基于数学形态学的病理性附加肺音时频谱图分析  

Spectrogram Analysis of Pathologic Adventitious Respiratory Sounds Based on Mathematical Morphology

作　　者：张柯欣[1,2] 王雪峰[3] 魏巍[3] 龙哲[1] 王春武[4] 赵宏 

机构地区：[1]东北大学中荷生物医学与信息工程学院,辽宁沈阳110819 [2]辽宁中医药大学,辽宁沈阳110847 [3]辽宁中医药大学附属医院,辽宁沈阳110032 [4]吉林师范大学,吉林四平136000

出　　处：《中华中医药学刊》2016年第7期1550-1552,I0001,共4页Chinese Archives of Traditional Chinese Medicine

基　　金：国家自然科学基金项目(81273800)

摘　　要：对病理性附加肺音进行即时识别和分析,是监控呼吸系统病情和不断改进治疗方法的十分必要的手段。提出了典型病理性附加肺音进行STFT和小波变换后的特征频谱图像模型,并根据这一模型设计出病理性附加肺音识别方法。这一方法与国际上各种基于学习或统计的病理性附加肺音识别方法比较,具有直观的优势和鲁棒的特点。The immediate recognition and analysis of pathological adventitious respiratory sounds are very necessary means for monitoring and treatment improvement of respiratory disease. In this paper,a new characteristic spectral image model of typical pathological with lung sounds after STFT and Wavelet transformations has been proposed. And a new recognition method of pathological with lung sounds based on this model has been designed. Compared to the varieties of learning and statistical methods for the typical pathological with lung sounds in the field,this method is more intuitive and robust.

关 键 词：肺音 水泡音 喘鸣音 数学形态学 小波 时频谱图 

分 类 号：R443.4[医药卫生—诊断学]