A Review of Object Detection and Tracking Methods

Abstract

Moving object tracking in videos is an actively researched area for the last two decades due to its practical
applications in many areas, such as event analysis, human computer interaction, crowd analysis, etc. Extensive research
reported some success applications in highly constrained domains, but there are still many challenges that need to be
resolved, such as abrupt object motion, changes in appearance of the object, non-rigid objects, occlusion, illumination,
etc. This paper presents a comprehensive review of various object tracking approaches reported in literature and
proposes a new categorization to group various tracking approaches to streamline future researches. It also discusses, in
detail, methods used in each category for tracking single and multiple objects. Finally, the paper concludes by setting
directions for further research.