摘要
得益于高性能计算机和深度学习算法的不断进步,生物特征识别技术得到快速发展。脚步声识别技术利用人行走时发出的声音或震动信号实现行走人的身份识别,具有隐蔽性、非接触式、不易被伪造和无需被识别人配合等优点,是一种非常有潜力的生物特征识别技术。本文概述了脚步声识别系统的组成部分、基本概念以及脚步声识别系统的性能评价指标,总结了脚步声识别技术中的信号采集方法及其使用的仪器设备、降噪和端点检测信号预处理关键技术、时频域和声学特征参数提取、各种模式识别方法在脚步声识别技术中的应用等方面的研究成果。最后,系统地分析了脚步声识别技术研究中尚未解决的问题,探讨了该研究领域需要进一步探索的热点问题和未来的发展趋势。
Footstep acoustic recognition,one biometric distinguishing technology,is very promising with the continuous advancement of high-performance computer and deep-learning algorithm.Such a technology adopts the sound or vibration signals from walking people to recognize the relevant individual.It is specifi c of concealment,non-contact,diffi culty to forge,and no requirement for cooperation of the recognized people.This article summarizes the footstep acoustic recognition system about its composition,basic concepts and performance evaluation indicators,with elucidation of the research development and progress.The focuses were paid onto the signal acquisition methods and eligible equipment,key technologies for noise reduction and endpoint detection signal preprocessing,temporal frequency domain and extraction of acoustic feature parameters,together with the applications of various pattern recognition approaches.Finally,the problems unknown of solution are systematically analyzed,with the discussions being made into both the hot issues necessary for further exploration and future trend.
作者
房玉杰
张松
刘晋
冯磊
FANG Yujie;ZHANG Song;LIU Jin;FENG Lei(Ministry of Education’s Key Laboratory of High Effi ciency and Clean Mechanical Manufacture&School of Mechanical Engineering,Shandong University,Jinan 250061,China;National Demonstration Center for Experimental Mechanical Engineering Education,Shandong University,Jinan 250061,China;Institute of Forensic Science,Ministry of Public Security,Beijing 100038,China)
出处
《刑事技术》
2021年第1期81-86,共6页
Forensic Science and Technology
基金
痕迹科学与技术公安部重点实验室开放课题(2019FMKFKT06)。
关键词
脚步声识别
信号采集
预处理
特征参数提取
模式识别
footstep acoustic recognition
signal acquisition
preprocessing
feature parameter extraction
pattern recognition
