一种数字乐谱音乐信息快速智能提取算法  

Fast and intelligent algorithm for extracting digital score music information

作　　者：王有能 WANG Youneng(Guangxi Modern Vocational and Technical College,Hechi 547000,China)

机构地区：[1]广西现代职业技术学院,广西河池547000

出　　处：《无线互联科技》2024年第18期56-58,共3页Wireless Internet Science and Technology

摘　　要：数字乐谱信息来源广泛且格式多样,导致提取目标信息花费时间较长,为此,文章提出一种数字乐谱音乐信息快速提取算法。从初始音乐信息中提取一级种子URL,将其加入待爬取的数字乐谱URL队列,识别并匹配与目标乐谱相关的特征。基于深度学习算法筛选出与目标乐谱相匹配的信息部分,实现快速准确的信息提取。经验证,该算法在确保FI值达到0.98的前提下,提取测试曲目乐谱的高音音符、长音音符以及节奏音符信息的整体时间开销仅为115 s,展现出了出色的信息提取速度。Digital score information has a wide source and various formats,which leads to a long time to extract target information.Therefore,this study proposes a fast algorithm for the extraction of digital score music information.The primary seed URL was extracted from the initial music information and added to the digital score URL queue to be crawled to identify and match the features related to the target score.Based on the deep learning algorithm,the information that matches the target score is selected to achieve fast and accurate information extraction.It is proved that on the premise of ensuring the FI value of 0.98,the overall time cost of extracting the test track score is only 115 s,showing excellent information extraction speed.

关 键 词：网络爬虫 数字乐谱 音乐信息 快速提取技术 深度学习 目标特征 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]