Blood pressure monitoring device and method

Abstract

This invention provides a blood pressure monitoring device. The blood pressure monitoring device may include an electrocardiogram (ECG) signal sensor, an photoplethysmography (PPG) sensor, and a processor. The ECG signal sensor is used to acquire an ECG signal. The PPG sensor is used to acquire an PPG signal. The processor is coupled to the ECG signal sensor and the PPG sensor. The processor can input the ECG signal and the PPG signal into a complex deep learning model to generate complex predicted blood pressure values, and input the complex predicted blood pressure values ​​generated by the complex deep learning model into a regression model to generate a blood pressure prediction result.

Claims

1. A blood pressure monitoring device, comprising: An electrocardiogram (ECG) signal sensor is used to acquire an ECG signal; A photovolume change mapping sensor is used to acquire a photovolume change mapping signal; A processor is coupled to the aforementioned electrocardiogram (ECG) signal sensor and the aforementioned photoplethysmography (PPG) sensor; and a blood pressure monitor is coupled to the processor and measures a blood pressure value at an initial time point within a time period. The processor inputs the ECG signal and the PPG signal into a complex deep learning model to generate complex predicted blood pressure values, and inputs the complex predicted blood pressure values ​​generated by the complex deep learning model into a regression model to generate a blood pressure prediction result. The processor further inputs the blood pressure value into the regression model to train the regression model. Within the aforementioned time period, the processor continuously generates the blood pressure prediction result using the trained regression model. The processor determines whether an abnormality has occurred based on the blood pressure value and the blood pressure prediction result. If an abnormality is detected within the aforementioned time period, the blood pressure monitor measures a new blood pressure value to update the regression model. The processor continues to generate the blood pressure prediction result within the aforementioned time period using the updated regression model. In the next time period, the blood pressure meter measures a new blood pressure value to update the regression model, and the processor continues to generate the blood pressure prediction result in the next time period using the updated regression model.

2. The blood pressure monitoring device of claim 1, wherein the processor further inputs physical feature information to at least one of the complex deep learning models, and the processing uses the at least one of the complex deep learning models to generate the predicted blood pressure value corresponding to the at least one of the complex deep learning models based on the electrocardiogram signal, the photoplethysmography signal and the physical feature information.

3. The blood pressure monitoring device of claim 1, wherein the processor determines whether the difference between the blood pressure value and the blood pressure prediction result is greater than a threshold value, and wherein when the difference value is greater than the threshold value, the processor determines that an abnormality has occurred.

4. A blood pressure monitoring method, applicable to a blood pressure monitoring device, comprising: An electrocardiogram (ECG) signal is acquired using an ECG signal sensor in the aforementioned blood pressure monitoring device; an photoplethysmography (PPG) signal is acquired using an PPG sensor in the aforementioned blood pressure monitoring device; the ECG signal and the PPG signal are input into a complex deep learning model using a processor in the aforementioned blood pressure monitoring device to generate a complex predicted blood pressure value; and the complex predicted blood pressure value generated by the complex deep learning model is input into a regression model using the processor to generate a blood pressure prediction result. The method further includes: measuring a blood pressure value at an initial time point within a time period using a sphygmomanometer in the aforementioned blood pressure monitoring device; inputting the blood pressure value into the regression model using the processor to train the regression model; continuously generating the blood pressure prediction result using the trained regression model within a time period; and determining whether an abnormality has occurred based on the blood pressure value and the blood pressure prediction result using the processor. If, during the aforementioned time period, the processor determines that an anomaly has occurred, it measures a new blood pressure value using the aforementioned blood pressure meter to update the aforementioned regression model; the processor continues to generate the aforementioned blood pressure prediction results using the updated regression model during the aforementioned time period; in the next time period, it measures a new blood pressure value using the aforementioned blood pressure meter to update the aforementioned regression model; and the processor continues to generate the aforementioned blood pressure prediction results using the updated regression model during the next time period.

5. The blood pressure monitoring method as requested in item 4 further includes: Using the processor described above, physical feature information is input to at least one of the complex deep learning models; and using the processor described above, the predicted blood pressure value corresponding to the at least one of the complex deep learning models is generated based on the electrocardiogram signal, the photoplethysmography signal, and the physical feature information.

6. The blood pressure monitoring method as requested in item 4 further includes: The processor determines whether the difference between the blood pressure value and the blood pressure prediction result is greater than a threshold value. If the difference value is greater than the threshold value, the processor determines that an anomaly has occurred.

Images