论文部分内容阅读
该项目是我国目前首次将计算机音频自动处理技术综合应用于大规模广播电视监测的现代化、高技术的自动化监测系统,充分体现了技术应用模式创新和集成创新两方面特点。首创了采用多种抗噪声技术,对信噪比较低的短波信道音质的广播语音实现了高可靠度的语种识别,在15个语种条件下识别准确率达到95%。同时应用基于似然比
This project is the first modern and high-tech automatic monitoring system that comprehensively applies computer audio automatic processing technology to large-scale radio and television monitoring in our country and fully embodies the characteristics of both technical application mode innovation and integrated innovation. Pioneered the use of a variety of anti-noise technology, the shortest signal to noise ratio of short-wave channel sound quality broadcast voice to achieve a high degree of reliability of the language recognition in 15 language conditions, the recognition accuracy of 95%. At the same time the application based on the likelihood ratio