TWI608824B - Wearable ECG measuring device - Google Patents

Description

Wearable ECG measuring instrument

The invention relates to a measuring device, in particular to a wearable electrocardiograph.

According to the heart, it can be said that it is one of the most important and complicated ones in human organs. With the development of science and technology, it is possible to observe the operation and beating of the heart through the instrument, so as to judge whether the heart condition is abnormal, and then measure the heart. The electrocardiogram produced by the beating is one of the main judgments. In order to accurately capture the activity of the heart, the system currently uses a twelve lead system, 12 leads ECG, which uses twelve leads on the front and horizontal planes to record twelve different hearts. The direction of the electrical and physiological activities, and get 12 pieces of ECG data, so that doctors can observe the heart pulse pulse running from twelve different angles to determine the heart activity or to determine the cause of heart disease.

However, due to the long-term observation and record of ECG data acquisition, and the need to measure the heartbeat condition of twelve different directions at the same time, the overall amount of data is quite large, which occupies a lot of hard disk space. Therefore, how to put the electrocardiogram After the data is compressed and decompressed, the data is still less distorted (also known as quality assurance (PRD), which is used as an indicator of distortion. Generally, the value of PRD is preferably between 2% and 7%) Having a better amount of compression becomes an important issue.

Please refer to the third figure, which is the general step of compressing the ECG data. Before the measurement, the PRD value (target value) is preset, and the compressed PRD value must meet the preset value. After the steps of frequency band decomposition, quantization, error control, and encoding, the ECG data must be subjected to decoding analysis, inverse quantization, and inverse frequency band decomposition, and the compressed data is compared with the original data to determine whether the PRD meets the target value. The positive and negative five percent change range. Also, since the decoding needs to be subjected to decoding analysis, inverse quantization, and inverse band decomposition, the PRD value can be compared, which causes a lot of time for data compression and increases the load on the processor.

At the same time, because the acquisition of ECG data requires a large amount of time to continuously observe the activity of the heart, the patient has to stay in the hospital for observation for a long time, causing considerable inconvenience. Accordingly, the creator of the present invention believes that there should be a means to obtain ECG data without affecting the living conditions of the patient, and at the same time, when the ECG data is subjected to frequency band decomposition, quantization, error control, coding, etc., By performing a PRD comparison, it will save about half of the data compression time to reduce the load on the instrument.

In view of the deficiencies described in the prior art, the present creator proposes a solution for a wearable electrocardiograph, comprising:

A heart sensing unit:

A cardiac sensing unit can be disposed on the outer side of the human skin for sensing the activity state of the human heart, and obtaining an electrocardiogram data.

A host:

The host is available for the user to carry around, and the host includes:

A data storage unit:

The data storage unit can store ECG compressed data.

One processor:

The heart sensing unit and the data storage unit are respectively electrically connected to the processor, and the processor writes a first to third electrocardiogram compression program and a distortion preset value, and the first electrocardiogram compression program is executable by the processor: The sensing data is decomposed into frequency bands to obtain complex frequency band data; the second electrocardiogram processing program is executable by the processor: the quantized scale is used to quantize each frequency band data to obtain a multi-quantization result, and each quantized result is obtained The integer part is obtained by multi-quantization data, and each quantized result is obtained as a fractional part to obtain a complex truncation error data, and the quantization scale is respectively accumulated with the product of each truncation error data to obtain a distortion value, if the distortion value is less than or equal to the distortion preset When the value, or the distortion value satisfies a distortion preset range value, the processor executes the third electrocardiogram compression program; if the distortion value is greater than the distortion preset value, or the distortion value does not satisfy a distortion preset In the range value, the value represented by each quantization scale is changed to re-execute the second electrocardiogram compression program; the third electrocardiogram Reduction program for the processor to perform: encoding the quantized data to obtain a data compressed ECG, and the ECG compressed data transmitted to the data storage unit for storage.

The invention is mainly carried by the host, so that the patient can record the state of cardiac activity anytime and anywhere without having to stay in the hospital for a long time, so as to provide electrocardiogram information for the doctor to diagnose. At the same time, the processor passes the second electrocardiogram compression program to make the electrocardiogram data in the process of compression, and only uses an integer operation to pass the quantization scale and the truncation error data formed in the quantization process as a judgment. Whether the distortion value satisfies the preset value (academic ECG distortion is often measured by PRD. Generally, the value of PRD is preferably between 2% and 7%), so compared with traditional ECG data. The compression program has less steps of decoding analysis, inverse quantization, and anti-band decomposition, and the time required for the processor to be operated is greatly reduced, thereby reducing the load of the processor, thereby saving power.

In addition, according to the broad invention of the same patent law, the invention further proposes a compression method for an electrocardiogram, the compression method of the electrocardiogram comprising the following steps:

(A) Taking an electrocardiogram data for band decomposition, and obtaining complex frequency band data.

(B) taking the quantitative quantization scale to quantize the data of each frequency band to obtain the complex quantization result, and taking the quantized result as an integer part to obtain the multi-quantization data, taking the decimal part to obtain the complex truncation error data, and then performing error control. The quantization scale is respectively accumulated with the product of the truncated error data to obtain a distortion value. If the distortion value satisfies a distortion preset range value, or the distortion value is less than or equal to a distortion preset value, then the step is performed ( C); if the distortion value does not satisfy the distortion preset range value, or the distortion value is greater than a distortion preset value, then changing the value represented by each quantization scale and performing step (B) again.

(C) Encoding the quantized data to obtain an electrocardiogram compressed data.

(1)‧‧‧Heart sensing unit

(2) ‧‧‧Host

(21)‧‧‧Data storage unit

(22) ‧‧‧ Processor

(23)‧‧‧Antenna unit

The first figure is a schematic diagram of the execution steps of each electrocardiogram compression program created by the present invention.

The second figure is a schematic diagram of the connection of each component of the creation of the present invention.

The third figure is a schematic diagram of the execution steps of the conventional ECG data compression.

The construction, features and embodiments of the present invention will be described with the aid of the drawings, and the reviewers will have a better understanding of the present invention.

Referring to the first figure, the invention relates to a wearable electrocardiograph, comprising:

A heart sensing unit (1):

Referring to the second figure, the cardiac sensing unit (1) can be externally measured on the human skin to sense the activity state of the human heart, and an electrocardiogram data is obtained. Since the heart is also divided into a left ventricle, a left and right atrium, etc., in order to know the overall activity state of the heart, the cardiac sensing unit (1) preferably includes twelve sub sensing units, and each sub sensing unit is respectively disposed on the body. Different parts, such as arms, ribs, etc., observe the heart activity in various directions and angles, and the results measured by each sub-sensing unit together form the electrocardiogram data described in the present specification.

One host (2):

Referring to the second figure, the volume of the host (2) is portable for the user, and the host (2) further includes a data storage unit (21) and a processor (22): the data storage unit (21): Please refer to the second figure, the data storage unit (21) is suitable for storing ECG compressed data.

The processor (22)

Referring to the first figure and the second figure, the processor (22) is electrically connected to the cardiac sensing unit (1) and the data storage unit (21). The processor (22) writes a distortion preset value, and a first to third electrocardiogram compression program, the distortion preset The value can be set by electrically connecting an input unit to the processor (22), and can also be a basic setting when the host is manufactured. Among them, the academic quality assurance refers to the degree of distortion of the ECG within the range of 5% or more of the preset value. Generally speaking, in order to take into account the degree of distortion of the ECG and the size of the storage capacity, the preset value of the distortion is better. It is 2% to 7%.

The following is an introduction to the operation of each ECG compression program:

First, the first electrocardiogram compression program is executable by the processor (22) for band decomposition: the decomposition operation is performed by an integer operation, and the sensing data is band-decomposed to obtain a plurality of frequency band data, and the plurality of frequency band data is obtained. The system includes a first-order band data and a second-order band data. Generally, the plurality of band data preferably includes first-to-first-order band data to provide better compression performance of the electrocardiogram.

The frequency band decomposition may be performed by using a recursive wavelet transform method. First, the sensing data is multiplied by a first-order conversion matrix to obtain the first-order frequency band data, and then the first-order frequency band data is multiplied by one. The second-order conversion matrix obtains the second-order band data, and so on, and the first to eleventh band data can be obtained. However, when the wavelet transform method is used for frequency band decomposition, since the fractional part of the data of each order band is not deleted, the data of each order band has a higher bit quantity, and in particular, the data of each order band is analyzed sequentially, which will result in data of each order band. The amount of bits is continuously accumulated, so that the data of the band has a large amount of bits; or the recursive wavelet transform method deletes the fractional part of the data of each frequency band, and the fractional part forms an error amount, and sequentially analyzes each frequency band. After the data, the error amount of the data of each frequency band is continuously enlarged, and the compression quality of the electrocardiogram is difficult to control.

Therefore, the creation of the present invention can be implemented as follows: the frequency band data includes a first order band data and a second order band data; the processor (22) writes a filter matrix data, and the first electrocardiogram is compressed. The program is executable by the processor: multiplying the sensing data by the filter matrix to obtain the first-order band data and the second-order band data. The product of the sensing data and the filter matrix is not limited to obtaining the first-order band data and the second-order band data at the same time. Taking the creation of the present invention as an example, the product of the sensing data and the filter matrix is compared. The best system can obtain the first-order band data to the eleventh-order band data at the same time.

The invention is preferably based on the traditional recursive wavelet transform, and integrates the various order transformation matrices required in the traditional recursive wavelet transform to obtain the data of each order band to obtain a filter matrix, and thus the invention is simultaneously created. A non-returning wavelet transform is proposed, which can be implemented by: the processor (22) writing a filter matrix program, wherein the filter matrix program is available to the processor (22) according to the first conversion matrix, the second conversion The product of the matrix yields the filter matrix. For example, the first-order conversion matrix is, and the second-order conversion matrix is, and when the first-order band data and the second-order band data are to be obtained, the processor extracts the low-frequency one of the first-order conversion matrix. a matrix: and a matrix belonging to a high frequency in the second-order conversion matrix: the filter matrix is obtained by multiplying the two matrices, and then, when the sensing data is multiplied by the filter matrix, The first-order band data and the second-order band data are obtained at the same time. Similarly, if the first-order band data is to be obtained simultaneously to the tenth-order band data, or even any of the band data, the above-mentioned means can be used to obtain the filter matrix, and then the sensing data is multiplied by the The filter matrix is OK.

Therefore, the above preferred embodiment compares the data of each frequency band with respect to the recursive wavelet transform. The amount of bits is small, and the data of each frequency band is quantized to reduce the overall distortion rate. Moreover, the calculation amount of the processor (22) is also reduced, thereby improving the efficiency and power saving.

Furthermore, the second electrocardiogram compression program is executable by the processor (22) for quantization and error control: the quantized scale is used to quantize the data of each frequency band, and the multi-quantization result is obtained, and each quantized result is taken separately. The integer part is obtained by multi-quantization data, and each quantized result is taken as a fractional part to obtain a complex truncation error data. And accumulating the quantized scales and the products of the truncated error data respectively to obtain a distortion value. If the distortion value is less than or equal to the distortion preset value, or the distortion value satisfies a distortion preset range value, then the The processor (22) executes the third electrocardiogram compression program, and if the distortion value is greater than the distortion preset value, or the distortion value does not satisfy a distortion preset range value, the processor (22) changes each quantization scale. Represent the value and re-execute the second ECG compression program.

Wherein, each frequency band data corresponds to a quantization scale, for example, the first-order frequency band data corresponds to a first-order quantization scale, and the remaining various frequency band data and each order quantization scale are deduced by analogy. If the processor (22) executes the second electrocardiogram compression program and respectively changes the value represented by each quantization scale, so that the distortion value is less than or equal to the distortion preset value, or the distortion value satisfies a distortion preset range value. In the process, because the processor (22) has to continuously change the value represented by each quantization scale to achieve the above purpose, it is easy to cause the processor (22) to spend a lot of time. For this reason, the processor (22) is more Further, a quantization scale program is provided, and the quantization scale program is executable by the processor (22): each quantization scale corresponds to a one-ary quadratic equation, and a quantization scale coefficient is obtained by using each quadratic equation to obtain each quantization scale. The value represented.

For example, the first-order quadratic relationship between the first-order quantization scale (Q ₁ ), the second-order quantization scale (Q ₂ ), and the quantized scale coefficient (Q _F ) is:, and, the processor (22) The first-order quantization scale and the second-order quantization scale can be obtained simultaneously through only one quantized scale coefficient, and the other scale quantization scales are deduced by analogy. Therefore, when the distortion value is greater than the distortion preset value, or the distortion value does not satisfy a distortion preset range value, the processor (22) may change only the value of the quantization scale coefficient to obtain each order quantization. The scale, by which the processor (22) has better processing efficiency when changing the quantization scales.

Then, the third electrocardiogram compression program is executable by the processor (22): encoding the quantized data to obtain an electrocardiogram compressed data, and transmitting the electrocardiogram compressed data to the data storage unit (21) Save it.

Therefore, the following advantages can be achieved through the creation of the present invention:

(1) Easy to use:

Through the size of the host, patients can carry it with them. Therefore, patients can measure ECG anytime and anywhere without waiting in the hospital.

(2) Shorten the compression time:

When the second ECG compression program is used for quantization and error control, the quantization scale and the complex truncation error data obtained by the integer operation are mainly performed, and the second ECG compression program performs the PRD test while encoding. And accumulating the distortion value obtained by multiplying the quantization scale by the product of each truncation error data, and checking whether the distortion value satisfies the distortion preset value, thereby ensuring that the electrocardiogram compression data has better quality assurance (PRD), The steps can also be referred to as quality assurance mechanisms. Therefore, the present invention only needs to perform the frequency band decomposition, the quantization and the error control. The system and the code can obtain the electrocardiogram compression data, compared with the conventional ECG compression method (please refer to the third figure), the whole process needs to adopt floating point calculation, and also needs to perform inverse frequency band decomposition. Anti-quantization, decoding analysis, etc., is less than half the time. Therefore, the processor (22) created by the invention has small load capacity and fast execution speed, thereby further improving the energy-saving use effect.

(3) Using only one quantized scale factor, the quantized scales of each order can be obtained to facilitate error control:

According to the previous error control, the data of each frequency band are independently adjusted, so that each frequency band has an independent quantization scale. Once the quantization result of each frequency band needs to be adjusted, each quantization scale must be separately modulated to make the overall error control. More complicated and inefficient. Through the quantization scale program, the present invention can obtain the quantization scales of each order by using only a single quantization scale coefficient, thereby enabling the processor (22) to have better processing efficiency when changing the quantization scales. In addition, the present invention creates a first electrocardiogram program based on non-recursive wavelet transform, so that the PRD and the quantized scale have a linear relationship, and therefore, when error control is performed, the quantization scale can be quickly adjusted. Meet the PRD settings to improve overall operational efficiency.

In order to allow the physician to know the heart activity of the patient at any time, the invention can also be implemented as follows: the host (2) further comprises an antenna unit (23), wherein the antenna unit (23) can transmit the ECG compressed data to Cloud server.

In addition, please refer to the first figure, according to the broad invention of the same patent law, the invention further proposes a compression method of the electrocardiogram, the compression method of the electrocardiogram comprises the following steps:

(A) Taking an electrocardiogram data for band decomposition, and obtaining complex frequency band data.

(B) Re-quantizing the scale data to quantify the data of each frequency band, and taking the quantized results to obtain the complex quantized data, and taking the fractional part of each quantized result to obtain the complex truncation error data, and then performing error control. And accumulating the quantization scale and the product of each truncation error data to obtain a distortion value, if the distortion value satisfies a distortion preset range value, or the distortion value is less than or equal to a distortion preset value, then proceeding Step (C); if the distortion value does not satisfy the distortion preset range value, or the distortion value is greater than a distortion preset value, then changing the value represented by each quantization scale and performing step (B) again.

(C) Encoding the quantized data to obtain an electrocardiogram compressed data.

In addition, the step (A) is preferably implemented by: the frequency band data includes a first order band data, and a second order band data, and a filter matrix data is taken, and the electrocardiogram data is multiplied by the filter matrix. The data is obtained simultaneously with the first-order band data and the second-order band data.

The invention provides a method for compressing an electrocardiogram, and its advantages, features, and implementation manners can be referred to the present invention for creating a wearable electrocardiogram measuring device, which is not described herein.

In summary, the creation of the present invention is indeed in line with the industrial applicability, and is not found in the publication or public use before the application, nor is it known to the public, and has non-obvious knowledge, conforms to the patentable requirements, and is patented according to law. Application.

However, the above-mentioned statements are a preferred embodiment of the invention in the creation of the invention, and all the changes in the scope of the patent application according to the invention are within the scope of the claim.

Claims (2)

A wearable electrocardiograph, comprising: a cardiac sensing unit: can be arranged on the outer side of the human skin for sensing the activity state of the human heart, and obtaining an electrocardiogram data; a host: for the user to carry around, the The host includes: a data storage unit for storing the electrocardiogram compression data; a processor: electrically connecting the cardiac sensing unit and the data storage unit, wherein the processor writes a first to third electrocardiogram compression program, and a distortion preset value, the first electrocardiogram compression program is executable by the processor: performing frequency band decomposition on the sensing data to obtain a plurality of frequency band data; the second electrocardiogram processing program is executable by the processor: The data of each frequency band is quantized to obtain a multi-quantization result, and each quantized result is obtained as an integer part to obtain a multi-quantization data, and each quantized result is obtained as a fractional part to obtain a complex truncation error data, and the quantization scale is respectively accumulated with the product of each truncation error data. And obtaining a distortion value, if the distortion value is less than or equal to the distortion preset value, or When the true value satisfies a distortion preset range value, the processor executes the third electrocardiogram compression program; if the distortion value is greater than the distortion preset value, or the distortion value does not satisfy a distortion preset range value, then Rechanging the second electrocardiogram compression program by changing the value represented by each quantization scale; the third electrocardiogram compression program is executable by the processor: encoding the quantized data to obtain an electrocardiogram compression data, and transmitting the electrocardiogram compression data to The data storage unit performs storage; wherein the plurality of frequency band data includes a first order band data and a second order band resource The processor writes a filter matrix data, and the first electrocardiogram compression program is executable by the processor: multiplying the sensing data by the filter matrix to obtain the first-order frequency band data, and the second-order frequency band Data; wherein the processor writes a quantization scale program, wherein the quantization scale program is executable by the processing: each quantization scale corresponds to a one-dimensional quadratic equation, and a quantized scale coefficient is obtained by using each quadratic equation to obtain each quantization scale . The wearable electrocardiograph according to claim 1, wherein the host further comprises an antenna unit, wherein the antenna unit is configured to transmit the ECG compressed data to the cloud server.