YOLOv4: A Breakthrough in Real-Time Object Detection

YOLOv4 achieved the best performance on the COCO dataset by combining advanced techniques for regression (bounding box positioning) and classification (object class identification) using the Darknet framework. To enhance accuracy and adaptability, it employs Cross mini-Batch Normalization, Cross-Stage-Partial-connections, Self-Adversarial-Training, and Weighted-Residual-Connections, as well as CIoU loss, Mosaic data augmentation, and DropBlock regularization. With Mosaic augmentation and multi-resolution training, YOLOv4 achieves superior detection in diverse scenarios, attaining 43.5\% AP (in contrast, 65.7\% AP50) on a Tesla V100 at ~65 frames per second, ensuring efficiency, affordability, and adaptability for real-world environments.

Athulya Sundaresan Geetha · 2025