How to extract feature of an input image - Matlab wrapper?

[Cáp Hữu Quân]

Hi everyone,

I'm new user of Caffe.

I have download bvlc_reference_caffenet model and try to using it. I have an input image like a cat and I wanna extract feature of that image in a specific layer in caffenet model (fc6 or fc7).

I use

res = net.forward({im_data}); % run forward 

feature = res{1}; % get feature

But it return 1000-dims vector of softmax score.

Sorry about my bad English and thank you very much.

[Ahmed Ibrahim]

I guess you are using matlab (you should have mentioned that) , if you would like to extract certain layers output you should do something else, I wrote  a simple tutorial about that here

https://ahmedibrahimvt.wordpress.com/2016/04/18/caffe-in-matlab-matcaffe-simple-tutorial/

[Jan]

To make it a quick answer:

weights_FC6 = net.params(‘fc6’,1).get_data();
output_FC6 = net.blobs(‘fc6’).get_data();

(quoted from Ahmed's linked tutorial)

the return value of net.forward, which you called res, only contains the output blobs of the network (the topmost), not all intermediary blobs.

Jan

[Cáp Hữu Quân]

Hi @Ahmed Ibrahim

Thank you very much for your answer. It's help me to understand and fix my problems. :-)

Vào 20:33:02 UTC+7 Thứ Hai, ngày 18 tháng 4 năm 2016, Ahmed Ibrahim đã viết:

[Cáp Hữu Quân]

Hi @Jan

Thank you very much too ^_^

Vào 21:15:30 UTC+7 Thứ Hai, ngày 18 tháng 4 năm 2016, Jan đã viết: