Journal of Computer Applications ›› 2026, Vol. 46 ›› Issue (9): 2787-2792.DOI: 10.11772/j.issn.1001-9081.2025081009

• Artificial intelligence • Previous Articles    

Lightweight human pose estimation framework based on wavelet attention mechanism with enhanced low frequency

Zhangjian JI(), Siyuan WANG   

  1. School of Computer and Information Technology,Shanxi University,Taiyuan Shanxi 030006,China
  • Received:2025-09-03 Revised:2025-11-14 Accepted:2025-11-18 Online:2025-12-01 Published:2026-09-10
  • Contact: Zhangjian JI
  • About author:JI Zhangjian, born in 1983, Ph. D., associate professor. His research interests include computer vision, machine learning.
    WANG Siyuan, born in 1996, M. S. candidate. His research interests include computer vision, human pose estimation.
  • Supported by:
    Fundamental Research Program of Shanxi Province(202203021221002)

基于增强低频小波注意力机制的轻量化人体姿态估计框架

姬张建(), 王思源   

  1. 山西大学 计算机与信息技术学院,太原 030006
  • 通讯作者: 姬张建
  • 作者简介:姬张建(1983—),男,陕西澄城人,副教授,博士,CCF会员,主要研究方向:计算机视觉、机器学习
    王思源(1996—),男,山西临汾人,硕士研究生,主要研究方向:计算机视觉、人体姿态估计。
  • 基金资助:
    山西省基础研究计划项目(202203021221002)

Abstract:

Human pose estimation is a critical task in computer vision and High-Resolution Networks (HRNets) are widely used due to their outstanding performance. However, improvements in model accuracy are often accompanied by a surge in parameter count and computational cost. To address this issue, an innovative lightweight improved module, Frequency Domain Cross Attention Convolution (FCAC), was proposed. First, the basic convolutional module was reconstructed by fusing Wavelet Transform Convolutions (WTConvs) with Depthwise Separable Convolutions (DSCs) with WTConv's enhancing receptive field capability compensating for DSC's lack of contextual connections, thereby optimizing HRNet's performance while maintaining low complexity. Second, an attention mechanism was employed to guide low-frequency components to learn high-frequency components, and the learning results were aggregated dynamically through a dynamic channel reweighting mechanism. Furthermore, Squeeze-and-Excitation Network (SENet) was used to compensate for the current WTConv's learning of inter-channel feature relationships. Finally, FCAC was deployed at key nodes in the backbone network. Experimental results on the COCO dataset show that although the number of parameters of the improved HRNet-W32 model is increased by 22%, the model has the floating-point operations reduced by 7.0%, the mean Average Precision (mAP) improved by 3.5 percentage points, and the convergence ability greatly enhanced.

Key words: wavelet transform, cross-attention, Depthwise Separable Convolution (DSC), High-Resolution Network (HRNet), human pose estimation

摘要:

人体姿态估计在计算机视觉领域中是一项关键任务,高分辨率网络(HRNet)因具备出色的性能而被广泛应用。然而,模型精度的提高通常伴随着参数量与计算成本飙升。为缓解这一问题,提出一种轻量化改进模块——频域交叉注意力卷积(FCAC)。首先,通过融合小波变换卷积(WTConv)与深度可分离卷积(DSC)重构基础卷积模块,利用WTConv增大感受野的优势弥补DSC上下文联系不足的缺陷,在优化HRNet性能的同时保持较低的复杂度;其次,通过注意力机制引导低频分量学习高频分量,再通过动态通道重加权机制动态地聚合学习结果;此外,使用SENet(Squeeze-and-Excitation Network)弥补当前WTConv对于通道间特征关系的学习;最后,在主干网络的关键节点上部署FCAC。在COCO数据集上的实验结果表明,改进后的HRNet-W32模型虽然参数量提升22%,但是浮点运算量降低了7.0%,平均精度均值(mAP)提升了3.5个百分点,同时大幅增强了收敛能力。

关键词: 小波变换, 交叉注意力, 深度可分离卷积, 高分辨率网络, 人体姿态估计

CLC Number: