An Axial Compression Transformer for Efficient Human Pose Estimation

Tan, Wen; Zhang, Haixiang; Song, Xinyi

doi:10.3390/app15094746

This is an early access version, the complete PDF, HTML, and XML versions will be available soon.

Open AccessArticle

An Axial Compression Transformer for Efficient Human Pose Estimation

by

Wen Tan

,

Haixiang Zhang

^* and

Xinyi Song

School of Computer Science and Technology, Zhejiang Sci-Tech University, Hangzhou 310018, China

^*

Author to whom correspondence should be addressed.

Appl. Sci. 2025, 15(9), 4746; https://doi.org/10.3390/app15094746

Submission received: 22 March 2025 / Revised: 16 April 2025 / Accepted: 23 April 2025 / Published: 24 April 2025

(This article belongs to the Collection Advances in Image Processing, Analysis and Recognition Technology)

Download Versions Notes

Abstract

Transformer has a wide range of applications in human posture estimation. It can model the global dependence relationship of images through the self-attention mechanism to obtain key human body information. However, Transformer consumes a lot of computation. An axial compression pose transformer (ACPose) method is proposed to reduce part of the computational cost of Transformer by the axial compression of the input matrix, while maintaining the global receptive field by feature fusion. A Local Enhancement Module is constructed to avoid the loss of too much feature information in the compression process. In the COCO dataset experiment, there was a significant reduction in computational cost compared to those of state-of-the-art transformer-based algorithms.

Keywords: human pose estimation; transformer; axial compression

Share and Cite

MDPI and ACS Style

Tan, W.; Zhang, H.; Song, X. An Axial Compression Transformer for Efficient Human Pose Estimation. Appl. Sci. 2025, 15, 4746. https://doi.org/10.3390/app15094746

AMA Style

Tan W, Zhang H, Song X. An Axial Compression Transformer for Efficient Human Pose Estimation. Applied Sciences. 2025; 15(9):4746. https://doi.org/10.3390/app15094746

Chicago/Turabian Style

Tan, Wen, Haixiang Zhang, and Xinyi Song. 2025. "An Axial Compression Transformer for Efficient Human Pose Estimation" Applied Sciences 15, no. 9: 4746. https://doi.org/10.3390/app15094746

APA Style

Tan, W., Zhang, H., & Song, X. (2025). An Axial Compression Transformer for Efficient Human Pose Estimation. Applied Sciences, 15(9), 4746. https://doi.org/10.3390/app15094746

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Menu

An Axial Compression Transformer for Efficient Human Pose Estimation

Abstract

Share and Cite

Article Metrics

Article Access Statistics

Further Information

Guidelines

MDPI Initiatives

Follow MDPI