Object tracking techniques and performance measures — A conceptual survey

Abstract

Object tracking is an important computer vision chore which has key applications such as human computer interaction (HCI), visual surveillance, video compression, etc. In spite of widespread research on this topic, the majority of object tracking algorithms suffers from complications in treating complex object appearance changes caused by factors such as illumination variation, partial occlusion, shape deformation, and camera motion. Hence, operative modeling of 2D appearance of tracked objects is a crucial issue for the successful operation of a tracker. The main goal of computer vision is to permit computers to imitate the primary to complex functions of human vision to a minimum degree of acceptance. In order to achieve this goal, a significant effort has been made on object tracking which is a stimulating research topic in computer vision. A large number of techniques have been proposed for object tracking. Along with the techniques, different design metrics are also proposed. In this paper an exhaustive survey has been conducted on both the techniques and design metrics.