面向电视台录音采访文字校对软件  

Text Proofreading Software for TV Recording and Interview

作　　者：韦文熠 王沛文 赵莹鑫 WEI Wenyi;WANG Peiwen;ZHAO Yingxin(School of Computing,Beijing Information Science&Technology University,Beijing,China,100020)

机构地区：[1]北京信息科技大学计算机学院,北京100020

出　　处：《福建电脑》2021年第3期1-4,共4页Journal of Fujian Computer

基　　金：北京信息科技大学2020年大学生创新创业训练计划项目(No.5102010805)资助。

摘　　要：在目前的电视台采访和录音中,有大量的文本任务需要使用语音识别软件进行从语音向文字的转换。如今语音识别的准确率虽然已经足够出色,但对于电视台等严谨的专业领域效果一般,其结果还不能完全信任。由于缺少自动有效地对识别结果进行校对的方法,电视台需要花费大量的人力和物力进行人工校对。因此,本文希望设计并开发一个录音采访文字校对软件来解决此问题。该软件开发的主要工作是构建通用领域和专业领域的语言模型、融合基于统计方法的N-Gram模型和基于特征与学习的Seq2Seq模型相结合的查错纠错算法、构建新闻播报和电视台录音采访等专业领域的查错规则库。In TV interview and recording, a large number of text tasks need to use speech recognition software to convert from voice to text. Although the accuracy of speech recognition is high enough, the effect is still not ideal for professional fields such as TV station. Due to the lack of automatic and effective proofreading method, TV stations need to spend a lot of manpower and material resources for manual proofreading. In order to solve these problems, this paper designs and develops a text proofreading software. The software constructs the language model of general domain and professional domain, the error checking and error correcting algorithm of N-gram model and seq2 seq model, and the error checking rule base of professional domain.

关 键 词：新闻播报 电视台录音采访 通用领域 专业领域 N-GRAM模型 Seq2Seq模型 机器学习 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]