Face Detection and Segmentation Based on Improved Mask R-CNN

Deep convolutional neural networks have been successfully applied to face detection recently. Despite making remarkable progress, most of the existing detection methods only localize each face using a bounding box, which cannot segment each face from the background image simultaneously. To overcome...

Full description

Saved in:
Bibliographic Details
Main Authors: Kaihan Lin, Huimin Zhao, Jujian Lv, Canyao Li, Xiaoyong Liu, Rongjun Chen, Ruoyan Zhao
Format: Article
Language:English
Published: Wiley 2020-01-01
Series:Discrete Dynamics in Nature and Society
Online Access:http://dx.doi.org/10.1155/2020/9242917
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850177684130234368
author Kaihan Lin
Huimin Zhao
Jujian Lv
Canyao Li
Xiaoyong Liu
Rongjun Chen
Ruoyan Zhao
author_facet Kaihan Lin
Huimin Zhao
Jujian Lv
Canyao Li
Xiaoyong Liu
Rongjun Chen
Ruoyan Zhao
author_sort Kaihan Lin
collection DOAJ
description Deep convolutional neural networks have been successfully applied to face detection recently. Despite making remarkable progress, most of the existing detection methods only localize each face using a bounding box, which cannot segment each face from the background image simultaneously. To overcome this drawback, we present a face detection and segmentation method based on improved Mask R-CNN, named G-Mask, which incorporates face detection and segmentation into one framework aiming to obtain more fine-grained information of face. Specifically, in this proposed method, ResNet-101 is utilized to extract features, RPN is used to generate RoIs, and RoIAlign faithfully preserves the exact spatial locations to generate binary mask through Fully Convolution Network (FCN). Furthermore, Generalized Intersection over Union (GIoU) is used as the bounding box loss function to improve the detection accuracy. Compared with Faster R-CNN, Mask R-CNN, and Multitask Cascade CNN, the proposed G-Mask method has achieved promising results on FDDB, AFW, and WIDER FACE benchmarks.
format Article
id doaj-art-c39896de9ddc44fc97df1f3fbb43e543
institution OA Journals
issn 1026-0226
1607-887X
language English
publishDate 2020-01-01
publisher Wiley
record_format Article
series Discrete Dynamics in Nature and Society
spelling doaj-art-c39896de9ddc44fc97df1f3fbb43e5432025-08-20T02:18:55ZengWileyDiscrete Dynamics in Nature and Society1026-02261607-887X2020-01-01202010.1155/2020/92429179242917Face Detection and Segmentation Based on Improved Mask R-CNNKaihan Lin0Huimin Zhao1Jujian Lv2Canyao Li3Xiaoyong Liu4Rongjun Chen5Ruoyan Zhao6School of Computer Science, Guangdong Polytechnic Normal University, Guangzhou 510665, ChinaSchool of Computer Science, Guangdong Polytechnic Normal University, Guangzhou 510665, ChinaSchool of Computer Science, Guangdong Polytechnic Normal University, Guangzhou 510665, ChinaSchool of Computer Science, Guangdong Polytechnic Normal University, Guangzhou 510665, ChinaSchool of Computer Science, Guangdong Polytechnic Normal University, Guangzhou 510665, ChinaSchool of Computer Science, Guangdong Polytechnic Normal University, Guangzhou 510665, ChinaSchool of Computer Science, Guangdong Polytechnic Normal University, Guangzhou 510665, ChinaDeep convolutional neural networks have been successfully applied to face detection recently. Despite making remarkable progress, most of the existing detection methods only localize each face using a bounding box, which cannot segment each face from the background image simultaneously. To overcome this drawback, we present a face detection and segmentation method based on improved Mask R-CNN, named G-Mask, which incorporates face detection and segmentation into one framework aiming to obtain more fine-grained information of face. Specifically, in this proposed method, ResNet-101 is utilized to extract features, RPN is used to generate RoIs, and RoIAlign faithfully preserves the exact spatial locations to generate binary mask through Fully Convolution Network (FCN). Furthermore, Generalized Intersection over Union (GIoU) is used as the bounding box loss function to improve the detection accuracy. Compared with Faster R-CNN, Mask R-CNN, and Multitask Cascade CNN, the proposed G-Mask method has achieved promising results on FDDB, AFW, and WIDER FACE benchmarks.http://dx.doi.org/10.1155/2020/9242917
spellingShingle Kaihan Lin
Huimin Zhao
Jujian Lv
Canyao Li
Xiaoyong Liu
Rongjun Chen
Ruoyan Zhao
Face Detection and Segmentation Based on Improved Mask R-CNN
Discrete Dynamics in Nature and Society
title Face Detection and Segmentation Based on Improved Mask R-CNN
title_full Face Detection and Segmentation Based on Improved Mask R-CNN
title_fullStr Face Detection and Segmentation Based on Improved Mask R-CNN
title_full_unstemmed Face Detection and Segmentation Based on Improved Mask R-CNN
title_short Face Detection and Segmentation Based on Improved Mask R-CNN
title_sort face detection and segmentation based on improved mask r cnn
url http://dx.doi.org/10.1155/2020/9242917
work_keys_str_mv AT kaihanlin facedetectionandsegmentationbasedonimprovedmaskrcnn
AT huiminzhao facedetectionandsegmentationbasedonimprovedmaskrcnn
AT jujianlv facedetectionandsegmentationbasedonimprovedmaskrcnn
AT canyaoli facedetectionandsegmentationbasedonimprovedmaskrcnn
AT xiaoyongliu facedetectionandsegmentationbasedonimprovedmaskrcnn
AT rongjunchen facedetectionandsegmentationbasedonimprovedmaskrcnn
AT ruoyanzhao facedetectionandsegmentationbasedonimprovedmaskrcnn