欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
随着网络的发展,人们对音乐质量的要求越来越高,无损音乐的交流也从音乐发烧友间向普通大众过渡.但网络上的无损音乐资源质量得不到保证,对于音乐类应用公司来说,也无法通过人工过滤其千万级别的曲库来改善曲库质量.提出了一种基于贝叶斯网络的无损音乐检测算法,该算法将音乐频率和音乐类型的关系考虑在内,通过大量标记好的真假无损音乐数据进行训练,训练好的模块可应用于音乐质量的判断.实验证明,该算法具有较快的运行速度和较高的检测准确率.
英文摘要：
As the web evolves, the need for people to improve the quality of music will become more acute. The communication of lossless audio has transformed from music fans to the general public. However, there is no guaranteed for the quality of music resources on the network. For music applications, this is not even realistic if companies improve the quality of their ten million library with artificial methods. A lossless audio detection method is proposed based on Bayesian networks, which takes into consideration the relationships between different spectrum and the relationships musical quality. Is is a powerful methods for judging the quality of music through a large number of learning and training. The experimental results verify that the algorithm has a faster speed and higher accuracy.
同期刊论文项目
同项目期刊论文
期刊信息
位置：