项目名称: 面向连续语音的哈萨克语关键词识别技术研究
项目编号: No.61462084
项目类型: 地区科学基金项目
立项/批准年度: 2015
项目学科: 计算机科学学科
项目作者: 达吾勒·阿布都哈依尔
作者单位: 新疆大学
项目金额: 45万元
中文摘要: 本项目根据国家丝绸之路经济带战略构想及新疆信息化建设的迫切需要,研究面向新疆及中亚地区信息领域的面向连续语音的哈萨克语关键词识别关键技术。面向网络和手机短信语料,创建哈萨克语电话、手机、互联网以及口语对话语音语料库,提取并分析哈萨克语口语语音特征参数、噪音消除技术、特征提取方法、研究基于连续语音识别技术的哈萨克语关键词检索技术,搭建基于网络及通讯设备的哈萨克语关键词检索系统。该项成果不仅对哈萨克语语音文档内容进行情报搜集等提供强大的技术支撑,并且将来会在新疆和中亚地区创造深远的社会及经济价值。
中文关键词: 哈萨克语;关键词识别;特征提取;语音语料库
英文摘要: According to the Strategic Conception of the Silk Road Economic Belt and the urgent need of information construction of Xinjiang, we will research the key technologies of Kazakh Continuous Speech Keyword Spotting for Xinjiang and Central Asia region . Create a network and phone short message based Kazakh language speech corpus, which will be collected from phone, mobile and web. Extract and analyze the Kazakh oral speech feature parameters, the noise cancellation technologies, feature extraction methods, study continuous speech Kazakh keywords retrieval technology, construct network and communication equipments based Kazakh keyword retrieval system. This project not only provide a strong technical support for Intelligence collection from Kazakh speech document, but also create great social and economic value for the regions of Xinjiang and Central Asia in the future.
英文关键词: Kazakh;Keyword Spotting;Feature Extraction;Speech Corpus