based algorithm for detecting a moving object

Abstract

A machine-learning-based algorithm for detecting a moving object is proposed in this paper. It is composed of an object-detection model, and a machine-learning model. The object-detection model is used to detect the object by applying the random ferns method, while the machine-learning model is used to maintain the object expression in the object detecting model by online learning. Using the proposed algorithm, a moving object can be detected in real-time environment with robustness, such as illumination change, highspeed motion, object occlusion, object distortion, and noisy object. The effectiveness and the efficiency of the proposed algorithm are demonstrated by experiments in comparison with other models.