Face alignment recurrent network

作者：

Highlights：

• This paper presents a new recurrent based facial landmark detection method for images and videos under uncontrolled conditions.

• LSTM model is employed in our network to make full use of the spatial and temporal middle stage information in a natural way.

• The results show clear improvement on public image datasets and video datasets.

摘要

•This paper presents a new recurrent based facial landmark detection method for images and videos under uncontrolled conditions.•LSTM model is employed in our network to make full use of the spatial and temporal middle stage information in a natural way.•The results show clear improvement on public image datasets and video datasets.

论文关键词：Face alignment,Recurrent network

论文评审过程：Received 28 February 2017, Revised 11 September 2017, Accepted 18 September 2017, Available online 20 September 2017, Version of Record 5 October 2017.

论文官网地址：https://doi.org/10.1016/j.patcog.2017.09.028