论文部分内容阅读
根据研究可知,当疾病或者其他情况使得血管发生病变时,血流信号所包含的信息在靠近血管壁的位置会发生明显的变化。对超声多普勒血流信号进行频谱的分析具有非常重要的意义,因为,血流信号的变化可以由其时频分布谱体现出来。在本论文中,分别利用传统的STFT算法和基于随机字典的Matching Pursuit算法(MP算法)对同一血流信号的时频分布谱、平均频率曲线及最大频率曲线进行分析。结果表明:基于随机字典的Matching Pursuit算法和传统的STFT算法进行比较时,其估计出的时频分布谱以及从谱中提取出的曲线都是更加准确和精确的。
According to the study, when the disease or other conditions lead to vascular lesions, the information contained in the blood flow signal in the blood vessel wall near the location of significant changes occur. It is very important to analyze the spectrum of ultrasound Doppler blood flow signal because the change of blood flow signal can be reflected by its time-frequency distribution spectrum. In this thesis, the time-frequency distribution spectrum, average frequency curve and maximum frequency curve of the same blood flow signal are analyzed by using the traditional STFT algorithm and the Matching Pursuit algorithm based on random dictionary (MP algorithm) respectively. The results show that the Matching Pursuit algorithm based on random dictionary is more accurate and accurate than the traditional STFT algorithm in estimating the time-frequency distribution spectrum and the curve extracted from the spectrum.