Visual Object Tracking: A survey of methods and a comparative analysis on VOT2022

Visual object tracking is a fundamental problem in computer vision with primary surveillance and autonomous systems applications. Over the years, numerous tracking methods have been developed, leveraging different principles and architectures to enhance accuracy and robustness. This paper provides an overview of various tracking approaches, highlighting key trends and advancements in the field. Additionally, a comparative analysis is conducted between two prominent trackers from the family of adaptive correlation filters and Siamese neural networks. The discussion covers their general methodologies, advantages, and challenges, offering insights into their relative performance in different tracking scenarios.