Skip to content

Latest commit

 

History

History
18 lines (15 loc) · 931 Bytes

README.md

File metadata and controls

18 lines (15 loc) · 931 Bytes

vgg_face_matconvnet

VGG Face descriptor source code and models (MatConvNet)

This is my starting point for learning NN & DL.I would like to share learning experiences with the great Github community. I make use of Pretrained Model to accomplish Face Verification task.The result is the cosine similarity of two faces.

Step by step

1.download and install "MatConvNet: CNNs for MATLAB",see http://www.vlfeat.org/matconvnet/. I use version 1.0-beta17. 2.download this vgg_face_matconvnet and run it in MATLAB. I use MATLAB 2014b. Pretrained model vgg_face.mat (1GB) is not contained in my repository.Just use the one in "http://www.robots.ox.ac.uk/~vgg/software/vgg_face/". In MATLAB:

convNet = lib.face_feats.convNet('data\vgg_face.mat'); demo('1.jpg','2.jpg',convNet)

Acknowdgements

Thanks to "http://www.robots.ox.ac.uk/~vgg/software/vgg_face/",with which I have had the first intimate touch with NN & DL.