神经网络引入后，检测框架变得更快更准确。然而，大多数检测方法受限于少量物体。检测和训练数据上联合训练物体检测器，用有标签的检测图像来学习精确定位，同时用分类图像来增加词汇和鲁棒性。原YOLO系统上生成YOLOv2检测器；在ImageNet中超过9000类的数据和COCO的检测数据上，合并数据集和联合训练YOLO9-After the neural network is introduced, it is becoming faster and more accurate detection fr

非线性分叉求解软件，coco，适用于MATLAB 2015A(This is a summary of the instructions found at <https://sourceforge.net/p/cocotools/wiki/installation/>. Installing and using COCO ========================= In the current version, tutorials are available for