End-To-end cascade cnn for simultaneously face detection and alignment

  • Sanyuan Zhao*
  • , Hongmei Song
  • , Weilin Cong
  • , Qi Qi
  • , Hui Tian
  • *Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Citation (Scopus)

Abstract

Real-world face detection and alignment demand an advanced discriminative model to address challenges by pose, lighting and expression. Recent studies have utilized the relation between face detection and alignment to make models computationally efficiency, but they ignore the connection between each cascade CNNs. In this paper, we combine detection, calibration and alignment in each cascade structure and propose an End-To-End cascade Online Hard Example Mining (OHEM) for training, which expert in accelerating convergence. Experiments on FDDB and AFLW demonstrate considerable improvement on accuracy and speed.

Original languageEnglish
Title of host publicationProceedings - 2017 International Conference on Virtual Reality and Visualization, ICVRV 2017
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages35-40
Number of pages6
ISBN (Electronic)9781538626368
DOIs
Publication statusPublished - 2 Jul 2017
Event7th International Conference on Virtual Reality and Visualization, ICVRV 2017 - Zhengzhou, China
Duration: 21 Oct 201722 Oct 2017

Publication series

NameProceedings - 2017 International Conference on Virtual Reality and Visualization, ICVRV 2017

Conference

Conference7th International Conference on Virtual Reality and Visualization, ICVRV 2017
Country/TerritoryChina
CityZhengzhou
Period21/10/1722/10/17

Keywords

  • Cascade CNNs
  • End-To-End
  • Face detection
  • Facial key points detection

Fingerprint

Dive into the research topics of 'End-To-end cascade cnn for simultaneously face detection and alignment'. Together they form a unique fingerprint.

Cite this