【摘要】 采用生物电阻抗技术从人体上腹部体表提取的电阻抗信号,不但包含了常规的胃蠕动频率特征,而且携带有反映胃动力状况的更深层次的信息。提取并分析这些信息,对胃动力的检测与评价具有重要意义。对20名糜烂性胃炎患者的胃阻抗和胃电信号进行研究,经过小波滤波去噪后,进行多层小波包变换,计算小波能量熵并作为特征向量,采用3层BP神经网络进行模式分类。经一周治疗后,14名患者胃阻抗和胃电信号的小波能量熵值下降,以小波能量熵为特征向量的BP神经网络对治疗前后的识别正确率为80%。结果表明,小波能量熵能够从整体上表征胃动力信号时域和频域能量分布的复杂程度,可为胃肠病患者的疗效评价提供有效的特征描述。更多还原

【Abstract】 The impedance of the epigastric region not only contains frequency characteristic of gastric peristalsis but also includes information reflecting gastric motility at a deeper level.The extraction and analysis of this information are important to detect and evaluate gastric motility.The gastric impedance and electrogastrogram signals from 20 patients with erosive gastritis were collected and denoised using wavelet transform method.The wavelet energy entropy of the signals was proposed as feature vector to a three layered neural network classifier.The experiment results demonstrated that the wavelet energy entropy of most patients decreased after treatment and the recognition rate of the BP neural network was 80%.The wavelet energy entropy could describe the energy distribution complexity of gastric signals in both time and frequency domain and provided effective featured description for therapeutic evaluation of gastroenteropathy patient.更多还原