Two-Stream RGB-D Human Detection Algorithm Based on RFB Network

Two-Stream RGB-D Human Detection Algorithm Based on RFB Network

In order to effectively combine RGB image features with depth image features for human detection, this paper proposes a two-stream RGB-D human detection algorithm based on RFB network. The proposed algorithm mainly contains three parts: RGB-stream, Depth-stream and Channel Weight Fusion (CWF) strate...

Full description

Bibliographic Details
Main Authors: Wenli Zhang, Jiaqi Wang, Xiang Guo, Kaizhen Chen, Ning Wang
Format: Article
Language:English
Published: IEEE 2020-01-01
Series:IEEE Access
Subjects:
RGB-D
human detection
fusion features
two-stream
Online Access:https://ieeexplore.ieee.org/document/9134743/
id doaj-789893c8ecd6467fa80b42f2798adb44
record_format Article
spelling doaj-789893c8ecd6467fa80b42f2798adb442021-03-30T02:17:53ZengIEEEIEEE Access2169-35362020-01-01812317512318110.1109/ACCESS.2020.30076119134743Two-Stream RGB-D Human Detection Algorithm Based on RFB NetworkWenli Zhang0https://orcid.org/0000-0003-3151-5755Jiaqi Wang1https://orcid.org/0000-0002-1708-3573Xiang Guo2https://orcid.org/0000-0003-3395-1737Kaizhen Chen3https://orcid.org/0000-0001-6871-4091Ning Wang4https://orcid.org/0000-0002-9863-0275Faculty of Information Technology, Beijing University of Technology, Beijing, ChinaFaculty of Information Technology, Beijing University of Technology, Beijing, ChinaFaculty of Information Technology, Beijing University of Technology, Beijing, ChinaFaculty of Information Technology, Beijing University of Technology, Beijing, ChinaFaculty of Information Technology, Beijing University of Technology, Beijing, ChinaIn order to effectively combine RGB image features with depth image features for human detection, this paper proposes a two-stream RGB-D human detection algorithm based on RFB network. The proposed algorithm mainly contains three parts: RGB-stream, Depth-stream and Channel Weight Fusion (CWF) strategy. (1) The RGB-stream extracts RGB image features using RFB-Net as the backbone network. (2) By analyzing the results of depth features visualization, we build the Depth-stream, which can effectively extract the depth image features. (3) The improved CWF strategy can enhance the effectiveness of important channels in RGB-D fusion features and improve the capability of the network expression. The experimental results show that the proposed algorithm has a significant improvement compared with other algorithms on two common datasets.https://ieeexplore.ieee.org/document/9134743/RGB-Dhuman detectionfusion featurestwo-stream
collection DOAJ
language English
format Article
sources DOAJ
author Wenli Zhang
Jiaqi Wang
Xiang Guo
Kaizhen Chen
Ning Wang
spellingShingle Wenli Zhang
Jiaqi Wang
Xiang Guo
Kaizhen Chen
Ning Wang
Two-Stream RGB-D Human Detection Algorithm Based on RFB Network
IEEE Access
RGB-D
human detection
fusion features
two-stream
author_facet Wenli Zhang
Jiaqi Wang
Xiang Guo
Kaizhen Chen
Ning Wang
author_sort Wenli Zhang
title Two-Stream RGB-D Human Detection Algorithm Based on RFB Network
title_short Two-Stream RGB-D Human Detection Algorithm Based on RFB Network
title_full Two-Stream RGB-D Human Detection Algorithm Based on RFB Network
title_fullStr Two-Stream RGB-D Human Detection Algorithm Based on RFB Network
title_full_unstemmed Two-Stream RGB-D Human Detection Algorithm Based on RFB Network
title_sort two-stream rgb-d human detection algorithm based on rfb network
publisher IEEE
series IEEE Access
issn 2169-3536
publishDate 2020-01-01
description In order to effectively combine RGB image features with depth image features for human detection, this paper proposes a two-stream RGB-D human detection algorithm based on RFB network. The proposed algorithm mainly contains three parts: RGB-stream, Depth-stream and Channel Weight Fusion (CWF) strategy. (1) The RGB-stream extracts RGB image features using RFB-Net as the backbone network. (2) By analyzing the results of depth features visualization, we build the Depth-stream, which can effectively extract the depth image features. (3) The improved CWF strategy can enhance the effectiveness of important channels in RGB-D fusion features and improve the capability of the network expression. The experimental results show that the proposed algorithm has a significant improvement compared with other algorithms on two common datasets.
topic RGB-D
human detection
fusion features
two-stream
url https://ieeexplore.ieee.org/document/9134743/
work_keys_str_mv AT wenlizhang twostreamrgbdhumandetectionalgorithmbasedonrfbnetwork
AT jiaqiwang twostreamrgbdhumandetectionalgorithmbasedonrfbnetwork
AT xiangguo twostreamrgbdhumandetectionalgorithmbasedonrfbnetwork
AT kaizhenchen twostreamrgbdhumandetectionalgorithmbasedonrfbnetwork
AT ningwang twostreamrgbdhumandetectionalgorithmbasedonrfbnetwork
_version_ 1724185421037961216

Similar Items