Particle Filter Based Object Tracking with Sift and Color Feature

Abstract

Visual object tracking is an important topic in multimedia technologies. This paper presents robust implementation of an object tracker using a vision system that takes into consideration partial occlusions, rotation and scale for a variety of different objects. A scale invariant feature transform (SIFT) based color particle filter algorithm is proposed for object tracking in real scenarios. The Scale Invariant Feature Transform (SIFT) has become a popular feature extractor for vision based applications. It has been successfully applied for metric localization and mapping. Then the object is tracked by a color based particle filter. The color particle filter has proven to be an efficient, simple and robust tracking algorithm. Experimental results of applying this technique show improvement in tracking and robustness in recovering from partial occlusions, rotation and scale.