This chapterLocalization presents a thorough overview of object detection and localization, essential tasks in computer vision that allow machines to detect and localize objects in images or video frames. It discusses the historic development of these techniques, from the early days of handcrafted feature-based methods such as Haar cascadesHaar cascades and HOG to the game-changing rise of deep learningDeep learning. This chapter describes the role of a Convolutional NeuralNeural networks NetworkConvolutional Neural Networks (CNN) in automatically extracting features and supporting end-to-end learning, making models the foundation of modern approaches. Furthermore, this chapter provides an extensive taxonomy of recent architectures, including two-stage detectorsTwo-stage detectors (e.g., FasterR-CNN R-CNNFaster R-CNN) that are optimized for high accuracy, single-stage detectors (e.g., YOLOYOLO, SSDSingle shot multibox detector) that maximize real-time performance, and emerging anchor-free and transformerTransformers-based approaches (e.g., DECTER). The discussion also highlights several key challenges in the area, including scale variability, occlusionsOcclusion, real-time inference requirements, and the requisite for large annotated datasets. Additionally, the chapter demonstrates the significance of object detection for many real-world applications, such as autonomous drivingAutonomous driving, surveillance, and medical imagingMedical imaging. Lastly, it outlines the main differences between the related tasks of object detection–which detects objects within individual frames–and objectTracking trackingObject tracking, which preserves identity across a video sequence. This introductory picture is necessary to frame up in the forthcoming chapters where we consider in depth methodologies, applications, and future directions.