理解数字声音——基于一般音频/环境声的计算机听觉综述  

Understanding Digital Audio——A Review of General Audio/Ambient Sound based Computer Audition

作　　者：李伟[1,2] 李硕 LI Wei;LI Shuo(School of Computer Science and Technology,Fudan University,Shanghai 201203,China;Shanghai Key Laboratory of Intelligent Information Processing,Fudan University,Shanghai 200433,China)

机构地区：[1]复旦大学计算机科学技术学院,上海201203 [2]复旦大学上海市智能信息处理重点实验室,上海200433

出　　处：《复旦学报（自然科学版）》2019年第3期269-313,共45页Journal of Fudan University：Natural Science

基　　金：国家自然科学基金(61671156)

摘　　要：声音是人类获取信息的重要来源,对声音内容进行自动分析和理解具有重要意义.本文介绍声音的基本知识,从信号、听觉感受、声音特性等3个角度对声音进行分类,阐明各个分类之间的关系,明确基于一般音频/环境声的计算机听觉技术的研究对象和学科位置.之后,介绍计算机听觉技术的基本概念、原理、研究课题和技术框架.作者全面总结了计算机听觉技术在各个领域中:包括医疗卫生,安全保护,交通运输、仓储,制造业,农、林、牧、渔业,水利、环境和公共设施管理业,建筑业,其他采矿业、日常生活、身份识别、军事等的典型应用.分类总结了各领域计算机听觉应用中现有典型文献的基本原理、技术路线.最后总结计算机听觉领域存在的各方面问题,并展望未来发展趋势.Sound is one of the main sources for human to obtain information.It is of great importance to automatically analyze and understand the content of sound.This paper introduces the basic knowledge of sound,classifies various sounds from three digital perspectives,i.e.,signal,auditory perception and sound characteristics,and clarifies the relationship between each class,as well as makes clear the research object and discipline position of general audio based Computer Audition(CA).Next,we describe the basic concept,principles,research topics,and technical framework of computer audition.Typical applications of computer audition are comprehensively summarized in various fields,including health care,safety protection,transportation,storage,manufacturing,agriculture,forestry,animal husbandry,fishery,water conservancy,environment protection,construction,mining,daily life,identification,military etc.With each kind of CA application,the basic principles and technical route in typical papers are outlined.Finally,we analyze the problems that hinder the development of CA,and prospect the bright future.

关 键 词：数字声音 一般音频/环境声 计算机听觉 音频信号处理 人工智能 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]