基于PYNQ的心电监测系统实现

import numpy import scipy.io as scipy import math train_label_normal = '/home/xilinx/jupyter_notebooks/DDC/train_label_normal.mat' test_label_normal = '/home/xilinx/jupyter_notebooks/DDC/test_label_normal.mat' test_normal = '/home/xilinx/jupyter_notebooks/DDC/test_normal.mat' train_normal = '/home/xilinx/jupyter_notebooks/DDC/train.mat' test = '/home/xilinx/jupyter_notebooks/DDC/test.mat' test_label = '/home/xilinx/jupyter_notebooks/DDC/test_label.mat' train = '/home/xilinx/jupyter_notebooks/DDC/train.mat' train_label = '/home/xilinx/jupyter_notebooks/DDC/train_label.mat' train_label_normal = scipy.loadmat(train_label_normal) test_label_normal = scipy.loadmat(test_label_normal) test_normal = scipy.loadmat(test_normal) train_normal = scipy.loadmat(train_normal) test = scipy.loadmat(test) test_label = scipy.loadmat(test_label) train = scipy.loadmat(train) train_label = scipy.loadmat(train_label) train_label_normal = train_label_normal['train_label_normal'] test_label_normal = test_label_normal['test_label_normal'] test_normal = test_normal['test_normal'] train_normal = train_normal['train'] test = test['test'] test_label = test_label['test_label'] train = train['train'] train_label = train_label['train_label'] for itr in train_normal: itr = itr[0:292] for itr in test_normal: itr = itr[0:292] train = list(train) train_label = list(train_label) train_normal = list(train_normal) train_label_normal = list(train_label_normal) test = list(test) test_label = list(test_label) test_normal = list(test_normal) test_label_normal = list(test_label_normal) train.extend(train_normal) train_label.extend(train_label_normal) test.extend(test_normal) test_label.extend(test_label_normal) syn0 = 2*numpy.random.random((293,10)) - 1 syn1 = 2*numpy.random.random((10,1)) - 1 for j in range(60000): l1 = 1/(1+numpy.exp(-(numpy.dot(train,syn0)))) l2 = 1/(1+numpy.exp(-(numpy.dot(l1,syn1)))) l2_delta = (train_label - l2)*(l2*(1-l2)) l1_delta = l2_delta.dot(syn1.T) * (l1 * (1-l1)) syn1 += l1.T.dot(l2_delta) syn0 += X.T.dot(l1_delta)