融合音素的缅甸语语音识别文本纠错
Text error correction of Burmese speech recognition based on phoneme fusion
陈璐 1董凌 1王文君 1王剑 1余正涛 1高盛祥1
作者信息
- 1. 昆明理工大学信息工程与自动化学院,云南 昆明 650500;昆明理工大学云南省人工智能重点实验室,云南 昆明 650500
- 折叠
摘要
缅甸语语音识别文本中包含大量的同音和空格错误,使用通用的文本语义信息纠正错误字符,对缅甸语空格和同音错误定位和纠正不准确.考虑到缅甸语是一种声调语言,并且音素中包含了声调信息,因此提出融合音素的缅甸语语音识别文本纠错方法.通过参数共享策略对转录文本及其音素进行联合建模,利用音素信息辅助检测并纠正缅甸语同音和空格错误.实验结果表明,本文所提方法相比基线方法ConvSeq2Seq,在缅甸语语音识别纠错任务中的F1值提升了85.97%,达到了79.15%.
Abstract
The Burmese language speech recognition text contains a large number of homophones and space errors.General methods use text semantic information to correct erroneous characters,but they are not accurate in locating and correcting Burmese space and homophone errors.Considering that Bur-mese is a tonal language with tone information embedded within its phonemes,this paper proposes a method for correcting errors in Burmese language speech recognition text that incorporates phonemes.Parameter sharing strategy is used to jointly model the transcribed texts and theirs phonemes,phoneme information is used to assist in detecting and correcting Burmese homophones and space errors.Experi-mental results show that compared with ConvSeq2Seq method,the F1 value of the proposed method in the Burmese speech recognition correction task has increased by 85.97%,reaching 79.15%.
关键词
缅甸语/语音识别文本纠错/音素/共享参数/BERTKey words
Burmese language/speech recognition text correction/phoneme/shared parameter/bidi-rectional encoder representations from transformers(BERT)引用本文复制引用
基金项目
国家自然科学基金(U21B2027)
国家自然科学基金(61972186)
云南省高新技术产业发展项目(201606)
云南省科技重大专项(202103AA080015)
云南省科技重大专项(202302AD080003)
云南省基础研究计划(202001AS070014)
云南省学术和技术带头人后备人才项目(202105AC160018)
出版年
2024
国家科技期刊平台
NSTL
万方数据