GoDP: Globally Optimized Dual Pathway deep network architecture for facial landmark localization in-the-wild

作者：

Highlights：

• A dual-pathway deep network architecture is proposed for facial landmark estimation.

• The architecture relies on proposal-refinement structure and new loss functions.

• Significant improvement over current state-of-the-art algorithms is demonstrated.

摘要

•A dual-pathway deep network architecture is proposed for facial landmark estimation.•The architecture relies on proposal-refinement structure and new loss functions.•Significant improvement over current state-of-the-art algorithms is demonstrated.

论文关键词：Deep learning,Facial landmark localization,Face alignment,Face recognition

论文评审过程：Received 22 February 2017, Revised 21 October 2017, Accepted 8 December 2017, Available online 14 December 2017, Version of Record 11 April 2018.

论文官网地址：https://doi.org/10.1016/j.imavis.2017.12.002