Object detection and tracking have become fundamental research areas in computer vision due to their widespread applications in intelligent surveillance, autonomous vehicles, robotics, healthcare, traffic monitoring, sports analytics, and smart city systems. Traditional object detection techniques relied on handcrafted features and conventional machine learning algorithms, which often struggled under challenging environmental conditions. Recent advances in deep learning have significantly improved the accuracy and speed of object detection and multi-object tracking systems. This review presents a comprehensive analysis of conventional and modern object detection and tracking approaches reported in recent literature. Various image preprocessing techniques, feature extraction methods, object detection algorithms including R-CNN, Fast R-CNN, Faster R-CNN, SSD, YOLO, RetinaNet, EfficientDet, and transformer-based models such as DETR are examined. In addition, popular tracking algorithms including SORT, Deep SORT, ByteTrack, OC-SORT, and StrongSORT are discussed. Publicly available datasets and evaluation metrics such as Mean Average Precision (mAP), Intersection over Union (IoU), Multiple Object Tracking Accuracy (MOTA), Multiple Object Tracking Precision (MOTP), Precision, Recall, and F1-score are also reviewed. Furthermore, this paper highlights current challenges including occlusion, illumination variation, crowded scenes, real-time processing, and computational complexity while identifying future research opportunities involving explainable artificial intelligence, Vision Transformers, multimodal perception, edge computing, and lightweight object detection models. This review provides researchers with a comprehensive understanding of recent developments and emerging trends in automated object detection and tracking systems.
