视频运动放大技术发展分析  

Analysis of the Development of Video Motion Magnification Technology

作　　者：曹鸿博 CAO Hongbo(College of Computer and Infomation Technology,China Three Gorges University,Yichang 443002,China)

机构地区：[1]三峡大学计算机与信息学院,湖北宜昌443002

出　　处：《长江信息通信》2024年第4期80-82,共3页Changjiang Information & Communications

摘　　要：人们受限于身体机能,对许多重要信息无法直观获取。往往借助各种传感器,伴随着视频放大技术的兴起,人们借助计算机视觉技术起到了辅助桥梁监测的作用。这是由于视频运动放大技术可以将许多人眼无法直观看到的细微运动呈现出来,来实现对运动的观察和监测。随着计算机视觉领域的迅猛发张,视频运动放大技术,也从最初的拉格朗日放大方法、欧拉放大方法、相位放大方法,逐步过渡到深度学习模型和Transformer模型。之后,该技术更是被广泛应用于多个场景中,例如:微表情识别、动态行为检测、非接触心理检测等。People are limited by their physical functions and cannot intuitively access many important information.Often,they rely on various sensors,and with the rise of video magnification tcchnology,they usc computer vision tcchnology to play an auxiliary role in bridge monitoring.This is bccause vidco motion magnification technology can prcsent many subtle motions that human eyes cannot directly see,to achieve the observation and monitoring of motion.With the rapid development of the field of computer vision,video motion magnification technology has also gradually transitioned from the initial Lagrangian magnification method,Euler magnification method,phase magnification method,to deep learning models and Transformer models.Later,this technology was applied to multiple scenarios,such as:micro-expression recognition,dynamic bchavior detection,non-contact psychological detcction,ctc.

关 键 词：视频运动放大 欧拉方法 深度学习模型 Transformer模型 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]