Pedestrian Detection and Tracking

Authors:

Abstract
Pedestrian detection and tracking are critical components of modern surveillance systems, playing a vital role in various applications such as public safety, autonomous driving, and urban planning. This chapter delves into the fundamental concepts, methodologies, and technological advancements that have shaped the field of pedestrian detection and tracking. Beginning with an overview of traditional methods, including background subtraction and feature-based approaches, the chapter transitions into contemporary deep learning techniques that have significantly improved detection accuracy and robustness. Key algorithms, such as Convolutional Neural Networks (CNNs), Region-based CNNs (R-CNNs), and more recent advancements like Transformer-based models, are explored in detail. The chapter also addresses the integration of these algorithms into real-time tracking systems, discussing object association techniques, motion models, and multi-object tracking strategies.

This publication has 8 references indexed in Scilit: