A Human Pose Estimation Algorithm with Scale Invariant

Due to the existing issues in current human pose estimation algorithms, which struggle with accurately detecting small and large-sized human keypoints as well as having lower precision, this paper proposes a scale invariant convolution neural network to estimate human pose. First, resizing network i...

Full description

Saved in:
Bibliographic Details
Main Authors: SUN Ruiyang, YANG Huixin, ZHAO Lanfei
Format: Article
Language:zho
Published: Harbin University of Science and Technology Publications 2024-08-01
Series:Journal of Harbin University of Science and Technology
Subjects:
Online Access:https://hlgxb.hrbust.edu.cn/#/digest?ArticleID=2349
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849415702594715648
author SUN Ruiyang
YANG Huixin
ZHAO Lanfei
author_facet SUN Ruiyang
YANG Huixin
ZHAO Lanfei
author_sort SUN Ruiyang
collection DOAJ
description Due to the existing issues in current human pose estimation algorithms, which struggle with accurately detecting small and large-sized human keypoints as well as having lower precision, this paper proposes a scale invariant convolution neural network to estimate human pose. First, resizing network is constructed for resizing the input image to the standard resolution. This network would reduce feature loss caused by interpolation. Second, the receptive field of network is increased by introducing non-local convolution. Thirdly, resolution attention mechanism is introduced into multi-resolution feature fusion, leading to enhance invariance in scales. Finally, optimized network is designed to reduce quantisation error caused by sampling. Experimental results conducted on the COCO dataset indicate that the proposed algorithm achieves an average accuracy of 79. 2% , which is higher than other algorithms. Therefore, the proposed algorithm exhibits better scale invariance and accuracy than existing human pose estimation algorithms.
format Article
id doaj-art-927914eb4bc24b50b59a7b386d6ffda7
institution Kabale University
issn 1007-2683
language zho
publishDate 2024-08-01
publisher Harbin University of Science and Technology Publications
record_format Article
series Journal of Harbin University of Science and Technology
spelling doaj-art-927914eb4bc24b50b59a7b386d6ffda72025-08-20T03:33:26ZzhoHarbin University of Science and Technology PublicationsJournal of Harbin University of Science and Technology1007-26832024-08-012904596810.15938/j.jhust.2024.04.007A Human Pose Estimation Algorithm with Scale InvariantSUN Ruiyang0YANG Huixin1ZHAO Lanfei2College of National Traditional Sports, Harbin Sport University, Harbin 150008 , ChinaCollege of National Traditional Sports, Harbin Sport University, Harbin 150008 , ChinaThe Higher Educational Key Laboratory for Measuring and Control Technology and Instrumentations of Heilongjiang Province, Harbin University of Science and Technology, Harbin 150080 , ChinaDue to the existing issues in current human pose estimation algorithms, which struggle with accurately detecting small and large-sized human keypoints as well as having lower precision, this paper proposes a scale invariant convolution neural network to estimate human pose. First, resizing network is constructed for resizing the input image to the standard resolution. This network would reduce feature loss caused by interpolation. Second, the receptive field of network is increased by introducing non-local convolution. Thirdly, resolution attention mechanism is introduced into multi-resolution feature fusion, leading to enhance invariance in scales. Finally, optimized network is designed to reduce quantisation error caused by sampling. Experimental results conducted on the COCO dataset indicate that the proposed algorithm achieves an average accuracy of 79. 2% , which is higher than other algorithms. Therefore, the proposed algorithm exhibits better scale invariance and accuracy than existing human pose estimation algorithms. https://hlgxb.hrbust.edu.cn/#/digest?ArticleID=2349human pose estimationconvolutional neural networkscale invariancehuman keypoints detectionnon-local convolutionquantisation error
spellingShingle SUN Ruiyang
YANG Huixin
ZHAO Lanfei
A Human Pose Estimation Algorithm with Scale Invariant
Journal of Harbin University of Science and Technology
human pose estimation
convolutional neural network
scale invariance
human keypoints detection
non-local convolution
quantisation error
title A Human Pose Estimation Algorithm with Scale Invariant
title_full A Human Pose Estimation Algorithm with Scale Invariant
title_fullStr A Human Pose Estimation Algorithm with Scale Invariant
title_full_unstemmed A Human Pose Estimation Algorithm with Scale Invariant
title_short A Human Pose Estimation Algorithm with Scale Invariant
title_sort human pose estimation algorithm with scale invariant
topic human pose estimation
convolutional neural network
scale invariance
human keypoints detection
non-local convolution
quantisation error
url https://hlgxb.hrbust.edu.cn/#/digest?ArticleID=2349
work_keys_str_mv AT sunruiyang ahumanposeestimationalgorithmwithscaleinvariant
AT yanghuixin ahumanposeestimationalgorithmwithscaleinvariant
AT zhaolanfei ahumanposeestimationalgorithmwithscaleinvariant
AT sunruiyang humanposeestimationalgorithmwithscaleinvariant
AT yanghuixin humanposeestimationalgorithmwithscaleinvariant
AT zhaolanfei humanposeestimationalgorithmwithscaleinvariant