matlab faster-rcnn 测试与训练

编译与准备测试

（1）运行faster_rcnn_build.m。这是编译了nms和nms_gpu的mex，nvmex.m中的所有环境变量、VS是安装路径要和自己的一致：

（2）运行startup.m，这是设置基本环境。

测试demo

（1）运行fetch_data/fetch_faster_rcnn_final_model.m，下载训练好的模型。如果onedrive用不了或者太慢，去百度云。下载好以后把它解压到faster_rcnn-master文件夹下，可以看到是5个图片和output文件夹。 
（2）运行experiments/script_faster_rcnn_demo.m，测试单张图像。（注意，要在faster_rcnn-master文件夹下运行。我的GTX770对于VGG16网络完全不行，ZF网络还可以，但是屏幕闪的很厉害，而且好像用光了所有显存，电脑显示变得卡顿。）速度的确很快：

训练网络

（1）运行fetch_data/fetch_model_ZF.m下载在ImageNet上预训练好的ZF网络，运行fetch_data/fetch_model_VGG16.m下载在ImageNet上预训练好的VGG-16网络。解压后与faster_rcnn-master中的models文件夹合并。 
（2）下载VOC 2007和VOC 2012数据到./datasets中。 
（3）运行experiments/script_faster_rcnn_VOC2007_ZF.m训练ZF网络模型，它会运行如下四步： 

Resources

Note: This documentation may contain links to third party websites, which are provided for your convenience only. Such third party websites are not under Microsoft’s control. Microsoft does not endorse or make any representation, guarantee or assurance regarding any third party website, content, service or product. Third party websites may be subject to the third party’s terms, conditions, and privacy statements.

1. Experiment logs: OneDrive, DropBox, BaiduYun
2. Regions proposals of our trained RPN:
   • ZF net trained on VOC 07 trainval OneDrive, BaiduYun
   • ZF net trained on VOC 07/12 trainval OneDrive, BaiduYun
   • VGG net trained on VOC 07 trainval OneDrive, BaiduYun
   • VGG net trained on VOC 07/12 trainval OneDrive, BaiduYun
   • Note: the proposals are in the format of [left, top, right, bottom, confidence]

If the automatic "fetch_data" fails, you may manually download resouces from:

1. Pre-complied caffe mex:
   • Windows-based mex complied with VS2013 and Cuda6.5: OneDrive, DropBox, BaiduYun
2. ImageNet-pretrained networks:
   • Zeiler & Fergus (ZF) net OneDrive, DropBox, BaiduYun
   • VGG-16 net OneDrive, DropBox, BaiduYun
3. Final RPN+FastRCNN models: OneDrive, DropBox, BaiduYun