SANet: Smoothed Attention Network for Single Stage Face Detector

Lei Shi, Xiang Xu, Ioannis A. Kakadiaris

Research output: Chapter in Book/Report/Conference proceedingConference contribution

5 Scopus citations


Recently, significant effort has been devoted to exploring the role of feature fusion and enriching contextual information on detecting multi-scale faces. However, simply integrating features of different levels could lead to introducing significant noise. Moreover, recently proposed approaches of enriching contextual information are not efficient or ignore the gridding artifacts produced by dilated convolution. To tackle these issues, we developed a smoothed attention network (dubbed SANet), which introduces an Attention-guided Feature Fusion Module (AFFM) and a Smoothed Context Enhancement Module (SCEM). In particular, the AFFM applies an attention module to high-level semantic features and fuses attention-focused features with low-level semantic features to reduce the noise of the fused feature map. The SCEM stacks dilated convolution and convolution layers alternately to re-learn the relationship among completely separate sets of units produced by dilated convolution to maintain consistency of local information. The SANet achieves promising results on the WIDER FACE validation and testing datasets and is state-of-the-art on the UFDD dataset.

Original languageEnglish (US)
Title of host publication2019 International Conference on Biometrics, ICB 2019
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781728136400
StatePublished - Jun 2019
Event2019 International Conference on Biometrics, ICB 2019 - Crete, Greece
Duration: Jun 4 2019Jun 7 2019

Publication series

Name2019 International Conference on Biometrics, ICB 2019


Conference2019 International Conference on Biometrics, ICB 2019

ASJC Scopus subject areas

  • Computer Science Applications
  • Computer Vision and Pattern Recognition
  • Signal Processing
  • Statistics, Probability and Uncertainty
  • Demography


Dive into the research topics of 'SANet: Smoothed Attention Network for Single Stage Face Detector'. Together they form a unique fingerprint.

Cite this