一、摘要

We introduce YOLO9000, a state-of-the-art, real-time object detection system that can detect over 9000 object categories. First we propose various improvements to the YOLO detection method, both novel and drawn from prior work. The improved model, YOLOv2, is state-of-the-art on standard detection tasks like PASCAL VOC and COCO. Using a novel, multi-scale training method the same YOLOv2 model can run at varying sizes, offering an easy tradeoff between speed and accuracy. At 67 FPS, YOLOv2 gets 76.8 mAP on VOC 2007. At 40 FPS, YOLOv2 gets 78.6 mAP, outperforming state-of-the-art methods like Faster R-CNN with ResNet and SSD while still running signifificantly faster. Finally we propose a method to jointly train on object detection and classifification. Using this method we train YOLO9000 simultaneously on the COCO detection dataset and the ImageNet classifification dataset. Our joint training allows YOLO9000 to predict detections for object classes that don’t have labelled detection data. We validate our approach on the ImageNet detection task. 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. YOLO9000 predicts detections for more than 9000 differentobject categories, all in real-time.

准确率更高（提出了7个tricks：1. Batch Normalization 2. High Resolution Classifier 3.Convolutional With Anchor Boxes 4. Dimension Clusters 5. Direct location prediction 6. Fine-Grained Features 7. Multi-Scale Training）
能够允许任意尺度的图片的输入（7. Multi-Scale Training）
速度更快，提出新的骨干网络，Darknet-19
类别更多，能够识别9000多种类别（待定）

二、Better——更准确

针对Yolo v1以下的4个缺点：

MAP比较低
YOLOv1将输入图像分成7*7个grid cell，而每个grid cell 只能预测一个物体，说明YOLO v1 只能预测49个物体，也就是recall比较低（把全部目标检测出的能力比较弱）
即YOLO v1 对密集和小目标的检测效果不佳