Introduction - If you have any usage issues, please Google them yourself
One source is to identify the program, the other is the matrix class library. Has saved the trained network weights, so the first use of identification procedures can be run directly Release directory of the executable file, then the picture directory test picture read, recognition. Of course you can train on their own network using the training sample, but to pay attention to the choice of training samples, or it may identify very low. Training sample selection principle is that as much as possible representative, training time will not be too long in the case as much as possible the number of training samples