PoseAlign网络用于二维人类姿势估计中的混合结构
Jin Zhang1, Yabo Yin2, Wenzhong Yang3,4
1School of Software, Xinjiang University, Urumqi, 830091, China.
Scientific reports
|May 16, 2025
概括
本研究介绍了混合结构的PoseAlign网络 (PAN-HS),这是一个结合视觉转换器和卷积神经网络的新方法,用于改进人类姿势估计 (HPE). PAN-HS提高了关键点定位的准确性,在MPII数据集上实现了卓越的性能.
科学领域:
- 计算机视觉 计算机视觉
- 机器学习 机器学习
背景情况:
- 视觉转换器 (ViT) 和卷积神经网络 (CNN) 在人类姿势估计 (HPE) 中占据着重要地位.
- 混合方法可以利用ViT和CNN的优势来增强HPE.
研究的目的:
- 引入一种新的2D HPE方法,即混合结构的PoseAlign网络 (PAN-HS).
- 结合ViT和CNN的优势,以提高关键点定位精度.
主要方法:
- 设计的空间对齐和通道对齐块使用深度智能卷积来提取特征.
- 实施了一种高效的点重置注意力机制,以平衡本地和全球特征表示.
- 聚合的内层特征用于细粒度的局部表示.
主要成果:
- 在MPII数据集上获得了92.74%的平均PCKh.
- 在HPE任务中表现出卓越的性能和增强的关键点定位精度.
- 改善了中层空间和通道特征的表示.
结论:
- PAN-HS方法有效地整合了ViT和CNN,用于先进的人体姿势估计.
- 提出的注意力机制和特征提取块有助于精确的关键点定位.
- PAN-HS为混合HPE架构的未来研究提供了一个有希望的方向.
更多相关视频
05:41A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis
Published on: February 6, 2020
9.3K
10:23Author Spotlight: Three-Dimensional Cephalometric Landmark Annotation Demonstration on Human Cone Beam Computed Tomography Scans
Published on: September 8, 2023
2.6K
相关概念视频
02:06Newman Projections
16.2K
Different notations are used to represent the three-dimensional structure of molecules on two-dimensional surfaces. One of the most commonly used representations is the dash-wedge formula. The dashed wedges, solid wedges, and the plane lines indicate the groups situated behind the plane, coming out of the plane, and in the plane, respectively.
The organic molecules rotate across the single bonds leading to numerous temporary three-dimensional structures of varying energy known as...
The organic molecules rotate across the single bonds leading to numerous temporary three-dimensional structures of varying energy known as...
16.2K
03:35Hybridization of Atomic Orbitals II
31.4K
sp3d and sp3d 2 Hybridization
31.4K