Deep Face Recognition with Caffe Implementation

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

OCCLUSION ROBUST FACE RECOGNITION BASED ON MASK LEARNING[C]
Weitao, Wan and Jiansheng, Chen
2017 IEEE International Conference on Image Processing (ICIP)

The paper can be downloaded from here.

• Architecture
• Results
• Files
• Train_Model
• Contact
• Citation
• LICENSE
• README_Caffe

Architecture

Our network architecture is

Results

The generated masks on faces with occlusion are

Face verification on lfw validation set with synthesized square blocks with varying sizes.

Files

• Caffe with center loss imported from https://github.com/ydwen/caffe-face
• Mask Layer
  • src/caffe/layers/mask_layer.cu
  • src/caffe/layers/mask_layer.cpp (CPU mode not supported for now)
  • include/caffe/layers/mask_layer.hpp
• training network
  • face_example/train_centerMask2Pool2_ori.prototxt
  • face_example/face_solver.prototxt

Train_Model

1. Specify your mask size in 'num_output' (the value should equal to height x width of the mask)

    layer {
      name: "mask_ip"
      type: "InnerProduct"
      bottom: "mask_conv3"
      top: "mask_ip"
      param {
        name: "mask_ip_w"
        lr_mult: 1
      }
      param {
        name: "mask_ip_b"
        lr_mult: 2
      }
      inner_product_param {
        num_output: 572     # change it based on your network
        weight_filler {
          type: "constant"
        }
        bias_filler {
          type: "constant"
        }
      }
    }
   

2. Choose the location to insert the mask layer. I placed it after 'pool2'.

    layer {
      name: "mask"
      type: "Mask"
      bottom: "pool2"
      bottom: "mask_2d"
      top: "masked_pool2"
      mask_param {
        scale: 1
      }
    }
   

Contact

Weitao Wan([email protected])

Citation

Please consider citing the following paper if it helps your research.

@inproceedings{wan2017mask,
  title={OCCLUSION ROBUST FACE RECOGNITION BASED ON MASK LEARNING},
  author={Weitao, Wan and Jiansheng, Chen},
  booktitle={IEEE International Conference on Image Processing (ICIP)},
  pages={},
  year={2017},
  organization={IEEE}
}

License

Copyright (c) Weitao Wan

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}
}