Atrial fibrillation signal identification method and system based on machine learning
A signal recognition and machine learning technology, applied in character and pattern recognition, instruments, sensors, etc., can solve problems such as lack of good generalization ability, and achieve high robustness, generalization ability, and accurate recognition.
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0045] This embodiment provides a method for identifying atrial fibrillation signals based on machine learning, such as figure 1 shown, including the following steps:
[0046] Step 1: Obtain multiple ECG waveform signals as training samples, and convert the multiple ECG waveform signals into multiple spectrum images.
[0047] The 1-D ECG waveform was converted into a 2-D time-frequency image using the Modified Frequency Slicing Wavelet Transform (MFSWT). The modified frequency slice wavelet transform (MFSWT) is improved on the basis of the frequency slice wavelet transform. This method is more convenient to adjust the observation time and frequency, and can accurately locate the R wave, P wave, and T wave information in the ECG.
[0048] The Fourier transform of the signal f(t) is Then the frequency domain of the MFSWT model can be expressed as:
[0049]
[0050] Among them, t and ω are the set observation time and frequency respectively, "*" indicates the conjugate ope...
Embodiment 2
[0066] The purpose of this embodiment is to provide a system for identifying atrial fibrillation signals based on machine learning. The system includes:
[0067] The data acquisition module acquires a plurality of ECG waveform signals as training samples;
[0068] A data preprocessing module, converting the plurality of ECG waveform signals into a plurality of frequency spectrum images;
[0069] The model training module adopts the plurality of spectral images to train the atrial fibrillation / non-atrial fibrillation classification model based on a deep convolutional neural network to obtain model parameters;
[0070] Using the plurality of spectral images as the input of the model to obtain the corresponding prediction probability of each spectral image;
[0071] Combine the obtained multiple predicted probabilities to obtain a one-dimensional feature;
[0072] Carry out R-wave detection on the ECG signal and extract the RR interval, and calculate multiple RR interval featu...
Embodiment 3
[0076] The purpose of this embodiment is to provide a computer-readable storage medium.
[0077] A computer-readable storage medium, on which a computer program is stored for calculating the similarity of fingerprints. When the program is executed by a processor, the following steps are performed:
[0078] Obtaining multiple ECG waveform signals as training samples, converting the multiple ECG waveform signals into multiple spectrum images;
[0079] Using the plurality of spectral images, training an atrial fibrillation / non-atrial fibrillation classification model based on a deep convolutional neural network to obtain model parameters; then using the plurality of spectral images as input to the model to obtain the corresponding prediction probability of each spectral image;
[0080] Combine the obtained multiple predicted probabilities to obtain a one-dimensional feature;
[0081] Carry out R-wave detection on the ECG signal and extract the RR interval, and calculate multiple...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More