摘要
构音障碍影响儿童的语言表达能力,对其社会交往和心理健康产生不利影响。因此研究基于高斯混合模型-隐马尔可夫模型,提出构音障碍儿童语言矫治训练系统,以提高构音障碍儿童的发音准确率。研究结果表明,所提出的模型的损失值随迭代次数的增加而逐渐减小,随后逐渐趋于稳定,其最低损失值约为0.5。且在所有类型的构音障碍中,该模型表现出较高的发音准确率,几乎均在80.0%以上。尤其是在功能性构音障碍中表现最为突出,其发音准确率最高,为85.1%。研究表明,所提出的模型能显著提升构音障碍儿童的发音清晰度,且具有较高的准确率,为构音障碍儿童的语言矫治训练提供了一种具有广泛应用前景的技术方法,有助于推动相关领域的技术进步和应用发展。
Dysarthria affects children's language expression ability,and has adverse effects on their social interaction and mental health.Therefore,based on the training system of children with dysarthria to improve the pronunciation accuracy of children with dysarthria.The results show that the loss value of the proposed model decreases with the number of iterations and gradually stabilizes.Its lowest loss value is about 0.5.And in all types of dysarthria,the model showed a high pronunciation accuracy,almost all above 80%.Especially in the functional dysarthria,with the highest pronunciation accuracy of 85%.The research shows that the proposed model can significantly improve the pronunciation clarity of children with dysarthria,and has a high accuracy,which provides a promising technical method for the language correction training of children with dysarthria,and is helpful to promote the technological progress and application development in related fields.
作者
倪娜
NI Na(Xi’an Vocational and Technical College,Xi’an 710077,China)
出处
《自动化与仪器仪表》
2025年第10期258-261,266,共5页
Automation & Instrumentation
