Face and facial landmark detection

under construction ......

Face Detection and Facial Landmark Detection for Pose Estimation

1. Face Detection

• Paul Viola’s method: Haar like feature + Integral image + Cascaded Adaboost;

• HoG + SVM + Image Pyramid (dlib C++ lib);

• Variation of LBP and HoG for face detection: Local Gradient Pattern and Binary Histogram of Gradients;

• Integral/Aggregate Channel Features: HOG + LUV;

• DPM (Deformable Part Model) or Pictorial model plus SVM.

• DLib:

  • Max margin SVM;

  • Spatial Bag of Words;

  • HOG filter;

  • image pyramid-based.

2. Facial Landmark Detection

• View/Pose alignment + view/pose specific detectors;

• Facial feature (2d-landmark) detection for alignment: holistic or local

  • ASM, AAM: generative model;

  • Elastic graph matching;

  • Constrained local model (CLM): global shape constraints;

  • Explicit Shape Regression [Cao, Wei, Wen, Sun'12];

  • Robust cascaded pose regression (RCPR) [Burgos-Artizzu, Perona, Dollár'13];

  • Conditional regression forests [Dantone, Fanelli, Gall, Van Gool'12];

  • Tree Structured Part Model (TSPM): [Zhu, Ramandan’12];

  • Ensemble of regression trees (dlib C++ lib);

  • Supervised Descent Method [Intraface, Xiong, De la Torre'13];

  • Parts-based deformable shape model [Yu, Huang, Zhang, Yan, Metaxas'13];

  • Congealing or funneling [Huang, Jain, Learned-Miller'07]:

    • reduce the entropy by transform.

• DLib:

  • Face pose: 68 landmarks;

  • Feature selection: shape invariant;

  • Ensemble of regression tree: learning in cascade;

  • Gradient boosting: regularization.

References

1. D, King, Max-Margin Object Detection, arxiv: 1502.00046, 2015.

2. V Kazemi and J Sullivan, One Millisecond Face Alignment with an Ensemble of Regression Trees, CVPR 2014;

3. N Wang, X Gao, D Tao, X Li, Facial Feature Point Detection: A Comprehensive Survey, Int. J. Computer Vision, Oct. 2014;

4. S Zafeiriou, C Zhang, Z Zhang, A Survey on Face Detection in the wild: past, present and future, CVIU, Mar. 2015.