Supervised descent method with low rank and sparsity constraints for robust face alignment

2015 ◽  
Author(s):  
Yubao Sun ◽  
Bin Hu ◽  
Jiankang Deng ◽  
Xing Li
Keyword(s):  
Descent Method ◽  
Low Rank ◽  
Face Alignment ◽  
Supervised Descent Method ◽  
Sparsity Constraints

Multi-template supervised descent method for face alignment

2021 ◽  
Vol 65 ◽  
pp. 107-117
Author(s):  
Cheng Ding ◽  
Weidong Tian ◽  
Chao Geng ◽  
Xijing Zhu ◽  
Qinmu Peng ◽  
...  
Keyword(s):  
Descent Method ◽  
Face Alignment ◽  
Supervised Descent Method

scholarly journals Tightly-Coupled Data Compression for Efficient Face Alignment

2018 ◽  
Vol 8 (11) ◽  
pp. 2284
Author(s):  
Yehu Shen ◽  
Quansheng Jiang ◽  
Bangfu Wang ◽  
Qixin Zhu ◽  
Wenming Yang
Keyword(s):  
Data Compression ◽  
Parametric Method ◽  
Descent Method ◽  
Face Alignment ◽  
Expression Recognition ◽  
Model Data ◽  
Supervised Descent Method ◽  
Tightly Coupled ◽  
Original Size ◽  
Cascaded Regression

Face alignment is the key component for applications such as face and expression recognition, face based AR (Augmented Reality), etc. Among all the algorithms, cascaded-regression based methods have become popular in recent years for their low computational costs and satisfactory performances in uncontrolled environments. However, the size of the trained model is large for cascaded-regression based methods, which makes it difficult to be applied in resource restricted scenarios such as applications on mobile phones. In this paper, a data compression method for the trained model of supervised descent method (SDM) is proposed. Firstly, according to the distribution of the model data estimated with the non-parametric method, a K-means based data quantization algorithm with probability density-aware initialization was proposed to efficiently quantize the model data. Then, a tightly-coupled SDM training algorithm was proposed so that the training process reduced the errors caused by data quantization. Quantitative experimental results proved that our proposed method compressed the trained model to less than 19% of its original size with very similar feature localization performance. The proposed method opens the gates to efficient mobile face alignment applications based on SDM.

Sign in / Sign up

Export Citation Format

Share Document