How to test the model use LFW?

[taoyunuo]

Please provide specific test methods，thanks！

[Ldpe2G]

as far as i know, lfw is used for face verification tasks. If so, you just need the pcanet to extract the face feature.

steps:
1, prepare the training images for the pcanet algorithm to get the convolution filters. For this part, you can refer to the main.cpp from line 49 to 90.

2, after getting the filters, you can extract the feature of a test image. For this part, you can refer to main.cpp from line 167 to 178.

[taoyunuo]

Thanks very much!I will try to do it. If I have another problem, I might trouble you for help. Thanks again
@Ldpe2G

[Ldpe2G]

@taoyunuo your welcome :)

[taoyunuo]

I have two questions,

1. first question:
   I have used your code for the test of LFW, but I want to know how to directly use the trained models( filters.xml and svm.xml)to test, do not need to retrain every test. If it is convenient for you, please provide the test code that the model is used directly for testing.
2. second question,:
   Can you use the training picture size of 250*250?
   Does picture size affect the effect of the model?

Look forward to your reply！@Ldpe2G