/deep-face-recognition

One-shot Learning and deep face recognition notebooks and workshop materials

Primary LanguageJupyter Notebook

Deep Face Recognition using Tensorflow workshop material

(Thursday, August 30, 2018 )

  • One-shot & low-shot learning
  • Siamese network
  • What is metric learning & Face embedding?
  • Face Verification and Identification Challanges: Lfw, Megaface
  • Facenet Triplet Loss, Center Loss, Sphere Face, Amsoftmax, Arcface
  • How align a face with MTCNN Landmarks

Slides:

Slideshare | download pptx

video capture:

part 1 | part 2

Jupyter notebooks:

face Verification with face embedding

face detection and 5 point landmarks with MTCNN

face Identification & Recognition

Face alignment with 5 point landmarks

Important link

facenet pretrained model (Tensorflow)

davidsandberg/facenet/

Sphereface or Angular Softmax (Caffe)

wy1iu/sphereface

ArcFace (MXNet)

deepinsight/insightface

AMSoftmax (Caffe)

happynear/AMSoftmax

Iranian Face Dataset

iran-celeb.ir

Datasets

  1. CASIA WebFace Database. 10,575 subjects and 494,414 images
  2. Labeled Faces in the Wild.13,000 images and 5749 subjects
  3. Large-scale CelebFaces Attributes (CelebA) Dataset 202,599 images and 10,177 subjects. 5 landmark locations, 40 binary attributes.
  4. MSRA-CFW. 202,792 images and 1,583 subjects.
  5. MegaFace Dataset 1 Million Faces for Recognition at Scale 690,572 unique people
  6. FaceScrub. A Dataset With Over 100,000 Face Images of 530 People.
  7. FDDB.Face Detection and Data Set Benchmark. 5k images.
  8. AFLW.Annotated Facial Landmarks in the Wild: A Large-scale, Real-world Database for Facial Landmark Localization. 25k images.
  9. AFW. Annotated Faces in the Wild. ~1k images. 10.3D Mask Attack Dataset. 76500 frames of 17 persons using Kinect RGBD with eye positions (Sebastien Marcel)
  10. Audio-visual database for face and speaker recognition.Mobile Biometry MOBIO http://www.mobioproject.org/
  11. BANCA face and voice database. Univ of Surrey
  12. Binghampton Univ 3D static and dynamic facial expression database. (Lijun Yin, Peter Gerhardstein and teammates)
  13. The BioID Face Database. BioID group
  14. Biwi 3D Audiovisual Corpus of Affective Communication. 1000 high quality, dynamic 3D scans of faces, recorded while pronouncing a set of English sentences.
  15. Cohn-Kanade AU-Coded Expression Database. 500+ expression sequences of 100+ subjects, coded by activated Action Units (Affect Analysis Group, Univ. of Pittsburgh.
  16. CMU/MIT Frontal Faces . Training set: 2,429 faces, 4,548 non-faces; Test set: 472 faces, 23,573 non-faces.
  17. AT&T Database of Faces 400 faces of 40 people (10 images per people)

Other Face Dataset

Face Detection Dataset

FDDB

paper: http://vis-www.cs.umass.edu/fddb/fddb.pdf

dataset: http://vis-www.cs.umass.edu/fddb/index.html#download

Wider Face

extreme scale

paper: https://arxiv.org/pdf/1511.06523.pdf

dataset: http://mmlab.ie.cuhk.edu.hk/projects/WIDERFace/index.html

MAFA

occlusion

paper: http://openaccess.thecvf.com/content_cvpr_2017/papers/Ge_Detecting_Masked_Faces_CVPR_2017_paper.pdf

dataset: http://www.escience.cn/people/geshiming/mafa.html

4k face dataset

hight resolution

paper: https://arxiv.org/pdf/1804.06559.pdf

Unconstrained Face Detection Dataset (UFDD)

different weather

paper: https://arxiv.org/abs/1804.10275

dataset: https://github.com/hezhangsprinter/UFDD

wildest faces

paper: https://arxiv.org/pdf/1805.07566.pdf

Multi-Attribute Labelled Faces (MALF)

paper: http://www.cbsr.ia.ac.cn/faceevaluation/faceevaluation15.pdf

dataset: http://www.cbsr.ia.ac.cn/faceevaluation/#reference

IJB-A Dataset

paper: https://zhaoj9014.github.io/pub/IJBA_1N_report.pdf

dataset: https://www.nist.gov/itl/iad/image-group/ijb-dataset-request-form

Age Estimation Dataset

Adience dataset

dataset: https://talhassner.github.io/home/projects/Adience/Adience-data.html

statistic: Total number of images: 26,580 Total number of subjects: 2,284 Number of age groups: 8 (0-2, 4-6, 8-13, 15-20, 25-32, 38-43, 48-53, 60-) Gender labels: Yes In the wild: Yes Subject labels: Yes

UTK-Face

dataset: https://susanqq.github.io/UTKFace/

APPA-REAL (real and apparent age)

paper: http://openaccess.thecvf.com/content_cvpr_2018_workshops/papers/w48/Clapes_From_Apparent_to_CVPR_2018_paper.pdf

dataset: http://chalearnlap.cvc.uab.es/dataset/26/description/

Face Landmark Detection Dataset

300W

paper: https://ibug.doc.ic.ac.uk/media/uploads/documents/sagonas_iccv_2013_300_w.pdf

COFW

occluded to different degrees

paper: https://www.microsoft.com/en-us/research/wp-content/uploads/2013/12/BurgosArtizzuICCV13rcpr.pdf

AFLW

faces with large head pose up to 120◦ for yaw and 90◦ for pitch and roll.

paper: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.384.2988&rep=rep1&type=pdf

dataset: https://www.tugraz.at/institute/icg/research/team-bischof/lrs/downloads/aflw/

WFLW

from wider face dataset

paper: https://arxiv.org/pdf/1805.10483.pdf

dataset: https://wywu.github.io/projects/LAB/WFLW.html

papers:

Deep Face Recognition (2015)

FaceNet: A Unified Embedding for Face Recognition and Clustering (2015)

A Discriminative Feature Learning Approach for Deep Face Recognition (2016)

SphereFace: Deep Hypersphere Embedding for Face Recognition (2018)

Additive Margin Softmax for Face Verification (2018)

Ring loss: Convex Feature Normalization for Face Recognition(2018)

ArcFace: Additive Angular Margin Loss for Deep Face Recognition (2019)

Deep Face Recognition: A Survey (2019)

source:

https://github.com/jian667/face-dataset

https://github.com/jian667/Face-Resources/