基于人类活动的性别认同的多活动融合方法
概括
这项研究使用可穿戴传感器和机器学习进行性别识别,通过分析步行和登等多种日常活动,达到94.13%的准确性.
科学领域:
- 生物医学工程 生物医学工程
- 人与计算机的交互
- 机器学习 机器学习
背景情况:
- 在医疗保健,体育和可穿戴技术方面,性别认可越来越重要.
- 可穿戴式传感器提供了一种非侵入性的方法来收集生理数据.
- 监督机器学习模型可以分析复杂的活动数据来进行分类任务.
研究的目的:
- 开发和评估使用可穿戴惯性测量单位 (IMU) 的性别识别系统.
- 为了确定最佳的传感器放置和活动类型,以便准确的性别分类.
- 为了比较不同机器学习算法对此任务的性能.
主要方法:
- 使用了一种可穿戴传感器系统,五个IMU放置在上半身和下半身上.
- 记录了七种日常活动,包括站立,行走和爬的炼.
- 应用监督机器学习,特别是随机森林分类器 (RFC) 和支持矢量机器 (SVM),用于性别分类.
主要成果:
- 单个活动分类的最大准确度为92.06%,使用RFC在行走时使用脚传感器数据.
- 多项活动分类显著提高了准确性,在RFC中达到94.13%.
- 使用罗姆伯格测试 (眼睛开放),单腿姿势 (眼睛开放) 和爬楼梯活动的组合获得了最高的准确性.
结论:
- 穿戴式传感器数据与机器学习相结合,为性别识别提供了一种有效的方法.
- 与单一活动相比,分析多个活动可以提高分类准确性.
- 最佳的传感器放置和活动选择对于最大限度地提高性别识别性能至关重要.
相关概念视频
01:30Socioemotional Experience and Gender Development
Social-emotional experiences and cultural influences play significant roles in shaping gender development. During middle childhood, from ages 6 to 11, peer groups become dominant in reinforcing gender norms. Children in this age group often align with same-gender peer groups, which actively encourage behaviors that conform to traditional gender roles. For instance, boys may be discouraged from engaging in activities perceived as feminine, reinforcing culturally dictated norms about masculinity...
01:24Social Foundations of Self I: Play and Game
The development of self in children is deeply rooted in social interactions, mainly through stages of play and structured games. These stages, outlined by sociologist George Herbert Mead, illustrate how children progressively learn to understand and adopt social roles, forming a cohesive sense of self.The Play Stage: Imitation and Simple Role-TakingIn the early years of childhood, the play stage is characterized by imitative behavior, where children engage in role-playing based on familiar...