YOLOv4: Optimal Speed and Accuracy of Object Detection

23 Apr 2020  ·  Alexey Bochkovskiy, Chien-Yao Wang, Hong-Yuan Mark Liao ·

There are a huge number of features which are said to improve Convolutional Neural Network (CNN) accuracy. Practical testing of combinations of such features on large datasets, and theoretical justification of the result, is required. Some features operate on certain models exclusively and for certain problems exclusively, or only for small-scale datasets; while some features, such as batch-normalization and residual-connections, are applicable to the majority of models, tasks, and datasets. We assume that such universal features include Weighted-Residual-Connections (WRC), Cross-Stage-Partial-connections (CSP), Cross mini-Batch Normalization (CmBN), Self-adversarial-training (SAT) and Mish-activation. We use new features: WRC, CSP, CmBN, SAT, Mish activation, Mosaic data augmentation, CmBN, DropBlock regularization, and CIoU loss, and combine some of them to achieve state-of-the-art results: 43.5% AP (65.7% AP50) for the MS COCO dataset at a realtime speed of ~65 FPS on Tesla V100. Source code is at https://github.com/AlexeyAB/darknet

PDF Abstract

Results from the Paper


Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Object Detection COCO-O YOLOv4-P6 Average mAP 30.4 # 18
Effective Robustness 5.89 # 20
Object Detection COCO test-dev YOLOv4-608 box mAP 43.5 # 152
AP50 65.7 # 70
AP75 47.3 # 99
APS 26.7 # 82
APM 46.7 # 90
APL 53.3 # 115

Methods