应用技术与研究

面向AI的语音数据采集服务平台的设计与实现

展开
  • 中国传媒大学信息与通信工程学院数字媒体技术系
王子韵(1999-) ,男,重庆万州人,本科在读,研究方向为媒体融合。

网络出版日期: 2021-11-01

Design and Implementation of the AI-oriented Platform for Voice Data Acquisition Service

Expand
  • School of Information and Communication Engineering

Online published: 2021-11-01

摘要

为解决语音AI 的方言语音数据采集存在的数据量不够多、样本分布不均衡等问题,以语音数据收集、标注、数据交叉校验、数据集打包分享为目标,设计开发了一个语音数据采集与服务平台,提供语音数据采集、任务定制、语音与文本数据管理、数据标注、数据检索、数据下载等功能,通过微信小程序和手机APP吸引用户参与有趣的语音游戏,从而实现可定制的语音数据采集、标注、交叉校验等工作,在提升语音数据量的同时,有效解决数据采集过程中的样本分布不均衡问题,提升语音数据在方言人群和地域方面覆盖范围,提升数据质量,助力方言语音识别。

本文引用格式

王子韵 钮辰洋 .

面向AI的语音数据采集服务平台的设计与实现
[J]. 电脑与电信, 2021 , 1(11) : 69 -75 . DOI: 10.15966/j.cnki.dnydx.2021.11.008

Abstract

To solve the problems of insufficient data volume and unbalanced sample distribution in dialect voice data acquisition for
voice AI, a voice data acquisition service platform is designed and developed to provide voice data acquisition, task customization,
voice and text data management, data labeling, data retrieval and download available for users to participate in enjoyable voice
games in forms of WeChat mini program and mobile app, to realize customizable voice data acquisition, labeling and cross verification. It effectively solves the problem of imbalance of sample distribution of data acquisition while increasing the amount of voice data, improves the sampling coverage of dialect population and region, and enhances the voice data quality to boost dialect speech recognition.

Options
文章导航

/