Example of Caffe window interface

Caffe's own example of a new project, mainly the configuration include Lib DLLs are pits, but also divided into debug and release two versions.

and add input items to be aware of, but also need to be compiled caffe.lib and so on a series of things to copy under the current project.

Caffe's other pit is: F0519 14:54:12.494139 14504 layer_factory.hpp:77] Check failed:registry.count (t ype) = = 1 (0 vs. 1) Unknown Layer Type:input (known types:input) was originally to add header files! Http://blog.csdn.net/fangjin_kl/article/details/50936952#0-tsina-1-63793-397232819ff9a47a7b7e80a40613cfe1

Then it can be compiled through Caffe, the next pit:

String model_file = "D:\\caffe\\caffe-master\\mypower\\deploy.prototxt";//prototxt this must be depoly.

String trained_file = "D:\\caffe\\caffe-master\\mypower_iter_2000.caffemodel"; This is a well-trained model.
String mean_file = "D:\\caffe\\caffe-master\\mypower\\imagenet_mean.binaryproto";
String label_file = "D:\\caffe\\caffe-master\\mypower\\synset_words.txt"; This is a sample label.
Classifier Classifier (Model_file, Trained_file, Mean_file, label_file);

It was found that the small sample data was easy to fit through experiments.

Next two steps: The first step is to put the training data in the test data is a bit false

The second step is to use the data augmentation method (GO) to expand the Keras based on the

Example of Caffe window interface