KR101977745B1 - Method and apparatus for classifying cardiac arrhythmia - Google Patents

Abstract

Methods and devices for classifying cardiac arrhythmias using electrocardiogram (ECG) signals are provided. The method includes receiving the ECG signal indicative of an electrical activity of an individual's heart for a period of time, the ECG signal comprising an ECG bit. The method further includes converting each of the ECG bits to a symbol. The method further comprises identifying an arrhythmia class indicative of a type of cardiac arrhythmia associated with the individual from arrhythmia classes representing each type of heart arrhythmia based on the symbol representing each of the ECG bits.

Description

[0001] METHOD AND APPARATUS FOR CLASSIFYING CARDIAC ARRHYTHMIA [0002]

Lt; RTI ID = 0.0 > arrhythmia < / RTI >

An electrocardiogram (ECG) system monitors the electrical activity of the patient's heart. Normal electrocardiograms include P waves, QRS complexes, T-waves, and U-waves. ECG analysis is a well-established method for studying cardiac function in patients and for identifying disorders of the heart.

Doctors have been using the ECG system to monitor the cardiac activity of Chinese characters for decades. Currently, there are several other systems that monitor the electrical activity of the patient's heart and analyze the ECG signal to identify the type of cardiac arrhythmia the patient is experiencing. However, these systems are generally fixed and not suitable for portable use.

KR 19980067085 A (1998-10-15)

According to one aspect, there is provided a method of classifying cardiac arrhythmias using an electrocardiogram (ECG) signal, the method comprising receiving the ECG signal indicative of an electrical activity of an individual's heart for a period of time, The ECG signal includes ECG bits. The method further includes converting each of the ECG bits to a symbol. The method further comprises identifying an arrhythmia class indicative of a type of cardiac arrhythmia associated with the individual from arrhythmia classes representing each type of cardiac arrhythmia based on the symbol representing each of the ECG bits .

According to another aspect, there is provided a health monitoring device comprising a communication interface configured to receive an electrocardiogram (ECG) signal indicative of an electrical activity of an individual's heart for a period of time, the ECG signal comprising ECG bits. The health monitoring device further includes an arrhythmia classification module configured to convert each of the ECGs to a symbol. Wherein the arrhythmia classification module further identifies an arrhythmia class indicating the type of cardiac arrhythmia associated with the individual from arrhythmia classes representing each type of cardiac arrhythmia based on the symbol representing each of the ECG bits do.

According to another aspect, there is provided a system comprising an acquisition device configured to generate an electrocardiogram (ECG) signal indicative of an electrical activity of an individual's heart for a period of time, the ECG signal comprising ECG bits. The system further includes a health monitoring device configured to convert each of the ECG bits to a symbol. The health monitoring device is further configured to identify an arrhythmia class indicative of the type of cardiac arrhythmia associated with the individual from classes of arrhythmia indicative of each type of cardiac arrhythmia based on the symbol representing each of the ECG bits.

Other aspects may become apparent from the description, drawings, and claims.

1A is a block diagram illustrating one embodiment of a health monitoring environment in accordance with the present invention.
FIG. 1B is a block diagram illustrating an embodiment of the arrhythmia classification module of FIG. 1A.
2 is a process flowchart illustrating an embodiment of a method of classifying arrhythmias based on an electrocardiogram (ECG).
3 is a flow chart showing another embodiment of a method of classifying an arrhythmia based on an ECG signal.
4 is a flow chart illustrating still another embodiment of a method of classifying arrhythmias based on an ECG signal.
5 is a flow chart illustrating one embodiment of a method of configuring a training bit database that includes a training bit number.
6 is a flow chart illustrating one embodiment of a method for detecting an R-peak at each ECG bit of an ECG signal.
FIG. 7 is a block diagram illustrating one embodiment of a dedicated health monitoring device of FIG. 1A.
8A is a graph illustrating one embodiment of a segment of ECG bits of an ECG signal.
FIG. 8B is a graph illustrating one embodiment of a symbolic aggregate approximation (SAX) symbolization of ECG bits.
Throughout the drawings and the description, unless indicated otherwise, it is to be understood that the same reference numerals denote the same elements, functions, and structures. The relative size and description of these elements may be exaggerated for clarity, illustration, and convenience.

The following detailed description provides apparatus, and / or systems in accordance with some embodiments for purposes of clarity of understanding, but various changes, modifications, and equivalents of the systems, apparatus, and / And can be clearly understood by those skilled in the art to which the present invention belongs. Also, descriptions of well-known functions and structures may be omitted for clarity and brevity.

FIG. 1A is a block diagram illustrating one embodiment of a health monitoring environment 100. FIG. In FIG. 1A, the health monitoring environment includes a dedicated health monitoring device 102 and an acquisition device 108.

The acquisition device 108 includes a device coupled to one or more sensors disposed in the body of an individual (e.g., a patient) suffering from a cardiac disorder.

A dedicated health monitoring device 102 includes an electronic device that is wirelessly coupled to an acquisition device 108 via a wireless network 110. In one embodiment, a dedicated health monitoring device 102 may be provided to the physician to monitor cardiac activity of the patient.

For example, the dedicated health monitoring device 102 may include a smartphone or tablet PC associated with the physician.

In another embodiment, a dedicated health monitoring device 102 may be provided to an attendant associated with the patient. In this embodiment, the dedicated health monitoring device 102 may include a handheld device that is provided to the attest by the hospital personnel.

As shown, the health monitoring device 102 includes an arrhythmia classification module 104, a training beat database 106, and a display unit 105.

The acquisition device 108 periodically collects the cardiac electrical activity data of the individual from the sensor disposed in the body of the individual and generates an ECG signal based on the electrical activity data.

The acquiring device 108 transmits the electrical activity data to the dedicated health monitoring device 102 via the wireless network 110 via the ECG signal.

The ECG signal received at the health monitoring device 102 includes ECG bits, and each ECG bit represents an electrical activity of the heart over an a cardiac cycle. The arrhythmia classification module 104 filters the ECG signal using a bandpass filter and detects an R-peak at each ECG bit of the ECG signal. The arrhythmia classification module 104 segments each ECG bit region around each R-peak and converts the region of each ECG bit into a symbolic sum approximation (SAX) symbol.

The arrhythmia classification module 104 retrieves training bits corresponding to respective arrhythmia classes representing each type of arrhythmia from the training bit database 106.

The arrhythmia class may include a normal, a left bundle brunch block, a right bundle brunch block, an atrial premature beat, and / or a premature ventricular contraction ).

The arrhythmia classification module 104 then identifies the arrhythmia class that indicates the type of arrhythmia associated with the individual by comparing the SAX symbol representing each ECG bit with the SAX symbol representing each training bit. The arrhythmia classifying module 104 displays the arrhythmia class to the user through the display unit 105. [

FIG. 1B is a block diagram illustrating an embodiment of the arrhythmia classifying module 104 of FIG. 1A. 1B, the arrhythmia classification module 104 includes a bandpass filter 152, an R-peak detection module 154, a symbolic aggregate approximation module 156, An arrhythmia class identification module 158, and a display module 160.

A bandpass filter 152 filters the ECG signal representing the electrical activity of the heart of the individual using a high pass filter and a low pass filter. The R-peak detection module 154 detects the location of the R-peak at each ECG bit of the ECG signal.

The symbolic sum approximation module 156 segments the region of each ECG bit around each of the R-peaks (segments of each ECG beat that is around the respective R-peak). The symbolic sum approximation module 156 converts the region of each ECG bit into a SAX symbol.

An arrhythmia class identification module 158 identifies the arrhythmia class indicating the type of cardiac arrhythmia associated with the individual from the arrhythmia class.

In one embodiment, the arrhythmia class identification module 158 identifies the arrhythmia class based on a morphology similarity measure technique, as described in FIG. In another embodiment, an arrhythmia class identification module 158 identifies the arrhythmia class based on a K-mean clustering technique, as described in FIG. The display module 160 displays the arrhythmia class to the user through the display unit 105 of FIG. 1A.

2 is a flow chart illustrating one embodiment of a method 200 for identifying an arrhythmia based on an ECG signal. For example, the method may be performed by an arrhythmia classification module 104 of FIGS. 1A and 1B.

In step 202, the ECG signal indicative of an electrical activity of the heart of the individual over a period of time may be obtained from an ECG signal, such as an acquisition device 108, such as the acquisition device 108 of FIG. ≪ RTI ID = 0.0 & is obtained from an acquisition device. The ECG signal includes ECG bits.

In step 204, the ECG signal indicative of the electrical activity of the heart of the individual for the certain period of time is filtered. For example, the ECG signal may be filtered using a bandpass filter having cutoff frequencies of 0.5 and 40 Hz.

More specifically, the ECG signal can be filtered using a high-pass filter with a cut-off frequency of 0.5 Hz and a low-pass filter with a cut-off frequency of 40 Hz. In step 206, an R-peak corresponding to each ECG bit in the ECG signal is detected. The method of detecting the R-peak is described in more detail in the description of FIG.

In step 208, the area of each of the ECG bits around the detected R-peak in each ECG bit is segmented (a region of said ECG beat is detected) is segmented). The region of each ECG bit is segmented by moving in the backward direction and in the forward direction of the periphery of the R-peak (a segmented by a backward direction and a forward direction around the R-peak). Referring to FIG. 8A, the R-peak detected in each ECG bit is circled and the area of each ECG bit around each detected R-peak is indicated by a line (shown is lines ).

에 의해 w-차원의 공간(여기서 w << n)으로 나타낼 수 있다. 상기

의 i^th 요소는 수학식 1에 기초하여 결정될 수 있다.In step 210, the segmented region of each ECG bit is converted to SAX symbols. SAX is an approximation technique that reduces dimen- sions along the low bound on the Lp norm. The time series T of length n is a vector

Dimensional space (where w << n). remind

Of the i ^th element it may be determined based on equation (1).

To convert the time series T of length n into a string of arbitrary length w, the ECG bits are divided into w equal-sized windows. For each window, the mean value is determined. The vector of the average value of each window is a representation on a reduced scale.

This representation is known as Piecewise Aggregate Approximation (PAA). After conversion of the time series T to the PAA, a further approximation is applied for discretization. This discretization helps to generate symbols with equal probability.

Normalized subsequences may include a highly Gaussian distribution. Therefore, the " breakpoints " that can produce the same areas in the Gaussian curve can be easily determined. The breakpoints (B =? ₁ , ...,? _A-1 ) are lists of ordered numbers, such as areas in N (0, 1), where? ₀ and? _A are -∞ and? This breakpoint B can be determined from the statistical table. For values from 3 to 5, Table 1 shows the breakpoints of? ₁ to? ₄ .

After determining the breakpoint, the discretization of the time series is performed. The PAA coefficients less than the smallest breakpoint are mapped to symbol " a ", and the PAA coefficients greater than or equal to the small breakpoint are mapped to symbol " b " and so on.

8A, eight symbols "b", "c", "d", "e", "f", "g", "h", and "j" may be used to represent the original time series (used to represent an original time series). The concatenation of the SAX symbols representing the sequence is referred to as a word.

In step 212, an arrhythmia class that indicates the type of cardiac arrhythmia associated with the individual from a plurality of arrhythmia classes may be selected, for example, SAX (x, y) representing the training bits stored in the training bit database 106 of FIG. Symbol and the SAX symbol representing each ECG bit. The training bits correspond to respective arrhythmia classes representing each type of arrhythmia.

In step 214, the arrhythmia class, which indicates the type of cardiac arrhythmia associated with the individual, may be displayed on the display unit 105 of a dedicated health monitoring device, such as the display unit 105 of the dedicated health monitoring device 102 of FIG. Lt; / RTI &gt;

3 is a flow chart illustrating another embodiment of a method 300 for classifying arrhythmias based on an ECG signal. For example, the method may be performed by an arrhythmia classification module 104 of FIGS. 1A and 1B.

In step 302, the ECG signal indicative of the cardiac electrical activity of the individual for a period of time is obtained, for example, from an acquisition device, such as the acquisition device 108 of FIG. 1A. The ECG signal includes a plurality of ECG bits. In step 304, the ECG signal representing the electrical activity of the heart of the individual for the certain period of time is filtered. In step 306, the R-peak corresponding to each ECG bit in the ECG signal is detected.

In step 308, the area of each of the ECG bits around the detected R-peak in each ECG bit is segmented. In step 310, the segmented region of each ECG bit is converted to a SAX symbol.

In step 312, a morphology similarity measure is calculated between each of said training bits and said plurality of ECG bits corresponding to said plurality of arrhythmia classes representing each type of arrhythmia.

More specifically, the MINDIST function is used to calculate the morphology similarity measure between each of the ECG bits and the training bit. The MINDIST function returns a minimum distance between an original time series of two words (e. G., ECG bits and training bits) based on equation (2).

및

는 모폴로지 유사도 측정치가 결정되는 것의 상기 두 개의 단어를 나타내고,

및

는 두 개의 단어

및

의 SAX 심볼을 각각 나타낸다. 서브-함수(sub-function) dist()는 표 2에서 나타낸 룩업 테이블(lookup table)을 이용하여 구현될 수 있다.In Equation (2)

And

Represents the two words of which the morphology similarity measure is determined,

And

Two words

And

Respectively. The sub-function dist () can be implemented using a lookup table as shown in Table 2. &lt; tb &gt;&lt; TABLE &gt;

In step 314, the ECG bit is identified from the plurality of ECG bits of the ECG signal based on a training bit that includes the highest morphology similarity measure (lowest MINDIST value) with the ECG bit.

In step 316, it is determined whether the highest morphology similarity measure associated with the identified ECG bit is greater than a predetermined threshold value.

If the highest morphology similarity measure is determined to be less than the threshold, then in step 318, the identified ECG bit in the ECG signal is associated with an arrhythmia class corresponding to the training bit associated with the highest morphology similarity measure . In addition, the arrhythmia class corresponding to the training bit is identified as an arrhythmia class that indicates the type of cardiac arrhythmia associated with the individual.

In step 320, the arrhythmia class, which indicates the type of arrhythmia associated with the individual, may be displayed on a display unit of a dedicated health monitoring device, such as the display unit 105 of the dedicated health monitoring device 102 of FIG. do.

If the highest morphology similarity measure is greater than the threshold, then in step 322, the arrhythmia class that indicates the type of arrhythmia is assigned to the identified ECG bit (an arrhythmia class indicating a type of arrhythmia is assigned to the identified ECG beat). Additionally, the identified ECG bits that are assigned to the arrhythmia class, representing the type of arrhythmia, are stored in the training bit database 106, for example.

4 is a flow chart illustrating still another embodiment of a method 400 for classifying arrhythmias based on an ECG signal. For example, the method may be performed by the arrhythmia classification module 104 of FIGS. 1A and 1B.

In step 402, the ECG signal indicative of the cardiac electrical activity of the individual for a period of time is obtained, for example, from an acquisition device, such as acquisition device 108 of FIG. 1A. The ECG signal includes ECG bits. In step 404, the ECG signal indicative of the electrical activity of the heart of the individual for the predetermined period of time is filtered. In step 406, an R-peak corresponding to each ECG bit in the ECG signal is detected.

At step 408, the respective ECG bit region around the detected R-peak in each ECG bit is segmented. In step 410, the segmented region of each ECG bit is converted to a SAX symbol. In step 412, a normalized frequency of occurrence of the SAX symbols of the SAX symbol representing the respective ECG bits in the ECG signal is calculated.

In step 414, each ECG bit in the ECG signal is classified by belonging to one of the clusters of ECG bits based on the frequency of occurrence of the SAX symbol. In addition, the arrhythmia class corresponding to one of the clusters of ECG bits is identified as an arrhythmia class that indicates the type of cardiac arrhythmia associated with the individual.

For example, each ECG bit may be divided into three different ECG bits, such as normal ECG beats, premature ventricular contraction (PVC) ECG bits, and atrial premature complex (APC) ECG bits. It can be classified as either cluster or group (may be classified into any one of three different clusters or groups).

In step 416, the arrhythmia class corresponding to one of the clusters of ECG bits may be displayed on a display unit of a dedicated health monitoring device, such as the display unit 105 of the dedicated health monitoring device 102 of FIG. &Lt; RTI ID = do.

5 is a flow chart illustrating one embodiment of a method 500 of constructing a training bit database that includes a number of training beats. For example, the method may be performed by the arrhythmia classification module 104 of FIGS. 1A and 1B.

At step 502, an annotated ECG signal indicative of the cardiac electrical activity of the individual for a period of time is obtained, such as from the augmentation device 108 of FIG. 1A. In step 504, the ECG signal representing the electrical activity of the heart of the individual for the certain period of time is filtered. In step 506, an R-peak corresponding to each ECG bit in the ECG signal is detected.

In step 508, the area of each of the ECG bits around the detected R-peak in each ECG bit is segmented. In step 510, the training bits are selected for each arrhythmia class.

For example, the training bit may comprise a N bit, a Premature Ventricular Contraction (V) bit, an Atrial Premature Beat (A) bit, a Junctional Premature Contraction (J) ) Bits, a Junctional Escape (j) bit, and / or a Ventricular Escape (E) bit.

In step 512, the segmented region of each ECG bit is converted to a SAX symbol. In step 514, the SAX symbols corresponding to the segmented region of each ECG bit are stored as training bits in a training bit database.

6 is a flow chart illustrating one embodiment of a method 600 for detecting the R-peak at each ECG bit of the ECG signal. For example, the method may be performed by the R-peak detection module 154 of FIG. 1B.

In step 602, the filtered ECG signal representing the electrical activity of the individual's heart is windowed into portions. The filtered ECG signal may be represented by x0 [n], and the ECG signal may be sampled at a sampling frequency Fs Hz.

In order to determine the adaptive threshold used to detect the R-peak at each ECG bit, the window size is set so that the slowest acceptable heart rate corresponds to one bit in 2 seconds Based on a 30-bit assumption of 30 minutes per minute, it can be determined, for example, that T is 2 seconds (a window size may be determined to be T, eg, 2 seconds, based on an assumption that the slowest acceptable heart rate is 30 Beats Per Minute BPM) that corresponds to one beat in 2 seconds).

To further illustrate the situation where only the portion of the R-peak is detected in the windowed portion of the filtered ECG signal (to account for situations where only the portion of the R-peak is detected within a windowed portions in the filtered ECG signal, adjacent windowed portions are overlapped with the windowed portion by T0, e.g., 0.61 seconds.

In step 604, the adaptive threshold Th for each windowed potion of the filtered ECG signal is calculated based on an absolute difference signal for each windowed potion. The absolute difference signal comprises the absolute value of the difference determined between the ECG signals.

For example, the filtered ECG signal x0 [n] may be a combination of an averaging filter and a high pass filter to determine the absolute difference signal x1 [n] based on Equation (3) Is passed through an absolute differentiator component (not shown) that can act:

Further, the adaptive threshold Th may be determined based on Equation 4:

Here,? Represents a predetermined threshold value.

The z-transform of the filtered ECG signal may be determined based on equation (5): &lt; EMI ID =

In Equation (5), x ₀ (z) represents the z-transform of the original ECG signal. The Z-transform of the filtered ECG signal may be determined using two filters based on Equation (6): &lt; RTI ID = 0.0 &gt;

In equation (6) is _{x avg [n] = x 0} [n] + x 0 [n + 1] and the average filter to scale _hp x _{[n] = x avg [n} ] -x avg [n + 1] Scale a is associated with a high-pass filter (Equation 6 leads to a scaled averaging filter as x avg [n] = x 0 [n] + x 0 [n + 1], and a high pass filter as x hp [n] = x _avg [n] - x _avg [n + 1]).

The absolute differentiator determines the slope between two portions of the windowed portion of the ECG signal separated by a single windowed portion of the ECG signal. This helps to suppress the effects of intermediate variance. The adaptive threshold is based on a mean of the absolute difference signal for each windowed portion due to the fact that the magnitude of the gradient is necessary to locate the R-peak. of the absolute difference signal.

In step 606, a search for fiducial peaks in each windowed portion of the filtered ECG signal is performed based on the adaptive threshold Th for each windowed portion do.

More specifically, the adaptive threshold Th is used to eliminate and search for a fiducial peak such as P or T in each ECG bit because the slope of the R-peak is substantially higher than the slope of the P and T waves (is used to find and eliminate the fiducial peaks). Thus, the slope of the windowed portion of the filtered ECG signal above the adaptive threshold value indicates the presence of the R-peak in the windowed portion, and the R - &lt; / RTI &gt; peaks within the windowed potion that is smaller than the peak.

In step 608, false peaks are removed in each windowed portion of the filtered ECG signal. Because the windowed portions overlap, the same peak can be sensed twice, or the peak can cross the boundary of one window (may straddle a boundary of a windowed portion and as a result) As a result, the position of the peak corresponds to the maximum in the window and does not correspond to the actual maximum of the R-peak.

To avoid false detection, the locations of the two subsequent peaks are determined. If the distance between the two consecutive peaks is less than a predefined value (e.g., SKIP_WINDOW), the position of the minimum amplitude is discarded and the position of the maximum amplitude is declared as the R-peak. .

In step 610, the position of the R-peak corresponding to each ECG bit in the filtered ECG signal is detected in each windowed portion of the filtered ECG signal.

To detect the R-peak at each windowed portion of the filtered ECG signal, the sample is displayed as a reference where the value of the absolute difference signal x1 [n] intersects the adaptive threshold Th (where fiducially marked where the absolute difference signal x ₁ [n] crosses the adaptive threshold Th).

And wherein the maximum value of the peak is a maximum search window of the filtered ECG signal x ₀ [n] starting from the position of the sample marked with the reference in each said windowed portion of the filtered ECG signal ) (MAX_SR_WIN = 0.2 sec). Effectively, the maximum value represents a fiducial R-peak of a QRS complex of the QRS complex in each of the windowed portions of the filtered ECG signal.

Once an R-peak is detected, there may not be another QRS complex within the allowed time frame (e.g., SKIP_WIN = 0.25 sec) and thus a sample marked with another criterion within the allowed time frame Are discarded. By discarding the samples marked with different criteria within the allowed time frame, the computation in searching for the R-peak is substantially reduced.

FIG. 7 is a block diagram illustrating one embodiment of a dedicated health monitoring device 102 of FIG. 1A. 7, the dedicated health monitoring device 102 includes a processor 702, a memory 704, a read only memory (ROM) 706, a bus 708, a communication interface 710, a display 105, and an input device 712.

The processor 702 used herein may be a microprocessor, a microcontroller, a complex instruction set computing microprocessor, a reduced instruction set computing microprocessor, a much longer instruction microprocessor, a long instruction word microprocessor, an explicitly parallel instruction computing microprocessor, a graphics processor, a digital signal processor, and / or a processing circuit known in the art But is not limited to, all types of computing circuits, such as all other types of computing devices.

The processor 702 may also be implemented as a microprocessor or other programmable logic device or array, such as graphics or programmable logic devices or arrays, application specific integrated circuits, single-chip computers, / Other controllers known in the art to which the present invention pertains.

Memory 704 and ROM 706 may each include volatile memory and non-volatile memory. The memory 704 may comprise an arrhythmia classification module (not shown) that classifies cardiac arrhythmias using an ECG signal indicative of an electrical activity of the heart of the individual, based on one or more embodiments shown in Figures 1A- ) &Lt; / RTI &gt;

The memory also includes a training bit database 106 that stores training beats associated with each arrhythmia class. The memory 704 may be any suitable memory device such as a compact disk, digital video disk, diskettes, magnetic tape cartridges, memory cards, and / or other memory devices known to those of ordinary skill in the art to which the invention pertains. A removable media drive for handling compact disks, a hard drive, a read only memory, a random access memory, an erasable programmable read only memory, an electrically Erasable and programmable read only memory, and all such suitable memory device (s) for storing machine-readable instructions and data, , &Lt; / RTI &gt;

Embodiments of the present subject matter may also include functions, procedures, data structures, and / or applications that define low-level hardware contexts or abstract data or that perform tasks (or define abstract data types or low-level hardware contexts) May be implemented with a module containing a program.

The arrhythmia classification module 104 may be stored in the form of machine-readable instructions in any of the above-mentioned storage media and may be executed by the processor 702. For example, the computer program may be based on the embodiments and teachings described herein in a machine-readable manner capable of classifying cardiac arrhythmias using ECG signals obtained from the acquisition device 108 of FIG. 1A Command.

In yet another embodiment, the program may be included in a compact disc-read only memory (CD-ROM) and may be stored in the non-volatile memory from the CD- Lt; / RTI &gt;

The communication interface 710 receives the ECG signal representing the electrical activity of the heart of the individual from the acquisition device 108 of FIG. 1A. Display unit 105 displays an arrhythmia class that indicates the type of arrhythmia associated with the individual classified based on the ECG signal.

The bus 708 acts as an interconnect between the various components of the dedicated health monitoring device 102. Components such as input device 712 are well known in the art to one of ordinary skill in the art, and a description thereof is thus omitted.

The units and modules described herein may be implemented using hardware components and software components. For example, the hardware component may include a microphone, an amplifier, a band-pass amplifier, an audio-to-digital converter, and a processing device.

The processing device may be, for example, a processor, a controller and an arithmetic logic unit, a digital signal processor, a microcomputer, a field programmable array, a programmable logic unit, Or any other device capable of executing and responding to instructions in a defined manner.

The processing device may execute an operating system (OS) and one or more software applications executing on the OS. The processing device may also access, store, manipulate, and process data (may access, store, manipulate, process, and create data) in response to execution of the software. More simply, the description of the processing device is used as singular.

However, one of ordinary skill in the art to which the present invention belongs can understand that a processing device may include multiple processing elements and multiple types of processing elements. For example, the processing device may include multiple processors or one processor, or one controller (may include multiple processors or a processor and a controller). Other processing configurations, such as a parallel processor, are also possible.

The software may comprise a computer program, piece of code, instruction, or some combination thereof for independently or collectively designating or configuring the processing device, which is preferably operating. The software and data may be implemented in any type of machine, component, physical or virtual machine, computer storage media or device, or non-recurringly or temporally.

The software may also be distributed over a network coupled with a computer system and the software is stored and executed in a distributed fashion. The software and data may be stored by one or more computer readable recording media.

The computer-readable medium may include any data storage device capable of storing data that can be read by a computer system or a processing device. Embodiments of non-transitory computer readable recording media include read only memory (ROM), random-access memory (RAM), CD-ROMs, magnetic tape, floppy disks, and optical data storage devices.

It should also be noted that the functional programs, codes, and code segments that perform the embodiments presented above may be implemented using techniques such as those described in the related art And can be easily interpreted by a programmer having ordinary knowledge in the field.

The devices described above may be used in mobile phones, personal digital assistants (PDAs), digital cameras, portable game consoles, and MP3 players, portable / personal multimedia players (&quot; (HDTV), optical disk players, set-top boxes, home appliances, and mobile devices such as handheld e-books, PMPs, handheld e-books, portable laptops, global positioning systems May refer to devices such as those capable of wireless communication and network communication consistent with those disclosed in &lt; RTI ID = 0.0 &gt;

The computing system or computer may include a bus, a user interface, and a microprocessor in electrical communication with the memory controller. This further includes a flash memory device. The flash memory device may store N-bit data through the memory controller. The N-bit data will be processed or processed by the microprocessor, and N may be 1 or an integer greater than one (N may be 1 or an integer greater than 1). Wherein the computing system or computer is a mobile device and the battery may be additionally provided to provide an operating voltage of the computing system or computer.

It will be appreciated by those of ordinary skill in the art that the computing system or computer may further include an application chipset, a camera image processor (CIS), a mobile dynamic random access memory (DRAM), and the like. will be. The memory controller and the flash memory device may constitute a solid state drive / disk (SSD) using non-volatile memory to store data.

In the foregoing, a number of embodiments have been described by way of example. Nevertheless, it should be understood that this can be various modifications. For example, when the described techniques are performed in a different order and / or when the described systems, architectures, devices, or circuits are combined in different ways and / or are replaced or supplemented with other components or their equivalents Even if appropriate results can be achieved.

Claims (20)

A method for classifying cardiac arrhythmias using an electrocardiogram (ECG) signal, the method comprising:
Receiving the ECG signal from a sensor indicative of an electrical activity of the heart of the individual for a period of time, the ECG signal comprising an ECG bit;
Converting each of the ECG bits into a symbolic aggregate approximation (SAX) symbol; And
Identifying an arrhythmia class indicative of the type of cardiac arrhythmia associated with the individual from arrhythmia classes representing each type of cardiac arrhythmia based on the symbolic aggregate approximation (SAX) symbol corresponding to each of the ECG bits; Step
&Lt; / RTI &gt; The method according to claim 1,
Wherein the converting comprises:
Detecting respective R-peaks of the ECG bits;
Segmenting each region of the ECG bit around each R-peak; And
Converting each of the regions of the ECG bits into symbolic aggregate approximation (SAX) symbols;
&Lt; / RTI &gt; The method according to claim 1,
Wherein the identifying comprises:
Determining a morphology similarity measure between each of the training bits and the ECG bits corresponding to each of the arrhythmia classes, the similarity measure comprising a symbolic aggregate approximation (SAX) symbol representing each of the ECG bits And a symbolic aggregate approximation (SAX) symbol representing each of the training bits;
Identifying an ECG bit from the ECG bit based on the training bit, the ECG bit including the highest morphology similarity measure;
Determining whether the highest morphology similarity measure associated with the identified ECG bit is greater than a predetermined threshold;
Assigning to the identified ECG bit an arrhythmia class indicative of a type of cardiac arrhythmia if the highest morphology similarity measure is greater than the predetermined threshold value; And
Classifying the identified ECG bits by associating an arrhythmia class corresponding to the training bit associated with the highest morphology similarity measure if the highest morphology similarity measure is less than or equal to the predetermined threshold value
&Lt; / RTI &gt; The method according to claim 1,
Wherein the identifying comprises:
Determining a frequency of occurrence of the symbolic aggregate approximation (SAX) symbol representing each of the ECG bits; And
Classifying each of the ECG bits by belonging to one of the clusters of ECG bits based on the generation frequency of the symbolic aggregate approximation (SAX) symbol representing each of the ECG bits, The cluster corresponding to each of the arrhythmia classes;
&Lt; / RTI &gt; A health monitoring device comprising:
A communication interface configured to receive an electrocardiogram (ECG) signal from a sensor indicative of an electrical activity of an individual's heart for a period of time, the ECG signal comprising an ECG bit; And
Converting each of the ECG bits to a symbolic aggregate approximation (SAX) symbol, and calculating a respective type of cardiac arrhythmia based on the symbolic aggregate approximation (SAX) symbol representing each of the ECG bits An arrhythmia class module configured to identify an arrhythmia class indicative of a type of cardiac arrhythmia associated with the individual from an arrhythmia class indicative of
And a health monitoring device. 6. The method of claim 5,
The arrhythmia classification module comprises:
And a display module configured to display the arrhythmia class. 6. The method of claim 5,
The arrhythmia classification module comprises:
And a bandpass filter configured to filter the ECG signal. 6. The method of claim 5,
The arrhythmia classification module comprises:
And an R-peak detection module configured to detect each R-peak of the ECG bits. 9. The method of claim 8,
The arrhythmia classification module comprises:
A symbolic sum approximation configured to segment each region of the ECG bits around each R-peak and to convert the region of each of the ECG bits into the symbolic aggregate approximation (SAX) A health monitoring device comprising a module. 6. The method of claim 5,
Wherein the arrhythmia classifying module includes an arrhythmia class identifying module,
Wherein the arrhythmia class identification module comprises:
Determining a morphology similarity measure between each of the training bits and the ECG bits corresponding to each of the arrhythmia classes, the similarity measure comprising a symbolic aggregate approximation (SAX) symbol representing each of the ECG bits, And comparing symbolic aggregate approximation (SAX) symbols representing each of the training bits,
Identifying an ECG bit comprising the ECG bit and a highest morphology similarity measure from the ECG bit based on a training bit,
Determining whether the highest morphology similarity measure associated with the identified ECG bit is greater than a predetermined threshold,
Assigning an arrhythmia class indicative of the type of cardiac arrhythmia to the identified ECG signal if the highest morphology similarity measure is greater than the predetermined threshold value;
Configured to identify the identified ECG bit by associating an arrhythmia class corresponding to the training bit associated with the highest morphology similarity measure if the highest morphology similarity measure is less than or equal to the predetermined threshold value, device. 11. The method of claim 10,
Wherein the arrhythmia class identification module comprises:
And to store the identified ECG bits assigned to the arrhythmia class. 6. The method of claim 5,
Wherein the arrhythmia classifying module includes an arrhythmia class identifying module,
Wherein the arrhythmia class identification module comprises:
Determine the frequency of occurrence of the symbol representing each of the ECG bits, and
And classify each of the ECG bits by belonging to one of the clusters of ECG bits based on the frequency of occurrence of the symbol representing each of the ECG bits. 13. The method of claim 12,
Wherein the cluster of ECG bits corresponds to each of the arrhythmia classes. 6. The method of claim 5,
The arrhythmia class includes a normal, left bundle brunch block, or right bundle brunch block, or atrial premature beat, or premature ventricular contraction, or any combination thereof. A system, comprising:
An acquisition device configured to generate an electrocardiogram (ECG) signal indicative of an electrical activity of an individual's heart for a period of time, the ECG signal comprising an ECG bit; And
Health monitoring device
Lt; / RTI &gt;
The health monitoring device includes:
Converting each of the ECG bits into a symbolic aggregate approximation (SAX) symbol, and
To identify an arrhythmia class indicative of the type of cardiac arrhythmia associated with the individual from an arrhythmia class representing each type of cardiac arrhythmia based on the symbolic aggregate approximation (SAX) symbol representing each of the ECG bits System. 16. The method of claim 15,
Wherein the health monitoring device is further configured to display the arrhythmia class on a display panel. 16. The method of claim 15,
Wherein the health monitoring device is further configured to filter the ECG signal using a filter. 16. The method of claim 15,
The health monitoring device includes:
Detecting respective R-peaks of the ECG bits;
Segment each region of the ECG bit around each R-peak; And
And convert each of the regions of the ECG bits to a symbolic aggregate approximation (SAX) symbol. 16. The method of claim 15,
The health monitoring device includes:
Determining a morphology similarity measure between each of the training bits and the ECG bits corresponding to each of the arrhythmia classes; comparing the symbol representing each of the ECG bits with the symbol representing each of the training bits; - determined by
Identifying ECG bits from the ECG bits based on the training bits as training bits, the ECG bits including the ECG bit and the highest morphology similarity measure;
Determine whether the highest morphology similarity measure associated with the identified ECG bit is greater than a predetermined threshold;
If the highest similarity measure is greater than the predetermined threshold, assign an arrhythmia class indicative of the type of cardiac arrhythmia to the identified ECG bit; And
And classify the identified ECG bits by associating an arrhythmia class corresponding to the training bit associated with the highest morphology similarity measure if the highest similarity measurement is less than or equal to the predetermined threshold value. 16. The method of claim 15,
The health monitoring device includes:
Determine a frequency of occurrence of the symbolic aggregate approximation (SAX) symbol representing each of the ECG bits, and
Classifying each of the ECG bits by belonging to one of the clusters of ECG bits based on the generation frequency of the symbolic aggregate approximation (SAX) symbol representing each of the ECG bits, Corresponding to each of the arrhythmia classes.

Images