Publication Date: 2023/03/08
Abstract: Multiple Object Tracking (MOT) is the detection of unique objects and their movements through frames. There are various issues in Multiple Object Tracking like occlusions, similar appearance of different objects, interaction among multiple objects, etc. There are even more issues with Object Tracking in aerial image sequences due to weather conditions, small sizes of objects, distortion in proportions, etc. Even with the various techniques and methods developed for MOT, only a fraction are able to perform on aerial image sequences with the rest focusing on ground level image sequences. Research in Multiple Object Tracking has been gaining a lot of attention with an increasing trend in the number of Multiple Object Tracking research papers published each year. This paper focuses on documenting the advancements in Multiple Object Tracking in recent years, with extra attention paid to techniques that help with aerial image sequences. The various architectures and techniques are classified and are compared along with their advantages, disadvantages and performance scores, with detailed mention of the datasets generally used for various applications.
Keywords: Multiple object tracking, aerial images, attention networks, occlusion, pedestrian tracking.
DOI: https://doi.org/10.5281/zenodo.7707008
PDF: https://ijirst.demo4.arinfotech.co/assets/upload/files/IJISRT23FEB961_(1).pdf
REFERENCES