YOLOv3: An Incremental Improvement

Benchmark Model Rank Results
classification-on-indlDarknet53#7Average Recall: 88.53%
object-detection-on-coco-oYOLOv3 (DarkNet-53)#41Average mAP: 14.8Effective Robustness: -0.37
pedestrian-detection-on-dvtodYOLOv3 (Thermal)#2mAP: 82.7
pedestrian-detection-on-dvtodYOLOv3 (Visible)#6mAP: 34.5
real-time-object-detection-on-cocoYOLOv3-L#74box AP: 33.0
robust-object-detection-on-cityscapes-1Photometric distortion#10mPC [AP]: 16.9