基于得分歸一化和系統(tǒng)融合的語音關(guān)鍵詞檢測方法
發(fā)布時間:2018-06-23 13:39
本文選題:關(guān)鍵詞檢測 + 得分歸一化 ; 參考:《數(shù)據(jù)采集與處理》2017年02期
【摘要】:為了有效利用不同關(guān)鍵詞檢測系統(tǒng)的互補性,解決不同系統(tǒng)檢測結(jié)果置信度得分不在同一范圍的問題,提出了一種基于得分規(guī)整和系統(tǒng)融合的語音關(guān)鍵詞檢測方法。首先,為了克服連續(xù)語音識別系統(tǒng)中因剪枝錯誤而引起的關(guān)鍵詞丟失問題,應(yīng)用了關(guān)鍵詞相關(guān)的軟Beam寬度剪枝策略裁剪詞圖;其次,在系統(tǒng)融合前采用得分歸一化方法,使得不同系統(tǒng)關(guān)鍵詞檢測結(jié)果置信度得分在同一范圍;最后,通過系統(tǒng)融合處理將不同系統(tǒng)的關(guān)鍵詞輸出進行整合,得到最終的關(guān)鍵詞檢測結(jié)果。實驗結(jié)果表明,經(jīng)過得分歸一化處理后,關(guān)鍵詞檢測性能的實際查詢詞權(quán)重代價(Actual term-weighted value,ATWV)平均相對提升30%;系統(tǒng)融合后關(guān)鍵詞的檢測性能,相比于得分歸一化處理后的最佳單一系統(tǒng),得到了10%的提升。
[Abstract]:In order to effectively utilize the complementarities of different keyword detection systems and solve the problem that the confidence score of different detection results is not in the same range, a speech keyword detection method based on score regularity and system fusion is proposed. First of all, in order to overcome the loss of keywords caused by pruning errors in continuous speech recognition system, the keyword related soft Beam width pruning strategy is applied to clipping the word map. Secondly, the score normalization method is adopted before the system fusion. The confidence score of different system keyword detection results is in the same range. Finally, the key word output of different systems is integrated by system fusion processing, and the final keyword detection results are obtained. The experimental results show that after the normalized score processing, the actual query word weight cost (Actual term-weighted value) of keyword detection performance is improved by 30% on average, and the detection performance of keywords after system fusion is improved by 30%. Compared with the best single system after normalized score processing, a 10% improvement was obtained.
【作者單位】: 解放軍信息工程大學(xué)信息系統(tǒng)工程學(xué)院;
【基金】:國家自然科學(xué)基金(61175017)資助項目
【分類號】:TN912.34
【相似文獻】
相關(guān)期刊論文 前4條
1 劉鋒;徐會法;陶然;;分數(shù)階Fourier變換中量綱歸一化因子的選取[J];系統(tǒng)工程與電子技術(shù);2011年02期
2 趙興浩,鄧兵,陶然;分數(shù)階傅里葉變換數(shù)值計算中的量綱歸一化[J];北京理工大學(xué)學(xué)報;2005年04期
3 黃志勇;張恒;朱佳佳;張曼;;WCDMA無線負荷歸一化方法探討[J];郵電設(shè)計技術(shù);2012年12期
4 ;[J];;年期
相關(guān)會議論文 前2條
1 陳益明;;混響背景下混合塊歸一化匹配濾波算法[A];2013中國西部聲學(xué)學(xué)術(shù)交流會論文集(下)[C];2013年
2 黃志勇;張恒;朱佳佳;張曼;;WCDMA無線負荷歸一化方法探討[A];2012全國無線及移動通信學(xué)術(shù)大會論文集(下)[C];2012年
,本文編號:2057296
本文鏈接:http://sikaile.net/kejilunwen/xinxigongchenglunwen/2057296.html
最近更新
教材專著