基于语音流畅度与模糊聚类的精神分裂症自动识别

针对精神分裂症诊断周期长、缺乏客观诊断依据的问题,提出语音流畅度矩形参数过能熵积,结合模糊聚类双重定位停顿区域的算法辅助诊断精神分裂症.该算法综合了精神分裂症患者语音流畅度低及能量平缓的特征,定位精神分裂症语音的停顿区域,提出语音流畅度量化参数提取算法,结合SVM(支持向量机)分类器,实现精神分裂症的自动识别.提取28例精神分裂症患者和28例正常对照组语音的流畅度声学特征,自动识别精神分裂症,正确率为85％以上.提出的基于过能熵积与模糊聚类的精神分裂症自动识别算法,能为临床诊断精神分裂症提供客观、有效、无创的辅助依据.

Automatic recognition of schizophrenia based on speech fluency and fuzzy clustering

Clinically, the diagnosis of schizophrenia has the problems of long diagnosis cycle and lack of objective diagnosis basis. This paper proposed a new algorithm. Combined with the characteristics of low fluency and gentle energy of schizophrenic speech, the algorithm proposed the speech fluency rectangle parameter. Then the use of fuzzy clustering based two-step location algorithm could locate the pause area of schizophrenia speech. Finally, with SVM(support vector machine) classifier, the algorithm could realize the recognition of schizophrenic speech automatically. This paper tested the speech samples collected by 28 schizophrenic patients and 28 normal controls. The automatic detection accuracy of schizophrenia is over 85%. The experiment results indicate that this method can provide objective, effective and non-invasive auxiliary basis for clinical diagnosis of schizophrenia.