吉林大学学报（理学版）2024，Vol.62Issue(4)：943-950,8.DOI:10.13413/j.cnki.jdxblxb.2023298
基于融合特征ADRMFCC的语音识别方法
Speech Recognition Method Based on Fusion Feature ADRMFCC
摘要
Abstract
Aiming at the problem of low accuracy and poor robustness of speech recognition in complex noise environment,we proposed a speech recognition method based on Mel cepstrum fusion feature of increasing and decreasing residuals.This method first used the increase and decrease component method to screen the key speech features,and then mapped them to the Mel domain-residual domain spatial coordinate system to generate the increase and decrease residual Mel cepstral coefficients.Finally,these fusion features were used to train the end-to-end model.The experimental results show that the proposed method significantly improves the accuracy and performance of speech recognition under different noise types and signal-to-noise ratio conditions.Under the low signal-to-noise ratio condition of-5 dB,the speech recognition accuracy reaches 73.13％,while the average speech recognition accuracy under other noise conditions reaches 88.67％,which fully proves the effectiveness and robustness of the proposed method.关键词
语音识别/残差Mel倒谱系数/特征筛选/增减分量法Key words
speech recognition/residual Mel cepstral coefficient/feature screening/increase and decrease component method分类
信息技术与安全科学引用本文复制引用
朵琳,马建,韦贵香,唐剑..基于融合特征ADRMFCC的语音识别方法[J].吉林大学学报（理学版）,2024,62(4):943-950,8.基金项目
国家自然科学基金(批准号:61962032). （批准号:61962032）
