首页> 外国专利> DEVICE FOR ESTIMATING DETERIORATION FACTOR OF SPEECH RECOGNITION ACCURACY, METHOD FOR ESTIMATING DETERIORATION FACTOR OF SPEECH RECOGNITION ACCURACY, AND PROGRAM

DEVICE FOR ESTIMATING DETERIORATION FACTOR OF SPEECH RECOGNITION ACCURACY, METHOD FOR ESTIMATING DETERIORATION FACTOR OF SPEECH RECOGNITION ACCURACY, AND PROGRAM

机译:语音识别精度的估计因子的装置,语音识别精度的估计因子的方法和程序

摘要

The present invention provides a device for estimating the deterioration factor of speech recognition accuracy, the device capable of estimating an acoustic factor that leads to a speech recognition error. The device for estimating the deterioration factor of speech recognition accuracy comprises: an acoustic feature amount extraction unit for extracting an acoustic feature amount for each frame from a speech that is input; a posterior probability calculation unit for calculating, on the basis of a plurality of acoustic events preliminarily classified into either a deterioration factor class or a non-deterioration factor class, a posterior probability for each acoustic event for the acoustic feature amount for each frame; a filter unit for correcting the posterior probability by filtering the posterior probability for each acoustic event using a time-series filter with weighting coefficients developed in the time axis; a speech recognition unit for outputting a set of speech recognition results with a recognition score; a speech recognition result feature amount extraction unit for outputting a feature amount for the speech recognition results for each frame; and a deterioration factor output unit for calculating and outputting a principal deterioration factor class for the speech recognition accuracy for each frame on the basis of the corrected posterior probability, the feature amount for speech recognition results for each frame, and the acoustic feature amount for each frame.
机译:本发明提供了一种用于估计语音识别精度的劣化因素的设备,该设备能够估计导致语音识别错误的声学因素。用于估计语音识别精度的劣化因子的设备包括:声学特征量提取单元,用于从输入的语音中提取每帧的声学特征量;以及后验概率计算单元,用于基于预先被分类为劣化因素类别或非劣化因素类别的多个声学事件,针对每个帧的声学特征量,针对每个声学事件计算后验概率;滤波器单元,其通过使用在时间轴上具有加权系数的时间序列滤波器来对每个声音事件的后验概率进行滤波来校正后验概率;语音识别单元,用于输出具有识别分数的一组语音识别结果;语音识别结果特征量提取单元,用于针对每个帧输出语音识别结果的特征量;劣化因子输出单元,用于基于校正后验概率,每一帧的语音识别结果的特征量以及每一声学特征量,计算并输出每一帧的语音识别精度的主要劣化因子类别。帧。

著录项

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号