用耳机传感器实时监测低头姿势，精准预防科技颈
NeckCare: Preventing Tech Neck using Hearable-based Multimodal Sensing
- 通过耳机内置传感器融合姿态与声音数据
- 姿势识别准确率达99%，距离估算误差在毫米级
- 适合长期用手机/电脑人群，帮助改善体态
科技颈是由长时间使用电子设备引发的现代健康问题，可能导致严重颈部疲劳和不适。本文提出NeckCare系统，利用可穿戴耳机传感器（包括惯性测量单元和麦克风）实现非侵入式、无处不在的感知，实时监测科技颈姿势并估算用户与屏幕的距离。基于15名参与者的数据，仅使用IMU数据即可实现96%的姿势分类准确率，结合音频数据提升至99%。距离估计算法在嘈杂环境中仍保持毫米级精度。NeckCare能即时向用户提供反馈，促进健康姿势，缓解颈部压力。未来工作将探索个性化提醒、肌肉应变预测、颈部运动检测及数字眼疲劳预测的集成。
原文摘要 · Abstract (English)
Tech neck is a modern epidemic caused by prolonged device usage and it can lead to significant neck strain and discomfort. This paper addresses the challenge of detecting and preventing tech neck syndrome using non-invasive ubiquitous sensing techniques. We present NeckCare, a novel system leveraging hearable sensors, including IMUs and microphones, to monitor tech neck postures and estimate distance form screen in real-time. By analyzing pitch, displacement, and acoustic ranging data from 15 participants, we achieve posture classification accuracy of 96% using IMU data alone and 99% when combined with audio data. Our distance estimation technique is millimeter-level accurate even in noisy conditions. NeckCare provides immediate feedback to users, promoting healthier posture and reducing neck strain. Future work will explore personalizing alerts, predicting muscle strain, integrating neck exercise detection and enhancing digital eye strain prediction.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。