一种区分性互补系统构造与融合的语音识别方法  

A speech recognition method based on complementary system generation and combination

作　　者：陈斌[1] 陈琦[1] 张连海[1] 屈丹[1] 李弼程[1] 

机构地区：[1]解放军信息工程大学信息系统工程学院,郑州450001

出　　处：《声学学报》2016年第1期125-134,共10页Acta Acustica

基　　金：国家自然科学基金(61175017);国家高技术研究发展计划(863)(2012AA011603)资助

摘　　要：在区分性训练的框架下,提出了一种基于混淆信息加权的互补系统构造方法。首先通过统计音素对的混淆信息,利用混淆信息给音素对加以不同的惩罚权重,分别以基线系统中的3个最优识别结果作为参考,计算混淆信息加权后的音素准确率,同时以正确的标注为参考计算标准的音素准确率。然后通过同时最大化混淆信息加权后的音素准确率和最小化标准音素准确率,构建模型层互补系统,并进一步通过结合RDLT(region-dependent linear transform)特征变换过程构造特征层的互补系统。实验结果表明,与互补最小音素错误准则相比,融合模型层互补系统后识别率提高了0.76%,同时融合特征层和模型层的互补系统识别率提高了1.35%。本方法可以增大互补系统间的差异性,提高系统融合后的识别性能。A complementary system generation method based on confusing information weighting is proposed within the framework of discriminative training. Firstly, each pair of confusing phones is dynamically weighted according to the phone confusion information, and the weighted phone accuracy is calculated by referring to the three best hypothesis paths of the base system. Meanwhile, the standard phone accuracy is obtained using the true transcription as the reference. Then, a model space complementary system is constructed by maximizing the weighted phone accuracy, and by minimizing the standard phone accuracy simultaneously. Furthermore, through combining the model-space complementary system-generating method with the RDLT feature transform process, a feature space complementary system is constructed. Experimental results show that compared with the complementary minimum phone error criterion, the recognition rate is increased by 0.76% by combining the base system with the model space complementary system. The performance gain is increased to 1.35% when combining the base system with both the feature and model space complementary systems. The presented method can enlarge the diversity among the complementary systems and improve the recognition rate of the combined system.

关 键 词：基线系统 构造方法 识别方法 区分性 互补 语音 计算标准 构建模型 

分 类 号：TN912.34[电子电信—通信与信息系统]