YOLO9000: Better, Faster, Stronger

CVPR 2017 Joseph Redmon • Ali Farhadi

At 67 FPS, YOLOv2 gets 76.8 mAP on VOC 2007. YOLO9000 gets 19.7 mAP on the ImageNet detection validation set despite only having detection data for 44 of the 200 classes. On the 156 classes not in COCO, YOLO9000 gets 16.0 mAP.

PDF Abstract

Code


Evaluation


Task Dataset Model Metric name Metric value Global rank Compare
Object Detection COCO YOLO v2 + Darknet-19 Bounding Box AP 21.6 # 23
Object Detection PASCAL VOC 2007 YOLO v2 MAP 78.6% # 8