摘要
为了实现对英语翻译器准确性校正,提出基于音频信号识别的英语翻译器准确性校正方法。采用随机信道扩频方法进行英语翻译器音频信号特征检测处理,建立语音特征序列和文本特征序列的标量时间序列,结合英语翻译器音频信号的包络幅值检测和信息聚类方法,提取英语翻译器音频信号聚类特征参数,通过信号滤波检测去除音频信号的干扰分量,采用翻译字典集和语义上下文匹配,实现音频信号的调制和解调。建立英语翻译器音频信号的特征聚类和检测模型,通过多维谱分析和高阶累积量检测,实现英语翻译器音频信号的输出稳定性检测,根据输出信号与发音标准信号的差异性,实现英语翻译器的准确性检测识别。仿真结果表明,采用该方法进行英语翻译器准确性校正的准确性较高,检测性能较好。与对照方法相比,本方法提高了英语翻译器的输出信号信噪比与语音识别准确率。
In order to correct the accuracy of English translators, an accuracy correction method of English translators based on audio signal recognition is proposed. The random channel spread spectrum method is used to detect the audio signal characteristics of English translator, and the scalar time series of speech feature sequence and text feature sequence is established. Combining with the envelope amplitude detection and information clustering method of English translator audio signal, the clustering feature parameters of English translator audio signal are extracted, the interference components of audio signal are removed by signal filtering detection, and the modulation and demodulation of audio signal are realized by translation dictionary set and semantic context matching. The feature clustering and detection model of English translator’s audio signal is established. Through multi-dimensional spectrum analysis and higher-order cumulant detection, the output stability of English translator’s audio signal is detected. According to the difference between the output signal and pronunciation standard signal, the accuracy of English translator is detected and identified. The simulation results show that this method has higher accuracy and better detection performance. Compared with the control method, the proposed method improves the output signal signal-to-noise ratio and speech recognition accuracy of the English translator.
作者
谢丹
殷超
XIE Dan;YIN Chao(Xi’an Traffic Engineering Institute,Xi’an 7100000,China;Universiti Teknologi MARA(UiTM),Malaysia Shah Alam 40450;Shaanxi Normal University,Xi’an 710000,China;Xi An Tieyi Middle School,Xi’an 710000,China)
出处
《自动化与仪器仪表》
2022年第11期27-30,35,共5页
Automation & Instrumentation
基金
2022年西安交通工程学院校级中青年基金项目《生态翻译学视域下西安十四运公示语英译研究》(22KY-67)。
关键词
音频信号
英语翻译器
准确性
校正
检测
包络幅值
audio signal
english translator
accuracy
correction
testing
envelope amplitude
