相关实验视频

Updated: Sep 16, 2025

Evaluation of a Smartphone-based Human Activity Recognition System in a Daily Living Environment

Evaluation of a Smartphone-based Human Activity Recognition System in a Daily Living Environment

Published on: December 11, 2015

基于骨架的轻量级和高效的体育活动识别与ASTM-Net

Bin Wu¹, Mei Xue², Ying Jia³

¹College of Light Textile and Chemical Engineering, Binzhou Polytechnic, Shandong, China.

|July 8, 2025

概括

通过动态建模空间亲和关系和时间依赖关系,ASTM-Net通过使用骨架数据推进了人类活动识别. 这种新的方法显著提高了准确性,同时降低了计算成本,提高了阻塞的稳定性.

科学领域:

计算机视觉计算机视觉
人工智能的人工智能
机器学习机器学习

背景情况:

人类活动识别 (HAR) 对于视频理解至关重要,骨数据提供了对环境变化的稳定性.
现有的图形卷积网络 (GCNN) 方法难以处理动态的空间关系,并将时空信息集成到复杂的操作中.

研究的目的:

介绍ASTM-Net,一个新的活动意识的时空空间多分支图形卷积网络,旨在克服骨架HAR的局限性.
增强捕获动态节点亲和关系以及空间和时间特征之间的相互作用,以改进动作识别.

主要方法:

开发了活动意识空间图卷积模块 (ASGM),通过合并多种图型来动态建模活动意识相邻图 (3A图).
引入了时间多分支图形卷积模块 (TMGM),使用具有扩展卷积的并行分支和聚合来实现高效的时间建模.
集成ASGM和TMGM以共同捕获时空信息,降低计算复杂度.

主要成果:

在基准数据集 (NTU-RGB+D,NTU-RGB+D 120,Toyota Smarthome) 上,ASTM-Net实现了卓越的性能,超过了最先进的方法.
与MST-GCNN-ALL相比,显示了参数 (51.9%) 和FLOP (49.7%) 的显著降低,同时精度提高了0.82%.
在30%的随机节点封闭下,展示了高精度 (86.94%) ,表明了强度.

结论:

通过动态建模空间亲和关系和时间依赖关系,ASTM-Net有效地解决了骨 HAR 中现有的 GCNNs 的局限性.
拟议的多分支架构为复杂的活动识别任务提供了一个参数高效和计算有效的解决方案.
ASTM-Net代表了基于骨的人类活动识别的重大进步,提供了更高的准确性,效率和稳定性.

更多相关视频

An Objective and Child-friendly Assessment of Arm Function by Using a 3-D Sensor

An Objective and Child-friendly Assessment of Arm Function by Using a 3-D Sensor

Published on: February 12, 2018

An Inertial Measurement Unit Based Method to Estimate Hip and Knee Joint Kinematics in Team Sport Athletes on the Field

An Inertial Measurement Unit Based Method to Estimate Hip and Knee Joint Kinematics in Team Sport Athletes on the Field

Published on: May 26, 2020

相关概念视频

Carbon Skeletons

Carbon Skeletons

Life on Earth is carbon-based, as all macromolecules that make up living organisms contain carbon atoms. All organic compounds have a carbon backbone. Each carbon atom is tetravalent and can bond with four other atoms, making it an extraordinarily flexible component of biological molecules. Because carbon’s valence electrons are stable, it rarely becomes an ion. As the carbon chain increases in length, structural modifications such as ring structures, double bonds, and branching side...

Classification of Skeletal Muscle Fibers

Classification of Skeletal Muscle Fibers

Skeletal muscles continuously produce ATP to provide the energy that enables muscle contractions. Skeletal muscle fibers can be categorized into three types based on differences in their contraction speed and how they produce ATP, as well as physical differences related to these factors. Most human muscles contain all three muscle fiber types, albeit in varying proportions.
Slow-Twitch Muscle Fibers
Slow oxidative, muscle fibers appear red due to large numbers of capillaries and high levels of...