MobileFaceNet_Tutorial_Pytorch

• This repo illustrates how to implement MobileFaceNet and Arcface for face recognition task.
• Pretrained model is posted for tests over picture, video and cam
• Help document on how to implement MTCNN+MobileFaceNet is available
• Scripts on transforming MXNET data records in Insightface to images are provided
• Scripts on train and evaluation of MobileFaceNet model are provided

MobileFaceNet Video Demo

Test over Picture, Video and Cam

1. Test Picture

python MTCNN_MobileFaceNet.py -img {image_path}

2. Take Picture for Face Database

• over cam

  python take_picture.py -n {name}
  

• over photo

  python take_ID.py -i {image_path} -n {name}
  

3. Test Video

• over cam

  python cam_demo.py
  

• over video file

  python video_demo.py
  

4. Instruction

MobileFaceNet_Step_by_Step.ipynb

Train

Download training and evaluation data from Model Zoo. All training data has been cropped, aligned and resized as 112 x 112. Put images and annotation files into "data_set" folder. The structure should be arranged as follows:

data_set/
            ---> AgeDB-30
            ---> CASIA_Webface_Image
            ---> CFP-FP
            ---> faces_emore_images
            ---> LFW

1. The following script is provided to convert .bin and .rec file to images:

python data_set/load_images_from_bin.py

2. Generate the corresponding annotation files

python data_set/anno_generation.py

3. Train MobileFaceNet

python Train.py

4. Instruction

MobileFaceNet_Training_Step_by_Step.ipynb

The training results over faces_emore data (5822653 images / 85742 ids) are shown below:

Evaluation

python Evaluation.py

Here is the evaluation result. 'Flip' the image could be applied to encode the embedding feature vector with ~ 0.07% higer accuracy. L2 distance score slightly outperforms cos similarity (not necessarily the same trend for other cases, but it is what we conclude in this work)

Eval Type	Score	LFW	AgeDB-30	CFP-FP
Flip	L2	99.52	96.30	92.93
Flip	Cos	99.50	96.18	92.84
UnFlip	L2	99.45	95.63	93.10
UnFlip	Cos	99.45	95.65	93.10

Don't forget to star the repo if it is helpful for your research

Reference

• https://github.com/deepinsight/insightface
• https://github.com/wujiyang/Face_Pytorch
• https://github.com/TreB1eN/InsightFace_Pytorch