[reading notes]One Millisecond Face Alignment with an Ensemble of Regression Trees

Kazemi V, Sullivan J. One millisecond face alignment with an ensemble of regression trees[C]// Computer Vision and Pattern Recognition. IEEE, 2014:1867-1874.

https://www.cv-foundation.org/openaccess/content_cvpr_2014/papers/Kazemi_One_Millisecond_Face_2014_CVPR_paper.pdf

1. I recommend this paper for the following reasons

(a). open source, integrated into Dlib, also used in OpenFace.

(b). very fast and robust.

2. algorithm process 

Similar to previous works [8, 2] our proposed method utilizes a cascade of regression functions.

To train each regression function, we use the gradient tree boosting algorithm with a sum of square error loss as described in [10].

This paper is an improved version of previous works [8, 2]. So I decide to read Mircosoft's paper [2] first.

[2] X. Cao, Y. Wei, F. Wen, and J. Sun. Face alignment by explicit shape regression. In CVPR, pages 2887–2894, 2012.

[8] P. Dollar, P. Welinder, and P. Perona. Cascaded pose regression. In CVPR, pages 1078–1085, 2010. 
[10] T. Hastie, R. Tibshirani, and J. H. Friedman. The elements of statistical learning: data mining, inference, and prediction. New York: Springer-Verlag, 2001. (can not download this book. 555! we can refer to this cause for boosting)