Strain sensors with high stretchability, broad strain range, high sensitivity, and good reliability are desirable, owing to their promising applications in electronic skins and human motion monitoring systems. In this...Strain sensors with high stretchability, broad strain range, high sensitivity, and good reliability are desirable, owing to their promising applications in electronic skins and human motion monitoring systems. In this paper, we report a high- performance strain sensor based on printable and stretchable electrically con- ductive elastic composites. This strain sensor is fabricated by mixing silver-coated polystyrene spheres （PS@Ag） and liquid polydimethylsiloxane （PDMS） and screen-printed to a desirable geometry. The strain sensor exhibits fascinating comprehensive performances, including high electrical conductivity （1.65 × 104 S/m）, large workable strain range （〉 80%）, high sensitivity （gauge factor of 17.5 in strain of 0%-10%, 6.0 in strain of 10%-60% and 78.6 in strain of 60%-80%）, inconspicuous resistance overshoot （〈 15%）, good reproducibility and excellent long-term stability （1,750 h at 85℃/85% relative humidity） for PS@Ag/PDMS-60, which only contains - 36.7 wt.% of silver. Simultaneously, this strain sensor provides the advantages of low-cost, simple, and large-area scalable fabrication, as well as robust mechanical properties and versatility in applications. Based on these performance characteristics, its applications in flexible printed electrodes and monitoring vigorous human motions are demonstrated, revealing its tremendous potential for applications in flexible and wearable electronics.展开更多
针对三维人体姿态估计的便捷性与准确性提升需求,提出一种基于TM-Net网络估计算法。该算法以MediaPipe为中心,融合帧率计算、动作检测、动作计数和真实坐标解析等多功能模块,实现对人体运动的精准检测与计数。针对公共数据集LSP(Leeds S...针对三维人体姿态估计的便捷性与准确性提升需求,提出一种基于TM-Net网络估计算法。该算法以MediaPipe为中心,融合帧率计算、动作检测、动作计数和真实坐标解析等多功能模块,实现对人体运动的精准检测与计数。针对公共数据集LSP(Leeds Sports Pose)和自建校园健身房运动数据集使用关键点的正确性概率(Probability of Correct Keypoint,PCK)、关节位置误差平均值(Mean Per Joint Position Error,MPJPE)和普罗克鲁斯对齐后的平均关节位置误差(Procrustes-Aligned Mean Per Joint Position Error,PA-MPJPE)等指标对该算法进行评估,并与目前先进的TP-3D网络估计算法进行对比。结果表明,TM-Net具有更高的准确率。此外,以开合跳为例进行消融实验,结果表明,TM-Net具有更强的泛化能力,能适应不同个体及拍摄角度的变化,满足了运动监测的实际需求。展开更多
基金This work was supported by the National Key R＆D Project from Minister of Science and Technology of China （No. 2016YFA0202702）, National Natural Science Foundation of China （Nos. 61701488 and 21571186）, Leading Scientific Research Project of Chinese Academy of Sciences （No. QYZDY-SSW-JSC010）, Youth Innovation Promotion Association （No. 2017411）, Guangdong Provincial Key Laboratory （No. 2014B030301014）, Guangdong TeZhi Plan Youth Talent of Science and Technology （No. 2014TQ01C102）, Shenzhen Basic Research plan （Nos. JSGG20150512145714246 and JSGG20160229155249762） and SIAT Innovation Program for Excellent Young Researchers （No. 2016005）.
文摘Strain sensors with high stretchability, broad strain range, high sensitivity, and good reliability are desirable, owing to their promising applications in electronic skins and human motion monitoring systems. In this paper, we report a high- performance strain sensor based on printable and stretchable electrically con- ductive elastic composites. This strain sensor is fabricated by mixing silver-coated polystyrene spheres （PS@Ag） and liquid polydimethylsiloxane （PDMS） and screen-printed to a desirable geometry. The strain sensor exhibits fascinating comprehensive performances, including high electrical conductivity （1.65 × 104 S/m）, large workable strain range （〉 80%）, high sensitivity （gauge factor of 17.5 in strain of 0%-10%, 6.0 in strain of 10%-60% and 78.6 in strain of 60%-80%）, inconspicuous resistance overshoot （〈 15%）, good reproducibility and excellent long-term stability （1,750 h at 85℃/85% relative humidity） for PS@Ag/PDMS-60, which only contains - 36.7 wt.% of silver. Simultaneously, this strain sensor provides the advantages of low-cost, simple, and large-area scalable fabrication, as well as robust mechanical properties and versatility in applications. Based on these performance characteristics, its applications in flexible printed electrodes and monitoring vigorous human motions are demonstrated, revealing its tremendous potential for applications in flexible and wearable electronics.
文摘针对三维人体姿态估计的便捷性与准确性提升需求,提出一种基于TM-Net网络估计算法。该算法以MediaPipe为中心,融合帧率计算、动作检测、动作计数和真实坐标解析等多功能模块,实现对人体运动的精准检测与计数。针对公共数据集LSP(Leeds Sports Pose)和自建校园健身房运动数据集使用关键点的正确性概率(Probability of Correct Keypoint,PCK)、关节位置误差平均值(Mean Per Joint Position Error,MPJPE)和普罗克鲁斯对齐后的平均关节位置误差(Procrustes-Aligned Mean Per Joint Position Error,PA-MPJPE)等指标对该算法进行评估,并与目前先进的TP-3D网络估计算法进行对比。结果表明,TM-Net具有更高的准确率。此外,以开合跳为例进行消融实验,结果表明,TM-Net具有更强的泛化能力,能适应不同个体及拍摄角度的变化,满足了运动监测的实际需求。
