摘要
通过引入非线性研究方法 ,对超声多普勒血流信号的特征进行综合分析 .利用数学形态学方法提取声谱包络 ,并进行包络特征点的自动识别 .采用综合反映声谱包络形态特征的波形分类决策法和超声多普勒音频信号的分形特征表示法提取信号的新特征 ,其中波形分类决策法采用了非线性的人工神经网络分类器 .这些非线性特征的分析经临床应用 ,均取得了较为显著的效果 ,预期为胎儿宫内生长发育状况的判断和疾病的早期诊断提供更好。
In order to provide better and more sensitive indices for judgement of inter uterine fetal growth state and early disease diagnosis, nonlinear research methods were introduced to analyze the characteristics of Doppler ultrasound blood flow signal. The mathematical morphology method was used to extract the envelope of the spectrogram, and then to recognize automatically the character points of the graph. Two methods were proposed to obtain the new features of the signal. One is the spectrogram envelope classification and decision, in which the artificial neural network method was used. Another is the fractal analysis of Doppler ultrasound audio signals. The clinical application of these nonlinear characteristics analysis methods shows very good performance.
出处
《复旦学报（自然科学版）》
CAS
CSCD
北大核心
2001年第3期268-272,共5页
Journal of Fudan University：Natural Science
基金
教育部优秀年轻教师基金资助项目
上海市青年科技启明星计划资助项目 (97QD140 35 )
