playif / caffe-face

This branch is developed for deep face recognition

Deep Face Recognition with Caffe Implementation

This branch is developed for deep face recognition, the related paper is as follows.

A Discriminative Feature Learning Approach for Deep Face Recognition[C]
Yandong Wen, Kaipeng Zhang, Zhifeng Li*, Yu Qiao
European Conference on Computer Vision. Springer International Publishing, 2016: 499-515.

• Updates
• Files
• Train_Model
• Extract_DeepFeature
• Contact
• Citation
• LICENSE
• README_Caffe

Updates

• Oct 13, 2016
  • A demo for extracting deep feature by the given model is provided.
• Oct 12, 2016
  • The links of face model and features on LFW are available.
    model: google drive baidu skydrive
    feature: google drive baidu skydrive
  • The training prototxt of toy example on MNIST are released.
• Otc 9, 2016
  • The code and training prototxt for our ECCV16 paper are released.
  • If you train our Network on CAISA-WebFace, the expected verification performance of SINGLE MODEL on LFW should be ~99%.

Files

• Original Caffe library
• Center Loss
  • src/caffe/proto/caffe.proto
  • include/caffe/layers/center_loss_layer.hpp
  • src/caffe/layers/center_loss_layer.cpp
  • src/caffe/layers/center_loss_layer.cu
• face_example
  • face_example/data/
  • face_example/face_snapshot/
  • face_example/face_train_test.prototxt
  • face_example/face_solver.prototxt
  • face_example/face_deploy.prototxt
  • face_example/extractDeepFeature.m
• mnist_example
  • mnist_example/data/
  • mnist_example/face_snapshot/
  • mnist_example/mnist_train_test.prototxt
  • mnist_example/mnist_solver.prototxt
  • mnist_example/mnist_deploy.prototxt

Train_Model

1. The Installation completely the same as Caffe. Please follow the installation instructions. Make sure you have correctly installed before using our code.

2. Download the face dataset for training, e.g. CAISA-WebFace, VGG-Face, MS-Celeb-1M, MegaFace.

3. Preprocess the training face images, including detection, alignment, etc. Here we strongly recommend MTCNN, which is an effective and efficient open-source tool for face detection and alignment.

4. Creat list for training set and validation set. Place them in face_example/data/

5. Specify your data source for train & val

    layer {
      name: "data"
      type: "ImageData"
      top: "data"
      top: "label"
      image_data_param {
        source: "face_example/data/###your_list###"
      }
    }
   

6. Specify the number of subject in FC6 layer

    layer {
      name: "fc6"
      type: "InnerProduct"
      bottom: "fc5"
      top: "fc6"
      inner_product_param {
        num_output: ##number##
      }
    }
   

7. Specify the loss weight and the number of subject in center loss layer

    layer {
      name: "center_loss"
      type: "CenterLoss"
      bottom: "fc5"
      bottom: "label"
      top: "center_loss"
      loss_weight: ##weight##
      center_loss_param {
        num_output: ##number##
      }
    }
   

8. Train model

    cd $CAFFE-FACE_ROOT
    ./build/tools/caffe train -solver face_example/face_solver.prototxt -gpu X,Y
   

Extract_DeepFeature

1. Compile matcaffe by make matcaffe

2. Specify the correspinding paths in face_example/extractDeepFeature.m

    addpath('path_to_matCaffe/matlab');
    model = 'path_to_deploy/face_deploy.prototxt';
    weights = 'path_to_model/face_model.caffemodel';
    image = imread('path_to_image/Jennifer_Aniston_0016.jpg');
   

3. Run extractDeepFeature.m in Matlab

Contact

• Yandong Wen
• Kaipeng Zhang

Citation

You are encouraged to cite the following paper if it helps your research.

@inproceedings{wen2016discriminative,
  title={A Discriminative Feature Learning Approach for Deep Face Recognition},
  author={Wen, Yandong and Zhang, Kaipeng and Li, Zhifeng and Qiao, Yu},
  booktitle={European Conference on Computer Vision},
  pages={499--515},
  year={2016},
  organization={Springer}
}

License

Copyright (c) Yandong Wen

All rights reserved.

MIT License

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED AS IS, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

---

README_Caffe

Caffe

Caffe is a deep learning framework made with expression, speed, and modularity in mind. It is developed by the Berkeley Vision and Learning Center (BVLC) and community contributors.

Check out the project site for all the details like

• DIY Deep Learning for Vision with Caffe
• Tutorial Documentation
• BVLC reference models and the community model zoo
• Installation instructions

and step-by-step examples.

Please join the caffe-users group or gitter chat to ask questions and talk about methods and models. Framework development discussions and thorough bug reports are collected on Issues.

Happy brewing!

License and Citation

Caffe is released under the BSD 2-Clause license. The BVLC reference models are released for unrestricted use.

Please cite Caffe in your publications if it helps your research:

@article{jia2014caffe,
  Author = {Jia, Yangqing and Shelhamer, Evan and Donahue, Jeff and Karayev, Sergey and Long, Jonathan and Girshick, Ross and Guadarrama, Sergio and Darrell, Trevor},
  Journal = {arXiv preprint arXiv:1408.5093},
  Title = {Caffe: Convolutional Architecture for Fast Feature Embedding},
  Year = {2014}
}