caffe_windows使用mnist训练的效果测试

今天根据网上的文章，训练了一下mnist的数据

效果不是很好

配置的过程参考：

http://www.cnblogs.com/yixuan-xu/p/5858595.html 

 http://www.cnblogs.com/yixuan-xu/p/5862657.html

其实最后他是用的是classification.exe进行分类

我修改了一个下classification.cpp里面的代码,代码如下：

int main(int argc, char** argv) {
	/***
  if (argc != 6) {
    std::cerr << "Usage: " << argv[0]
              << " deploy.prototxt network.caffemodel"
              << " mean.binaryproto labels.txt img.jpg" << std::endl;
    return 1;
  }

  ::google::InitGoogleLogging(argv[0]);

  string model_file   = argv[1];
  string trained_file = argv[2];
  string mean_file    = argv[3];
  string label_file   = argv[4];
  Classifier classifier(model_file, trained_file, mean_file, label_file);

  string file = argv[5];
  **/
	if (argc != 2) {
		std::cerr << "Usage: " << argv[0]
			<< " deploy.prototxt network.caffemodel"
			<< " mean.binaryproto labels.txt img.jpg" << std::endl;
		return 1;
	}

	::google::InitGoogleLogging(argv[0]);

	string model_file = "lenet.prototxt";
	string trained_file = "lenet_iter_10000.caffemodel";
	string mean_file = "mean.binaryproto";
	string label_file = "synset_words.txt";
	Classifier classifier(model_file, trained_file, mean_file, label_file);

	string file = argv[1];

	IplImage *image = cvLoadImage(file.c_str());
	IplImage *desc;
	if (!image){
		printf(" No image data 
 ");
		return -1;
	}
	CvSize sz;
	sz.width = 28;
	sz.height = 28;
	desc = cvCreateImage(sz, image->depth, image->nChannels);
	cvResize(image, desc, CV_INTER_AREA);
	cv::Mat imge;
	imge=cv::cvarrToMat(desc);
	cv::Mat gray_image;
	cvtColor(imge, gray_image, CV_BGR2GRAY);
	imwrite("gray_image.bmp", gray_image);
	file = "gray_image.bmp";

  std::cout << "---------- Prediction for "
            << file << " ----------" << std::endl;

  cv::Mat img = cv::imread(file, -1);
  CHECK(!img.empty()) << "Unable to decode image " << file;
  std::vector<Prediction> predictions = classifier.Classify(img);

  /* Print the top N predictions. */
  for (size_t i = 0; i < predictions.size(); ++i) {
    Prediction p = predictions[i];
    std::cout << std::fixed << std::setprecision(4) << p.second << " - ""
              << p.first << """ << std::endl;
  }
}

 最后生成的可以执行的文件包如下

 博客园这货上传文件容量那么少，只能找度娘了

http://pan.baidu.com/s/1gfcHkhP

 直接点击里面的bat就可以查看效果！