Search research articles

关于 JoVE

概览领导团队博客 JoVE 帮助中心

作者

出版流程编辑委员会范围与政策同行评审常见问题投稿

图书馆员

用户评价订阅访问资源图书馆顾问委员会常见问题

研究

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments 存档

教育

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual 教师资源中心教师网站

使用条款与条件

相关概念视频

您也可能阅读

相关文章

通过共同作者、期刊和引用图与本文相关的文章。

排序

Same author

PANDA: Patch-based unsupervised deep learning for brain anomaly detection via age prediction in fetal MRI.

Imaging neuroscience (Cambridge, Mass.)·2026

Same author

High dynamic range structured illumination microscopy based on multi-channel fusion and dual-phase binary fringe.

Optics express·2026

Same author

M1 macrophage-derived exosomal miR-155-5p exacerbates aortic dissection via SMAD5-Mediated regulation of vascular smooth muscle cell phenotype.

Bioactive materials·2026

Same author

Synthetic Chimeric Antigen Lysosome-Associated Receptor Redirects Targeted Protein to Degradation.

Journal of the American Chemical Society·2026

Same author

EAGP: an efficient generative augmentation framework for phage protein classification under severe class imbalance.

Bioinformatics (Oxford, England)·2026

Same author

A high-linearity, low-power 10-mm MEMS fast steering mirror with enhanced shock robustness for space laser communications.

Optics letters·2026

Search research articles

相关实验视频

Updated: Jun 2, 2025

Optical Recording of Suprathreshold Neural Activity with Single-cell and Single-spike Resolution

Optical Recording of Suprathreshold Neural Activity with Single-cell and Single-spike Resolution

Published on: September 5, 2012

Spiking-PhysFormer:基于相机的远程光电解剖学与并行尖驱动的变压器.

Mingxuan Liu¹, Jiankai Tang¹, Yongli Chen²

¹Tsinghua University, Beijing, China.

Neural networks : the official journal of the International Neural Network Society

|January 16, 2025

概括

本研究介绍了SNN (尖端神经网络) 用于在移动设备上进行节能远程光电显微镜 (rPPG). 新的Spiking-PhysFormer模型显著降低了功耗,同时保持了从面部视频中准确的生理信号测量.

关键词:

生物医学信号生物医学信号大脑启发的神经网络.遥控光电解声学遥控光电解声学变压器变压器变压器

更多相关视频

A Simple Stimulatory Device for Evoking Point-like Tactile Stimuli: A Searchlight for LFP to Spike Transitions

A Simple Stimulatory Device for Evoking Point-like Tactile Stimuli: A Searchlight for LFP to Spike Transitions

Published on: March 25, 2014

Using Neuron Spiking Activity to Trigger Closed-Loop Stimuli in Neurophysiological Experiments

Using Neuron Spiking Activity to Trigger Closed-Loop Stimuli in Neurophysiological Experiments

Published on: November 12, 2019

相关实验视频

Last Updated: Jun 2, 2025

Optical Recording of Suprathreshold Neural Activity with Single-cell and Single-spike Resolution

Optical Recording of Suprathreshold Neural Activity with Single-cell and Single-spike Resolution

Published on: September 5, 2012

A Simple Stimulatory Device for Evoking Point-like Tactile Stimuli: A Searchlight for LFP to Spike Transitions

A Simple Stimulatory Device for Evoking Point-like Tactile Stimuli: A Searchlight for LFP to Spike Transitions

Published on: March 25, 2014

Using Neuron Spiking Activity to Trigger Closed-Loop Stimuli in Neurophysiological Experiments

Using Neuron Spiking Activity to Trigger Closed-Loop Stimuli in Neurophysiological Experiments

Published on: November 12, 2019

科学领域:

生物医学工程生物医学工程
计算机科学计算机科学
人工智能的人工智能

背景情况:

基于摄像头的远程光电显微镜 (rPPG) 使用人工神经网络 (ANN) 来准确测量面部视频中的生理信号.
现有的基于ANN的rPPG方法的高计算需求限制了它们在资源有限的移动设备上部署.
尖端神经网络 (SNN) 通过它们的二进制事件驱动架构提供节能深度学习.

研究的目的:

首次将SNN引入rPPG,开发一种名为Spiking-PhysFormer的节能混合神经网络 (HNN) 模型.
为了减少实际移动设备应用的rPPG系统的功耗.
与现有方法相比,保持或提高生理信号提取 (如心率,呼吸率) 的准确性.

主要方法:

提出了一个混合神经网络 (HNN) 模型,Spiking-PhysFormer,集成ANN和SNN组件.
开发了一个并行尖峰变压器块,以增强SNN中的时空特征聚合.
引入了一种简化的尖端自我注意机制,省略了值参数,以减少计算负载而不会降低性能.

主要成果:

与PhysFormer模型相比,Spiking-PhysFormer实现了10.1%的整体功耗降低.
变压器块内的电力消耗被减少了12.2.2的系数.
该模型在四个基准数据集 (PURE,UBFC-rPPG,UBFC-Phys,MMPD) 中表现出与PhysFormer和其他基于ANN的模型相当的性能.

结论:

拟议的Spiking-PhysFormer有效地利用SNN进行节能rPPG,解决移动设备上传统ANN的局限性.
混合方法和简化的SNN机制为低功耗,准确的生理监测提供了一个有希望的方向.
这项工作为在具有有限计算资源的边缘设备上部署先进的rPPG技术铺平了道路.