Method and system for quickly detecting an abnormal concentration of potassium ions in blood from an electrocardiogram

1. A method for quickly detecting an abnormal concentration of potassium ions in blood from an electrocardiogram comprising the steps of:

building at least model data (MD) by using at least one reference ECG to provide a corresponding reference concentration of potassium in blood (KR), wherein the reference concentration of potassium ions in blood is in a selected range of K1 to K2, with an interval unit of q, wherein the selected range of the reference concentration of potassium ions in blood is converted into a binary sequence that comprises r digits of “0” and “1”, wherein the reference concentration of potassium in blood (KR) is converted into a binary sequence that comprises i digits of “1” according to first and second equations:

r =( K 2- K 1)/ q   the first equation; and

i =( KR - K 1)/ q   the second equation;

providing an ECG of a monitored person;

converting the ECG of the monitored person into a probability vector that comprises i digits of “1” in a binary sequence that comprises r digits by operating a model in the form of a convolutional neural network based on the reference concentration of potassium ions in blood (KR) of the binary sequence of the model data; and

converting the probability vector into a predicted concentration of potassium ions in blood (KP) by a third equation as follows:

KP=K 1+ qi   the third equation.

2. The method according to claim 1 , wherein the ECG of the monitored person is derived from ECG signals with 12 leads.

3. The method according to claim 1 , further comprising the step of model learning that learns by using a convolutional neural network to identify the eigenvalue of an ECG in an unsupervised manner to produce new MD for diagnose.

4. A system for quickly detecting abnormal concentration of potassium ions in blood from electrocardiogram comprising at least one vector converter comprising a transmitter, at least one memory unit, at least one storage unit and a processor unit connected to the transmitter, the memory unit and the storage unit, wherein the storage unit comprises a model data module and a predicting and converting module, wherein the model data module comprises model data comprising an ECG and a corresponding reference concentration of potassium ions in blood, wherein the predicting and converting module executes the method of claim 1 to convert the ECG of the monitored person into the corresponding predicted concentration of potassium ions in blood.

5. The system according to claim 4 , further comprising at least one ECG device connected to the vector converter and at least one monitor connected to the vector converter so that the vector convertor converts the ECG of the monitored person provided by the ECG device into the predicted concentration of potassium ions in blood, and shows the predicted concentration of potassium ions in blood on the monitor.

6. The system according to claim 5 , wherein the transmitter comprises at least one acquisition interface connected to the ECG device and at least one output interface connected to the monitor.

7. The system according to claim 5 , wherein the vector converter is a wearable physiologic monitor made in one piece with the ECG device.

8. The system according to claim 5 , wherein the ECG device, the vector converter and the monitor are made in one piece.

9. The system according to claim 5 , wherein the monitor comprises a display unit for showing a predicted concentration of potassium ions in blood and an ECG derived from ECG signals.

10. The system according to claim 5 , wherein the monitor further comprises an alerting unit for transmitting the predicted concentration of potassium ions in blood and the ECG derived from the ECG signals to another monitor.

11. The system according to claim 4 , wherein the vector converter further comprises a graphic processor unit connected to the processor unit to increase speed of calculation.

12. The system according to claim 4 , wherein the vector converter further comprises a model-learning module that learns by using a convolutional neural network to identify an ECG in an unsupervised manner.

13. The system according to claim 12 , wherein the model data module, the predicting and converting module and the model-learning module of the vector converter are stored in a same storage unit.

14. The system according to claim 4 , wherein the memory unit comprises a read only memory unit and a random access memory unit for storing programs and commands of the system and for temporarily storing an operation system and any other programs that are executed.

15. The system according to claim 4 , wherein the storage unit is connected to the processor unit via cables or in a wireless manner.

16. The system according to claim 15 , wherein the storage unit is an internal storage device or an external storage device.

17. The system according to claim 16 , wherein the vector converter is an online server.

18. A predicting and converting module for providing a predicted concentration of potassium ions in blood from an ECG, making a computer execute the following tasks:

using a convolutional neural network to build model data that comprises an ECG and a corresponding concentration of potassium ions in blood and to convert the concentration of potassium ions in blood into a reference concentration of potassium ions in blood in the form of a binary sequence;

obtaining an ECG of a monitored person from an ECG device and using the convolutional neural network to derive a probability vector from the ECG;

converting the probability vector into a predicted concentration of potassium ions in blood according to the binary sequence of the reference concentration of potassium ions in blood; and

showing results of analysis of the predicted concentration of potassium ions in blood and using a monitor to control and process the results.

19. The predicting and converting module according to claim 18 , wherein the building of the model data further comprises learning by using a convolutional neural network to identify eigenvalues of the ECG in an unsupervised manner.