Feature agglomeration networks for single stage face detection

Recent years have witnessed promising results of exploring deep convolutional neural network for face detection. Despite making remarkable progress, face detection in the wild remains challenging especially when detecting faces at vastly different scales and characteristics. In this paper, we propos...

Full description

Saved in:
Bibliographic Details
Main Authors: ZHANG, Jialiang, WU, Xiongwei, HOI, Steven C. H., ZHU, Jianke
Format: text
Language:English
Published: Institutional Knowledge at Singapore Management University 2020
Subjects:
Online Access:https://ink.library.smu.edu.sg/sis_research/5098
https://ink.library.smu.edu.sg/context/sis_research/article/6101/viewcontent/1712.00721.pdf
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Singapore Management University
Language: English
Description
Summary:Recent years have witnessed promising results of exploring deep convolutional neural network for face detection. Despite making remarkable progress, face detection in the wild remains challenging especially when detecting faces at vastly different scales and characteristics. In this paper, we propose a novel simple yet effective framework of “Feature Agglomeration Networks” (FANet) to build a new single-stage face detector, which not only achieves state-of-the-art performance but also runs efficiently. As inspired by Feature Pyramid Networks (FPN) (Lin et al., 2017), the key idea of our framework is to exploit inherent multi-scale features of a single convolutional neural network by aggregating higher-level semantic feature maps of different scales as contextual cues to augment lower-level feature maps via a hierarchical agglomeration manner at marginal extra computation cost. We further propose a Hierarchical Loss to effectively train the FANet model. We evaluate the proposed FANet detector on several public face detection benchmarks, including PASCAL face, FDDB, and WIDER FACE datasets and achieved state-of-the-art results2. Our detector can run in real-time for VGA-resolution images on GPU.