Machine learning-enhanced cardiac electronic skin system and method thereof
The BreaCARES system addresses the need for real-time, accurate cardiac monitoring by integrating liquid metal circuits and deep neural networks for electrocardiographic signal analysis, providing enhanced cardiac event detection and classification.
Patent Information
- Authority / Receiving Office
- HK · HK
- Patent Type
- Applications
- Current Assignee / Owner
- THE HONG KONG POLYTECHNIC UNIV
- Filing Date
- 2026-05-08
- Publication Date
- 2026-07-17
AI Technical Summary
Existing cardiac monitoring technologies lack real-time, accurate, and wearable solutions for continuous electrocardiographic signal analysis and cardiac event classification.
A machine learning-enhanced cardiac electronic skin system (BreaCARES) with liquid metal circuits, biopotential electrodes, and a deep neural network for processing electrocardiographic signals to provide real-time visual representations and cardiac event classification.
Enables continuous, accurate, and real-time cardiac monitoring through wearable biopotential sensing, enhancing cardiac event detection and classification.
Smart Images
Figure 00000000_0000_ABST
Abstract
Description
Abstract A machine learning-enhanced cardiac electronic skin system comprises a breathable cardiac electronic skin (BreaCARES) device configured to be worn on a user’s skin, an electronic device, and a server. The BreaCARES device comprises liquid metal (LM) circuits comprising biopotential electrodes for contacting the user’s skin, a biopotential sensing chip configured to sample electrocardiographic (ECG) signals at a predetermined frequency via the biopotential electrodes, and a microcontroller unit (MCU) communicatively coupled to the biopotential sensing chip for receiving the ECG signals from the biopotential sensing chip. The electronic device is configured to receive the ECG signals from the BreaCARES device and generate real-time visual representations of the ECG signals. The server is configured to process the ECG signals using a deep neural network (DNN) trained for cardiac event classification.
Claims
1. A machine learning-enhanced cardiac electronic skin system, comprising: A breathable cardiac electronic skin BreaCARES device, the BreaCARES device being configured to be worn on a user's skin, the BreaCARES device comprising: A liquid metal LM circuit, the LM circuit including biopotential electrodes for contact with a user's skin; A biopotential sensing chip, the biopotential sensing chip being configured to sample electrocardiogram (ECG) signals through the biopotential electrodes; A microcontroller unit (MCU) is communicatively connected to the biopotential sensing chip and is used to receive the ECG signal from the biopotential sensing chip. An electronic device configured to receive the ECG signal from the BreaCARES device and generate a real-time visual representation of the ECG signal; and The server is configured to process the ECG signal using a deep neural network (DNN) trained for cardiac event classification.
2. The system according to claim 1, wherein, The server is configured to perform a short-time Fourier transform (STFT) on the ECG signal to convert the ECG signal from a time-domain signal into a time-spectrum graph.
3. The system according to claim 2, wherein, The server is configured to adjust the size of the time-spectrum to obtain a resized output image.
4. The system according to claim 3, wherein, The server is configured to apply data augmentation to the resized output image to obtain an augmented dataset.
5. The system according to claim 4, wherein, The server is configured to input the augmented dataset into the DNN for state classification of cardiac events.
6. The system according to claim 5, wherein, The DNN mentioned is a Convolutional Neural Network (CNN).
7. The system according to claim 6, wherein, The CNN is configured to be trained using a stochastic gradient descent (SGDM) optimizer with momentum.
8. The system according to claim 1, wherein, The electronic device is a smartphone, and the smartphone includes: The BreaCARES application is configured to receive the ECG signal via Bluetooth Low Energy and generate a real-time visualization of the ECG signal; and A display screen configured to display the visual representation.
9. The system according to claim 1, wherein, The biopotential electrode comprises an LM and a hydrogel.
10. The system according to claim 1, wherein, The MCU is an MCU with integrated Bluetooth BLE, which enables the ECG signal to be transmitted to the electronic device via Bluetooth Low Energy.
11. The system according to claim 10, wherein, The BreaCARES device also includes: A first low-dropout regulator (LDO) is configured to power the MCU with integrated BLE. A second LDO, configured to supply power to the biopotential sensing chip; and A crystal, which is electrically connected to the biopotential sensing chip and configured to provide a timing reference for sampling the ECG signal.
12. The system according to claim 1, wherein, The thickness of the BreaCARES device is in the range of 180 μm to 1.1 mm.
13. A method for monitoring a user's cardiac activity, the method comprising: ECG signals associated with the user are obtained through the BreaCARES device, a breathable cardiac electronic skin worn on the user's skin. A real-time visual representation of the ECG signal is generated on the display screen of the electronic device; as well as The ECG signal is processed by a server to classify cardiac events using a deep neural network (DNN).
14. The method according to claim 13, wherein, Processing the ECG signal through the server includes: Perform a short-time Fourier transform (STFT) on the ECG signal received from the BreaCARES device to convert the ECG signal from a time-domain signal into a time-spectrum graph. The size of the time-spectrum image is adjusted to obtain a resized output image; Data augmentation is applied to the resized output image to obtain an augmented dataset; and The augmented dataset is processed using the DNN to classify cardiac events by state.
15. The method according to claim 14, wherein, Performing the STFT includes performing a discrete STFT on the ECG signal with a window length of 36.
16. The method of claim 14, wherein, Adjusting the size of the time spectrogram includes adjusting the size of the time spectrogram to generate a size-adjusted output image with a resolution of 224×224 pixels.
17. The method according to claim 16, wherein, Application data enhancement includes: Add salt and pepper noise with a noise density of 15% to the resized output image; and The resized output image is randomly scaled along the X-axis to a size between 90% and 110% of its original size to simulate variations in the ECG signal.
18. The method according to claim 14, wherein, Processing the augmented dataset using the DNN includes using a convolutional neural network (CNN) for state classification.
19. The method according to claim 18, wherein, Processing the augmented dataset includes dividing it into a first part for model training and a second part for model validation.
20. The method of claim 18, further comprising training the CNN using a stochastic gradient descent (SGDM) optimizer with momentum.