Speech recognition test method and apparatus

CN116665647BActive Publication Date: 2026-06-19MASHANG CONSUMER FINANCE CO LTD

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
MASHANG CONSUMER FINANCE CO LTD
Filing Date
2022-02-18
Publication Date
2026-06-19

AI Technical Summary

Technical Problem

The existing performance testing equipment for speech recognition systems has a complex and incomplete testing process, resulting in low testing efficiency.

Method used

By setting active speech detection parameters in the server's configuration file, combining the client's configuration file and annotation file, test metrics are calculated, and the parameter values ​​of active speech detection are adjusted according to the metrics to achieve automated speech recognition testing between the server and the client.

Benefits of technology

It improves the accuracy and efficiency of speech recognition testing, automates the speech recognition testing process, reduces testing time, and improves testing precision.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116665647B_ABST
    Figure CN116665647B_ABST
Patent Text Reader

Abstract

This specification provides a speech recognition testing method and apparatus. One speech recognition testing method includes: setting a first parameter value for an active speech detection parameter in a server configuration file for use by the server and client in performing speech recognition testing; reading speech recognition test results from a test log and calculating test metrics based on the test results and a labeling file, wherein the test log is generated on the client; determining a second parameter value for the active speech detection parameter within its value range based on the test metrics; and updating the first parameter value based on the second parameter value. Using this embodiment, the speech recognition testing process performed by the server and client can be automated, which helps improve the efficiency of speech recognition testing.
Need to check novelty before this filing date? Find Prior Art