基于大模型的外呼呼叫中心语音识别方法及系统

By calculating the feature values ​​of speech signal frames and obtaining adaptive subtraction factors, the problem of decreased recognition accuracy caused by non-stationary noise and speech aliasing is solved, and efficient speech recognition in noisy environments is achieved.

CN121811863BActive Publication Date: 2026-07-17JINAN YUNSHANG ELECTRONIC TECH CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
JINAN YUNSHANG ELECTRONIC TECH CO LTD
Filing Date
2026-01-31
Publication Date
2026-07-17

AI Technical Summary

Technical Problem

Existing spectral subtraction methods cannot effectively separate and reduce noise when faced with non-stationary noise and highly aliased speech, resulting in a decrease in speech recognition accuracy.

Method used

A large model-based approach is adopted to obtain an adaptive subtraction factor by calculating the first and second feature values ​​of the speech signal frame, and then use the speech model to reduce noise and recognize the speech in the call.

Benefits of technology

It improves speech recognition accuracy in noisy environments, effectively separates and reduces noise, and enhances call quality for outbound call centers.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121811863B_ABST
    Figure CN121811863B_ABST
Patent Text Reader

Abstract

本发明涉及噪音环境语音识别技术领域，提出了基于大模型的外呼呼叫中心语音识别方法及系统，包括：采集通话语音并进行分帧处理，获取语音信号帧；获取语音信号帧的不同IMF分量，计算语音信号帧的第一特征值；将语音信号帧转换为梅尔语谱图，确定语音信号帧的第二特征值；根据语音信号帧的第一特征值和第二特征值，计算语音信号帧的自适应减法因子，根据自适应减法因子获取降噪的通话语音，使用语音模型对同一外呼呼叫的通话对应的所有降噪的通话语音进行语音识别。本发明可提升非平稳噪声影响下语音识别的准确率。
Need to check novelty before this filing date? Find Prior Art