JP6430144B2 - ECG waveform detection apparatus, ECG waveform detection method, ECG waveform detection program, and imaging apparatus - Google Patents

Description

  Embodiments described herein relate generally to an electrocardiogram waveform detection apparatus, an electrocardiogram waveform detection method, an electrocardiogram waveform detection program, and an imaging apparatus.

  An electrocardiograph is a device that attaches electrodes to a living body and measures a potential difference between the electrodes. Information measured by an electrocardiograph is called an electrocardiogram (ECG) and is widely used in the medical field. Examples of information obtained from an electrocardiogram include a P wave (P-wave), an R wave (R-wave), a QRS complex wave (QRS complex), and a T wave (T-wave). These waveforms are used for diagnosis of various heart diseases, and are also used for a synchronization signal of a medical diagnostic apparatus capable of performing electrocardiogram synchronous imaging. Therefore, automatic waveform detection is important for industrial applications.

A. Ruha et al., “A Real-time Microprocessor QRS Detector System with a 1-ms Timing Accuracy for the Measurement of Ambulatory HRV”, IEEE Transactions on Biomedical Engineering, 1997; 44: 159-167

  An electrocardiogram waveform detection apparatus, an electrocardiogram waveform detection method, an electrocardiogram waveform detection program capable of reducing erroneous detection of a waveform not detected, even when a waveform not detected is superimposed on an ECG signal, There is a need for an imaging device.

  An electrocardiographic waveform detection apparatus according to an embodiment emphasizes at least one first ECG signal acquired from a subject, and generates a second ECG signal, and a template corresponding to a specific detection target waveform And a calculation unit that calculates an evaluation value by collating with the second ECG signal, and a detection unit that detects the detection target waveform based on the evaluation value.

The figure which shows an ECG signal typically. The figure which shows the electrocardiogram waveform detection apparatus of 1st Embodiment. The figure which shows the hardware constitutions of an electrocardiogram waveform detection apparatus. The figure which shows typically the waveform of the R wave and disturbance signal in the case where there is no high region emphasis processing, and when it exists. The flowchart which shows the process example of an electrocardiogram waveform detection apparatus. The figure which illustrates an ECG signal and a high frequency emphasis ECG signal. The figure which shows the structural example of the FIR filter of a high region emphasis part. The figure explaining an example of the production | generation method of a high region emphasis template. The figure which shows the concept of the collation process with a high region emphasis ECG signal and a high region emphasis template. The figure explaining the operation | movement concept of a detection part. The figure which shows an example of the evaluation result for confirming the effectiveness of an electrocardiogram waveform detection apparatus. The figure which shows the electrocardiogram waveform detection apparatus of 2nd Embodiment. The figure which shows the electrocardiogram waveform detection apparatus of 3rd Embodiment. The figure which shows the electrocardiogram waveform detection apparatus of 4th Embodiment. The figure which shows an electrocardiogram synchronous imaging device. The figure which shows the structural example of an electrocardiogram waveform detection apparatus. The figure which shows the structural example of an electrocardiogram waveform detection apparatus. The figure which shows the structural example of an electrocardiogram waveform detection apparatus.

  Embodiments of an electrocardiogram waveform detection apparatus, an electrocardiogram waveform detection method, an electrocardiogram waveform detection program, and an electrocardiogram synchronization imaging apparatus according to embodiments will be described with reference to the accompanying drawings. Note that in the following embodiments, the same reference numerals are assigned to the same operations, and duplicate descriptions are omitted as appropriate.

(First embodiment)
FIG. 1 is a diagram schematically showing an ECG signal (a signal corresponding to the shape of an electrocardiogram is called an ECG signal) that is a detection target of the electrocardiogram waveform detection apparatus 1 according to the present embodiment. As shown in FIG. 1, the ECG signal has specific waveforms such as a P wave, an R wave, a QRS composite wave (a composite wave of a Q wave, an R wave, and an S wave), and a T wave.

  In the following embodiments, an example of detecting an R wave among specific waveforms will be described. The example of detecting the R wave is an example, and the electrocardiographic waveform detection apparatus 1 of the embodiment can detect a waveform other than the R wave (eg, P wave, QRS composite wave, T wave, etc.). .

  When the electrocardiogram waveform detection device 1 detects an R wave, the electrocardiogram waveform detection device 1 outputs a heartbeat synchronization signal (synchronization signal) to the electrocardiogram synchronization imaging device (imaging device) 200. Examples of the electrocardiographic imaging apparatus 200 that can image in synchronization with the heartbeat include a CT (Computed Tomography) apparatus and an MRI (Magnetic Resonance Imaging) apparatus. For example, the electrocardiogram synchronous imaging apparatus 200 uses an imaging method (electrocardiographic synchronous imaging method) that determines the start timing of data collection with reference to an R wave generation position. The electrocardiogram-synchronized imaging apparatus 200 acquires a heartbeat synchronization signal corresponding to the position of the R wave, and determines a data collection start timing based on the acquired heartbeat synchronization signal.

  For example, the case of an MRI apparatus will be described as an example. In an MRI apparatus, various non-contrast MRA (Magnetic Resonance Angiography) such as FBI (Fresh Blood Imaging) method and Time-SLIP (Time-Spatial Labeling Inversion Pulse) method are used. ) Method is used. In data collection by the FBI method, the MRI apparatus collects diastolic images and systolic images by controlling the timing of data collection on the basis of, for example, a heartbeat synchronization signal, and calculates a difference image between them to obtain an artery. Can be obtained.

  Further, in data collection by the Time-SLIP method, the MRI apparatus can obtain an image of blood flow by controlling the timing of applying a labeling pulse and the timing of data collection with reference to a heartbeat synchronization signal, for example. . As described above, the MRI apparatus controls the data collection timing and the application timing of various pulses using the heartbeat synchronization signal generated from the ECG signal as a reference. In addition, the MRI apparatus performs imaging based on the heartbeat synchronization signal in other imaging, such as various imaging for the heart and the like, imaging using a contrast agent, and the like.

  For example, the heartbeat synchronization signal is also used when imaging the heart using the segmented method, or when imaging the heart in a specific cardiac phase. The heartbeat synchronization signal is also used when data of a basic cross-sectional image (for example, a left ventricular short axis image) of the heart is acquired in synchronization with the heartbeat synchronization signal, for example, in the diastole. Further, volume data including the entire heart is collected in synchronization with the heartbeat synchronization signal, for example, in the diastole. Then, image processing such as volume rendering and MIP is performed on the collected volume data.

  Also in the CT apparatus, imaging synchronized with the heartbeat synchronization signal is performed. For example, an imaging method is known in which a cardiac phase period with a large fluctuation in heart size is obtained from a heartbeat synchronization signal, and X-ray irradiation is stopped during the cardiac phase period with a large fluctuation, thereby reducing the exposure dose. Yes.

  FIG. 2 is a block diagram illustrating a configuration of the electrocardiogram waveform detection device 1 according to the first embodiment and a configuration of a device connected to the electrocardiogram waveform detection device 1. The electrocardiograph 100 generates an ECG signal and sends it to the electrocardiogram waveform detection apparatus 1. The pulse wave meter 400 generates a pulse wave signal and sends it to the electrocardiographic waveform detection apparatus 1. The electrocardiogram waveform detection apparatus 1 generates a heartbeat synchronization signal from the ECG signal and sends it to the electrocardiogram synchronization imaging apparatus 200.

  The electrocardiograph 100 includes electrodes 101a and 101b, an amplifier 110, and an AD converter 120. The electrodes 101a and 101b are attached to the human body. The amplifier 110 amplifies a weak potential difference between the electrodes 101a and 101b. The AD converter 120 converts the analog signal amplified by the amplifier 110 into a digital signal.

  The electrocardiograph 100 illustrates two electrodes 101a and 101b, but the number of electrodes is not limited to two. For example, in order to obtain a 12-lead electrocardiogram, a configuration including four electrodes attached to the limbs and six electrodes attached to the chest may be used. Further, instead of a method of obtaining a potential difference between two points on the body, a method of recording a potential difference between a predetermined reference and an electrode mounting point may be used.

  The electrocardiographic waveform detection apparatus 1 includes an input unit 10, a high frequency emphasis unit 20, a calculation unit 30, a detection unit 40, and a template generation unit 50.

  The input unit 10 acquires an ECG signal from the AD converter 120. The high frequency emphasizing unit 20 performs high frequency emphasis processing on the ECG signal to generate an ECG signal in which the high frequency component is emphasized, that is, a high frequency emphasized ECG signal. The calculation unit 30 compares the high frequency emphasis ECG signal with the high frequency emphasis template to calculate an evaluation value. Here, the high-frequency emphasis template is a template corresponding to a waveform obtained by performing high-frequency emphasis processing on a specific detection target waveform. In this example, the template corresponds to the waveform of the R wave that has been subjected to the high frequency emphasis process. Hereinafter, the high frequency emphasis template may be simply referred to as a template.

  The template generation unit 40 generates the above high frequency emphasis template based on the high frequency emphasis ECG signal generated by the high frequency emphasis unit 20. The detection unit 50 detects an R wave based on the evaluation value and generates a heartbeat synchronization signal. Further, the detection unit 50 sends the generated heartbeat synchronization signal to the electrocardiogram synchronization imaging apparatus 200.

  Each part of the electrocardiogram waveform detection apparatus 1 may be configured by hardware such as an application specific integration circuit (ASIC) or a field-programmable gate array (FPGA), or may be realized by software processing. Further, a combination of hardware and software processing may be realized. When realized by software processing, the operation of each part of the electrocardiogram waveform detection apparatus 1 can be realized by causing the computer 300 illustrated in FIG. 3 to execute a predetermined program.

  A computer 300 illustrated in FIG. 3 includes an input / output interface 301, a processor 302, a communication interface 303, a RAM (Random Access Memory) 304, a nonvolatile memory 305, and a disk drive 306.

  The non-volatile memory 305 is a storage device such as a hard disk or a flash memory, and stores various programs and data. The processor 302 is stored in the nonvolatile memory 305. The processor 302 reads a program for realizing the operation of each unit of the electrocardiogram waveform detection apparatus 1 from the nonvolatile memory 305 to the RAM 304 and executes it. In addition, a program stored in a recording medium such as a magnetic disk, an optical disk, or a USB memory may be read from the disk drive 306 or the input / output interface 301. Alternatively, it may be downloaded from an external server via the communication interface 303.

  The electrocardiographic waveform detection apparatus 1 according to the embodiment detects an R wave by comparing a high-frequency emphasized ECG signal with a high-frequency emphasized template. Hereinafter, this reason will be described.

  The waveform of the R wave that is the waveform to be detected in the ECG signal has a unique waveform for each subject. However, when viewed within the time to be measured, there is little temporal change in each R wave. Therefore, in order to improve the detection rate of the R wave, a unique waveform for each subject may be used as a template. On the other hand, among the waveforms other than the detection target in the ECG signal and the waveform of the disturbance signal superimposed on the ECG signal, there are those similar to the waveform of the R wave that is the detection target. For example, depending on conditions such as the position of the electrode, the R wave and the T wave show similar waveforms. In many cases, a disturbance signal similar to the R wave is superimposed on the ECG signal by electromagnetic induction in the MRI apparatus. Even in such a case, the disturbance signal may be erroneously detected as being an R wave.

  If the difference between the waveform other than the detection target and the waveform of the disturbance signal and the waveform of the R wave is small, it is difficult to stably detect the R wave. Therefore, in order to further improve the detection rate of the R wave, it is desirable to make the difference between the R wave and the other waveforms as large as possible.

  FIG. 4A is a diagram schematically showing an R wave waveform (solid line) and a disturbance signal waveform (broken line) similar to the R wave when the high-frequency emphasis process is not performed. When the normalized cross-correlation between the R wave and the disturbance signal shown in FIG. 4A is calculated, 0.89 is obtained. This value is smaller than 1 of the normalized cross-correlation value for completely the same waveform, but is a large value close to 1.

  Conventionally, a template corresponding to an R wave waveform and an input ECG signal are collated by a matching process such as a cross-correlation calculation. Then, the R wave is detected by comparing the cross-correlation value obtained as a comparison result with a predetermined threshold value. In this case, in order to avoid erroneous detection of the disturbance signal shown in FIG. 4A, the threshold value needs to be set larger than 0.89. However, when the threshold value is set to such a large value, it is difficult to stably detect the R wave when the detection target R wave has a fluctuation. On the contrary, if the threshold value is lowered in order to detect the R wave stably, the erroneous detection of the disturbance signal increases.

  On the other hand, in the electrocardiogram waveform detection apparatus 1 according to the embodiment, the high-frequency emphasis process is performed prior to the matching by the matching process. Then, the verification target waveform is converted into a waveform that is more sensitive to the width and inclination of the waveform.

  FIG. 4B shows waveforms obtained by performing high frequency emphasis processing on the R wave and the disturbance signal shown in FIG. In the waveform with the high frequency emphasis processing, the positive peak position and the positive peak value respectively correspond to the maximum positive inclination position and the maximum value of the positive inclination of the waveform without the high frequency emphasis processing. Further, in the waveform with the high frequency emphasis processing, the negative peak position and the negative peak value respectively correspond to the maximum negative inclination position and the maximum negative inclination of the waveform without the high frequency emphasis processing.

  In the case of no high frequency emphasis processing, as shown in FIG. 4A, even if the waveforms of the R wave and the disturbance signal are similar, in many cases, the width of the waveform, the slope of the rising or falling of the waveform Is different. For this reason, in the waveform subjected to the high frequency emphasis processing (FIG. 4B), either one or both of the positive and negative peak positions and peak values are different between the R wave and the disturbance signal. That is, the difference between the waveform of the R wave and the waveform of the disturbance signal is increased by performing the high frequency emphasis process. For example, the value of the normalized cross-correlation between the R wave and the disturbance signal is 0.89 when there is no high frequency emphasis processing, and decreases to 0.70 when there is high frequency emphasis processing. .

  As a result, a cross-correlation value between the template corresponding to the R wave waveform after the high frequency emphasis processing and the ECG waveform after the high frequency emphasis processing is obtained, and the R wave is detected by comparing the obtained cross correlation value with a threshold value. In this case, the probability (correct detection rate) of correctly detecting the R wave can be increased. Moreover, the probability (error detection rate) of erroneously detecting the disturbance signal can be suppressed even for the disturbance signal similar to the R wave.

  FIG. 5 is a flowchart showing an outline of processing of the electrocardiogram waveform detection apparatus 1. In step ST100, the input unit 10 of the electrocardiogram waveform detection apparatus 1 inputs an ECG signal as a time series signal. The ECG signal is, for example, a signal sampled at a constant period (for example, at 1 millisecond).

  In step ST102, the high frequency emphasis unit 20 performs high frequency emphasis processing on the input ECG signal. The high frequency emphasis unit 20 outputs a high frequency enhanced ECG signal, that is, a high frequency enhanced ECG signal. FIG. 6A is a diagram illustrating an ECG waveform input to the high frequency emphasizing unit 20. FIG. 6B is a diagram illustrating a high frequency emphasis ECG waveform output from the high frequency emphasizing unit 20. A positive maximum peak and a negative maximum peak occur at a position corresponding to the R wave of the high-frequency emphasized ECG waveform. The positive peak occurs at the maximum tilt position in the R wave rising region. Further, the negative peak occurs at the maximum tilt position in the falling region of the R wave. The zero crossing position between the positive maximum peak and the negative maximum peak corresponds to the peak position of the R wave.

  The method of high-frequency emphasis processing performed by the high-frequency emphasis unit 20 is not particularly limited. For example, differential processing realized by a multi-tap FIR (Finite Impulse Response) filter may be used. FIG. 7 shows an example in which the high frequency emphasis unit 20 is configured by a 5-tap FIR filter. In the FIR filter shown in FIG. 7, “τ” is a delay element, “W1” to “W5” are filter coefficients, and “+” is an addition element.

When the FIR filter shown in FIG. 7 is configured as a filter for high-frequency emphasis processing (differentiation processing), each filter coefficient (W1, W2, W3, W4, W5) is set as follows, for example.
(W1, W2, W3, W4, W5) = (-1, -2, 0, 2, 1)

The number of taps of the FIR filter may be other than 5. For example, when the number of taps is 4, each filter coefficient (W1, W2, W3, W4) is set as follows, for example.
(W1, W2, W3, W4) = (-1, -2, 2, 1)

When the number of taps is 3, each filter coefficient (W1, W2, W3) is set as follows, for example.
(W1, W2, W3) = (-1, 0, 1)

When the number of taps is 2, each filter coefficient (W1, W2) is set as follows, for example.
(W1, W2) = (− 1, 1)

  The filter coefficient of the FIR filter may be obtained by machine learning. As shown in FIGS. 4A and 4B, in order to reduce erroneous detection of disturbance signals, it is desirable that the normalized cross-correlation between the R wave as the detection target signal and the disturbance signal is as small as possible. . Therefore, the waveform data of the R wave and the waveform data of the disturbance signal are prepared in advance as learning data, and the filter coefficient is set so that the average normalized cross-correlation becomes small with respect to the FIR filter having a predetermined number of taps. Is calculated by machine learning. By using the filter coefficients calculated in this way for the FIR filter, it is possible to realize an FIR filter that performs high-frequency emphasis processing.

  Returning to FIG. 5, in step ST <b> 104, the template generation unit 50 selects a template corresponding to the R wave that is the detection target signal, that is, a high frequency enhancement template from the high frequency enhancement ECG signal generated by the high frequency enhancement unit 20. Generate.

  FIG. 8 is a diagram for explaining an example of a method for generating a high frequency emphasis template. The high frequency emphasis template is generated by the template generation unit 50 extracting a high frequency emphasis ECG signal in a period corresponding to the R wave, as shown in FIGS. 8A and 8B. The extraction position of the high frequency emphasis template is determined by the template generation unit 50 based on the pulse wave signal input from the pulse wave meter 400.

  The pulse wave signal is a signal obtained by measuring peripheral blood vessel movement with the pulse wave meter 400. As shown in FIG. 8C, in the pulse wave signal, a minimum point usually appears at a position slightly delayed from the R wave. Therefore, the position of the minimum point of the pulse wave signal is detected, and the high frequency emphasis ECG signal at a position slightly before the minimum point is cut out (extracted) to generate a high frequency emphasis template corresponding to the R wave. be able to.

  However, there are individual differences in the time from the R wave to the minimum point of the pulse wave signal. Therefore, an R wave search range is set in the previous period from the minimum point of the pulse wave signal, and a position where the absolute value of the ECG signal is maximum within the search range is determined as a peak position of the R wave. Then, a high frequency emphasis template corresponding to the R wave may be generated by extracting a high frequency emphasis ECG signal corresponding to a predetermined time width centered on the peak position of the R wave. Note that the time width of the extracted high-frequency emphasis template can be determined in advance from the assumed time width of the R wave.

  The high frequency emphasis template may be generated using an external signal other than the pulse wave signal. For example, a heart sound may be detected by a heart sound meter, and an R wave search range may be set using the detection result.

  The template generation unit 50 may update the high frequency emphasis template based on the latest high frequency emphasis ECG signal input as a time series signal. For example, a high frequency emphasis template may be generated every time an R wave arrives, and a past high frequency emphasis template may be replaced with the latest high frequency emphasis template. Alternatively, the update interval of the high frequency emphasis template may be set to a longer period composed of a plurality of R waves, and the past high frequency emphasis template may be replaced with the latest high frequency emphasis template for each update period. Alternatively, a high frequency emphasis template may be generated by performing a simple moving average or a weight moving average of a plurality of latest high frequency emphasizing ECG signals cut out each time an R wave arrives.

  When the high frequency emphasis template is generated as described above, in the next step ST106, the calculation unit 30 compares the high frequency emphasis ECG signal with the high frequency emphasis template and calculates an evaluation value. FIG. 9 is a diagram illustrating a concept of a collation process between the high frequency emphasis ECG signal and the high frequency emphasis template by the calculation unit 30.

  Hereinafter, some examples of collation processing by the calculation unit 30 will be described. In the following description, T (i) represents a high frequency emphasis template, S (i) represents a high frequency emphasis ECG signal, and E represents an evaluation value. N represents the length of the high frequency emphasis template, and i represents the index.

In the first example of the matching process, a matched filter is used. That is, the cross correlation between the high frequency emphasis template T (i) and the high frequency emphasis ECG signal S (i) is calculated, and the value of the cross correlation is set as the evaluation value E. In this case, the evaluation value E increases as the shape of the high frequency emphasis template T (i) and the high frequency emphasis ECG signal S (i) are similar, and as the strength of the high frequency emphasis ECG signal S (i) increases. . The evaluation value E is given below.

In the second example of the collation processing, the normalized correlation between the high frequency emphasis template T (i) and the high frequency emphasis ECG signal S (i) is calculated, and the value of the normalized correlation is set as the evaluation value E. The evaluation value E is normalized by the intensity of the high frequency emphasis template T (i) and the high frequency emphasis ECG signal S (i). In this case, the evaluation value E increases as the shape of the high frequency emphasis template T (i) and the high frequency emphasis ECG signal S (i) are similar. The evaluation value E is given below.

  Instead of the above formula, an average value of T (i) from each T (i), and a zero average normalized cross-correlation obtained by subtracting the average value of S (i) from each S (i) may be used. .

In the third example of the collation process, the sum of squared differences or the square root thereof is used as the evaluation value E. In these cases, the evaluation value E decreases as the shape of the high frequency emphasis template T (i) and the high frequency emphasis ECG signal S (i) are similar. In the case of the sum of squared differences, the evaluation value E is given by

In the fourth example of the collation process, the sum of absolute differences is used as the evaluation value E, or more generally, the sum of the difference Lp norms (p> 0) to the (1 / p) power. The sum of absolute differences is equivalent to p = 1. Assuming that p <2 in the Lp norm, if an outlier due to a measurement error or the like is included in a part of the high frequency emphasizing ECG signal S (i) or the high frequency emphasizing template T (i), the influence of the outlier is reduced. To achieve a robust effect. Expressions using the sum of absolute differences and the sum of Lp norms of the differences to the (1 / p) power are given below. In any case, the evaluation value E decreases as the shape of the high frequency emphasis ECG signal S (i) and the high frequency emphasis template T (i) are similar.

In the fifth example of the matching process, an error function is used in calculating the evaluation value E. The error function is a function that returns a value smaller than the square of the input value as a function value when the input value is large. By using the error function, the robustness, that is, the influence of the outlier value can be reduced. When the value of the error function for x is represented by R (x), the evaluation value E can be calculated from the following equation using the error function R (x). Also in this case, the evaluation value E becomes smaller as the shapes of the high frequency emphasis ECG signal S (i) and the high frequency emphasis template T (i) are similar.

Three examples of the error function R (x) are shown below. The shape of the error function is not limited to the following example. The parameter p of the error function may be determined in advance by conducting a preliminary experiment so that a good result can be obtained, for example.

  When the evaluation value is obtained as described above, in step ST108, the detection unit 40 detects an R wave based on the evaluation value and outputs a heartbeat synchronization signal to the electrocardiogram synchronization imaging apparatus 200.

  FIG. 10 is a diagram for explaining the concept of operation of the detection unit 40. As described above, in the first and second examples of the collation process, the evaluation value is calculated as a cross correlation or a normalized cross correlation. In this case, as shown in FIG. 10A, the evaluation value increases as the shape of the high frequency emphasis ECG signal and the high frequency emphasis template are similar.

  On the other hand, in the third to fifth examples of the collation processing, the evaluation value is calculated based on the difference. In this case, as shown in FIG. 10B, the evaluation value decreases as the shape of the high frequency emphasis ECG signal and the high frequency emphasis template are similar.

  In both cases of FIGS. 10A and 10B, a threshold value is set, and the position where the evaluation value is larger or smaller than the threshold value can be set as the R wave detection position. The threshold value may be a fixed value, but can be changed adaptively. For example, in the case of the third to fifth examples of the collation processing (in the case of FIG. 10B), the minimum value and median value of the evaluation value in a predetermined period immediately before detection are obtained, and the minimum value and median value are separately determined. A value obtained by adding a load according to the weight may be used as a threshold value.

  The detection by the detection unit 40 is not limited to a method using a threshold value. For example, the minimum point of the evaluation value may be obtained, and the position of the minimum point may be detected as the position of the R wave.

  Regardless of whether or not the threshold is used, a waveform similar to the R wave is detected in duplicate by excluding the time separately determined from the time when the R wave was detected from the detection target period. May be prevented.

  FIG. 11 is a diagram illustrating an example of an evaluation result for confirming the effectiveness of the above-described electrocardiographic waveform detection apparatus 1. An ECG signal disturbed by the gradient magnetic field of the MRI apparatus was acquired, and an R wave was detected for the ECG signal. The acquired ECG signal contained 1108 R waves. Compare the number of correctly detected R waves (number of positive detections) and the number of erroneously detected non-R waves as R waves (number of false detections) with and without high frequency enhancement processing did.

  As shown in FIG. 11A, the number of erroneous detections was 201 when there was no high frequency emphasis processing, but decreased to 134 when there was high frequency emphasis processing. On the other hand, as shown in FIG. 11B, the number of positive detections was 723 when there was no high-frequency enhancement processing, but increased to 916 when there was high-frequency enhancement processing. Thus, in the electrocardiogram waveform detection apparatus 1 using the high-frequency emphasis process, it was confirmed that the number of false detections was reduced while the number of positive detections was increased.

(Second Embodiment)
FIG. 12 is a block diagram illustrating a configuration of an electrocardiogram waveform detection device 1b according to the second embodiment. The difference from the electrocardiographic waveform detection apparatus 1 (FIG. 2) of the first embodiment is that it further includes a determination unit 60.

  The determination unit 60 determines whether or not a disturbance signal is superimposed on the ECG signal output from the electrocardiograph 100. Then, the determination result is output to the template generation unit 50. During the period in which the determination unit 60 determines that the disturbance signal is superimposed on the ECG signal, the template generation unit 50 temporarily stops generating or updating the high frequency emphasis template. On the contrary, the template generation unit 50 generates or updates the high frequency emphasis template only during the period when the determination unit 60 determines that the disturbance signal is not superimposed on the ECG signal. As a result, it is possible to prevent generation of a distorted high-frequency emphasis template caused by a disturbance signal.

  Whether or not a disturbance signal is superimposed on the ECG signal can be determined, for example, by monitoring the waveform of the ECG signal output from the input unit 10. For example, the determination may be made from the number of signals exceeding a predetermined threshold during a predetermined determination period.

  Further, instead of monitoring the ECG signal, or in parallel with the monitoring of the ECG signal, the operation state of the electrocardiogram synchronous imaging apparatus 200 is monitored, and the ECG is changed according to the operation state of the electrocardiogram synchronous imaging apparatus 200. It may be determined whether a disturbance signal is superimposed on the signal.

  When the electrocardiogram synchronous imaging apparatus 200 is an MRI apparatus, there is a high possibility that the waveform of the ECG signal is disturbed during application of an RF pulse or a gradient magnetic field pulse. In addition, the waveform of the ECG signal may be disturbed even during the period when the bed on which the patient is placed is moved. The disturbance of the ECG waveform during the movement of the bed can also occur when the electrocardiogram synchronous imaging apparatus 200 is a CT apparatus.

  The determination unit 60 determines that the disturbance signal is superimposed on the ECG signal during the period in which the electrocardiogram synchronous imaging apparatus 200 applies the RF pulse or the gradient magnetic field pulse or the period in which the electrocardiogram synchronous imaging apparatus 200 is moving on the bed. On the contrary, it is determined that the disturbance signal is not superimposed on the ECG signal during the period in which the ECG-synchronous imaging apparatus 200 does not apply the RF pulse or the gradient magnetic field pulse or the period in which the ECG signal is not moved. And the template production | generation part 50 produces | generates or updates a high region emphasis template in the period determined as the disturbance signal not being superimposed.

  For example, when the electrocardiographic synchronous imaging apparatus 200 is an MRI apparatus, the electrocardiographic synchronous imaging apparatus 200 has a plurality of protocols (for example, T1-weighted images) according to the imaging conditions input from the operator at the imaging planning stage in one examination. , A plurality of protocols for collecting images of a plurality of imaging types such as T2-weighted images) may be sequentially executed. The electrocardiogram synchronous imaging apparatus 200 continuously executes a plurality of protocols, but there may be a certain interruption time between the protocols. The determination unit 60 receives, for example, information indicating that the protocol is started, stopped, or being executed from the electrocardiogram synchronous imaging apparatus 200 as the operation state information, and based on this information, between the protocols. The interruption time is determined that the disturbance signal is not superimposed on the ECG signal, and it is determined that the disturbance signal is superimposed on the ECG signal during the execution of the protocol.

(Third embodiment)
FIG. 13 is a block diagram illustrating a configuration of an electrocardiographic waveform detection apparatus 1c according to the third embodiment. The difference from the electrocardiographic waveform detection apparatus 1 (FIG. 2) of the first embodiment is that a template storage unit 70 is provided instead of the template generation unit 50.

  In the first and second embodiments, a part of the high frequency emphasis ECG signal output from the high frequency emphasizing unit 20 in real time is cut out to generate or update the high frequency emphasis template. In contrast, the electrocardiographic waveform detection apparatus 1c according to the third embodiment generates a high-frequency emphasis template in advance prior to imaging of the subject, and stores this in the template storage unit 70. Since the high-frequency emphasis template is generated before imaging, for example, even when the electrocardiogram synchronous imaging apparatus 200 is an MRI apparatus, the high-frequency emphasis template is not affected by disturbance signals caused by application of RF pulses or gradient magnetic field pulses. The region enhancement template can be provided to the calculation unit 30.

(Fourth embodiment)
In the first to third embodiments described above, the ECG signal output from the electrocardiograph 100 to the electrocardiogram waveform detection devices 1, 1b, 1c is one-dimensional (the number of ECG signals is one).

  On the other hand, for example, a 12-lead type electrocardiograph outputs 12 (12-dimensional) ECG signals such as I, II, III, aVR, aVL, aVF, and V1 to V6. For example, in the case of a 4-terminal electrocardiograph used for electrocardiogram synchronization, 2 to 3 ECG signals (2 to 3 dimensions) are output. For example, the vector cardiogram uses three ECG signals (three-dimensional) generated from a plurality of electrode signals (these three ECG signals are called X, Y, and Z).

  In the fourth embodiment, R waves are detected from these multidimensional ECG signals output from the electrocardiograph 100. FIG. 14 is a block diagram illustrating a configuration of an electrocardiographic waveform detection apparatus 1d according to the fourth embodiment. The electrocardiogram waveform detection apparatus 1d includes a dimension reduction unit 80 and an evaluation value synthesis unit 90. In addition, a plurality of high frequency emphasis units 20, a calculation unit 30, and a template generation unit 40 are provided corresponding to the number of dimensions of the signal to be processed.

  The dimension reduction unit 80 reduces the number (dimension number) of ECG signals output from the electrocardiograph 100. For example, when the number of ECG signals output from the electrocardiograph 100 is n (n-dimensional signal), the ECG signal is converted into a signal having a dimension number smaller than n (m dimension: (m <n)). For example, a principal component vector is calculated by performing principal component analysis on an n-dimensional ECG signal input in the past for a predetermined time. In addition, by projecting an input n-dimensional ECG signal to a subspace generated from m (m <n) principal component vectors having a large contribution rate, the number of dimensions is changed from n dimensions to m dimensions. Can be reduced.

  The ECG signals with reduced dimensions, that is, ECG signals with the number of signals reduced from n to m are each input to one of the m high frequency emphasizing units 20. Note that a plurality of ECG signals output from the electrocardiograph 100 may be input to each high frequency emphasizing unit 20 without reducing the dimension. In this case, the dimension reduction unit 80 is unnecessary.

Each high frequency emphasis unit 20 performs high frequency emphasis processing similar to that of the first embodiment on each input ECG signal. Now, when m ECG signals are represented by an m-dimensional vector signal S (t), the vector signal S (t) is as follows.
S (t) = (S1 (t), S2 (t), S3 (t), ..., Sm (t))

  The high frequency emphasizing unit 20 performs, for example, an FIR filter having the above-described filter coefficient independently of S1 (t), S2 (t), S3 (t),..., Sm (t). Apply to generate a high frequency enhanced ECG signal. In this case, each high frequency emphasis unit 20 is configured as a 1-input 1-output FIR filter. Then, the m high frequency emphasizing units 20 output m high frequency emphasizing ECG signals as a whole.

  For example, in the case of a vector cardiogram, the above-described three signals, X, Y, and Z are input to the three high-frequency emphasizing sections 20 as S1 (t), S2 (t), and S3 (t), respectively. Then, three high frequency emphasized signals are output from each of the three high frequency emphasizing units 20.

  Note that the high frequency emphasis unit 20 may be configured by a multi-dimensional FIR filter having a plurality of inputs and a plurality of outputs. For example, when the input signal is an X, Y, Z three-dimensional signal corresponding to the vector cardiogram, the high frequency emphasis unit 20 may be configured as a three-input three-output three-dimensional FIR filter. .

  The m high frequency emphasis ECG signals generated by the high frequency emphasis unit 20 are input to the m template generation units 50, respectively. Each template generation unit 50 generates a high frequency emphasis template. Each template generation unit 50 outputs the generated high frequency emphasis template to m calculation units 30. The operation of each template generation unit 50 is the same as in the first to third embodiments.

  Each calculation unit 30 evaluates a cross correlation or a sum of squared differences based on the high frequency emphasis ECG signal output from each high frequency emphasis unit 20 and the high frequency emphasis template output from each template generation unit 50. Calculate the value. The operation of each calculation unit 30 is the same as in the first to third embodiments.

  The evaluation value synthesis unit 90 synthesizes a total of m evaluation values output from the m calculation units 30 to calculate one total evaluation value. For example, the total evaluation value may be calculated by simply adding m evaluation values. Alternatively, m evaluation values may be appropriately weighted, and the weighted addition value may be set as a comprehensive evaluation value.

  The detection unit 40 detects an R wave based on the comprehensive evaluation value output from the evaluation value synthesis unit 90. The operation of the detection unit 40 is the same as in the first to third embodiments.

  According to the electrocardiogram waveform detection apparatus 1d of the fourth embodiment, it is possible to detect an R wave using information of a multidimensional ECG signal output from the electrocardiograph 100.

  In each of the above-described embodiments, the example in which the electrocardiographic waveform detection device 1 is configured separately from the electrocardiogram synchronous imaging device 200 has been described. The electrocardiogram waveform detection apparatus 1 may be an internal configuration of the electrocardiogram synchronous imaging apparatus 200.

  FIG. 15 is a diagram illustrating an example in which the electrocardiogram synchronous imaging apparatus 200a includes the electronic waveform detection apparatus 1. In addition to the electrocardiographic waveform detection device 1 that generates a heartbeat synchronization signal, the electrocardiogram synchronization imaging device 200a includes a data collection unit 210 that collects imaging data from a subject in synchronization with the heartbeat synchronization signal, and the collected imaging imaging The image generation unit 220 generates an image of the subject from the data.

  In each of the above-described embodiments, the example in which the electrocardiographic waveform detection apparatus 1 is configured separately from the electrocardiograph 100 is shown. The electrocardiograph 100 may be the internal configuration of the electrocardiogram waveform detection apparatus 1.

  FIGS. 16 to 18 are diagrams showing an example in which the electrocardiographic waveform detection apparatus 1 includes an electrocardiograph 100.

  The ECG waveform detection apparatus 1 shown in FIG. 16 inputs the ECG signal received from the AD converter 120 to the input unit 10 by wire.

  The electrocardiographic waveform detection apparatus 1 illustrated in FIG. 17 further includes a transmission unit 130 and a reception unit 132. The transmission unit 130 and the reception unit 132 can perform wireless communication with each other. The receiving unit 132 receives an ECG signal that the transmitting unit 130 transmits wirelessly. According to this configuration, wiring to the outside can be eliminated from the electrocardiograph 100 attached to a patient lying on the inside of the bore of the electrocardiogram synchronous imaging apparatus 200.

  The electrocardiographic waveform detection apparatus 1 illustrated in FIG. 18 further includes a storage unit 140 and a reading unit 142. The storage unit 140 stores the ECG signal in a portable recording medium such as a CD, DVD, or USB memory. The reading unit 142 reads a recording medium. The reading unit 142 sends the read ECG signal to the input unit 10.

  Note that the various processing procedures described above (for example, FIG. 5) are not necessarily limited to the illustrated processing procedures. Processing procedures that can be executed in parallel may be executed in parallel, and processing procedures that do not depend on the order may be executed by changing the order.

  According to at least one embodiment described above, erroneous detection of a waveform outside the detection target can be reduced even when the waveform outside the detection target is superimposed on the ECG signal.

  Although several embodiments of the present invention have been described, these embodiments are presented by way of example and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the scope of the invention. These embodiments and modifications thereof are included in the scope and gist of the invention, and are included in the invention described in the claims and the equivalents thereof.

DESCRIPTION OF SYMBOLS 1 ECG waveform detection apparatus 10 Input part 20 High region emphasis part 30 Calculation part 40 Detection part 50 Template production | generation part 60 Determination part 70 Template memory | storage part 80 Dimension reduction part 90 Evaluation value synthetic | combination part 100 Electrocardiograph 200 Electrocardiogram synchronous imaging device

Claims (14)

An emphasis unit that performs high frequency emphasis on at least one first ECG signal acquired from the subject and generates a second ECG signal;
A template corresponding to a waveform that is generated based on the second ECG signal and subjected to high-frequency emphasis processing on a specific waveform to be detected is compared with the second ECG signal to obtain an evaluation value. A calculation unit for calculating,
A detection unit for detecting the detection target waveform based on the evaluation value;
An electrocardiographic waveform detection apparatus comprising: A determination unit that determines whether to generate the template;
The generation unit generates the template based on a determination result of the determination unit.
The electrocardiographic waveform detection apparatus according to claim 1 . The determination unit acquires information on the operating state of the MRI apparatus, determines that the template is not generated during a period in which the MRI apparatus collects the magnetic resonance signal from the subject, and determines the magnetic resonance signal. During the period not collected, it is determined that the template is generated.
The electrocardiographic waveform detection apparatus according to claim 2 . The generation unit extracts a signal of a predetermined period of the first ECG signal based on the pulse wave signal of the subject, and generates the template based on the extracted signal of the predetermined period.
The electrocardiographic waveform detection apparatus according to claim 1 . The evaluation value is a correlation between the second ECG signal and the template.
The electrocardiographic waveform detection apparatus according to any one of claims 1 to 4 , wherein The evaluation value is a sum of squares of differences between the second ECG signal and the template, a sum of absolute differences, an L-p norm (p> 0) with respect to the difference, or when the difference is larger than a predetermined value, A value obtained based on at least one of the sum of function values calculated by applying a function having a value smaller than the square to the difference,
The electrocardiographic waveform detection apparatus according to any one of claims 1 to 5 , wherein The enhancement unit performs high frequency enhancement on each of the plurality of first ECG signals acquired simultaneously from the subject, and generates a plurality of the second ECG signals,
The calculation unit obtains a plurality of evaluation values corresponding to a plurality of the second ECG signals,
A synthesis unit that adds a plurality of evaluation values or calculates a total evaluation value by weighted addition; and
The detection unit detects the detection target waveform based on the comprehensive evaluation value;
The electrocardiographic waveform detection apparatus according to claim 1 , wherein The plurality of first ECG signals are a plurality of signals for generating a vector cardiogram.
The electrocardiographic waveform detection apparatus according to claim 7 . When the number of the first ECG signals is N and the plurality of first ECG signals are N-dimensional time series signals,
A dimension reduction unit that performs a dimension reduction process for reducing the number of the first ECG signals to a number smaller than N before the high-frequency emphasis process;
The dimension reduction unit includes:
A principal component analysis is performed on the N-dimensional time series signal to calculate a principal component vector, and the N-dimensional time series signal is generated in a subspace generated from a principal component vector having a contribution rate larger than a predetermined value. Reducing the number of first ECG signals to a number less than N by projecting;
The electrocardiographic waveform detection apparatus according to claim 7 . A storage unit for storing the template;
The template is generated before acquiring the first ECG signal and stored in the storage unit.
The electrocardiographic waveform detection apparatus according to claim 1. An emphasis unit that performs high frequency emphasis on at least one first ECG signal acquired from the subject and generates a second ECG signal;
A template corresponding to a waveform that is generated based on the second ECG signal and subjected to high-frequency emphasis processing on a specific waveform to be detected is compared with the second ECG signal to obtain an evaluation value. A calculation unit for calculating,
A detection unit that detects the detection target waveform based on the evaluation value and generates a synchronization signal;
A data collection unit for collecting data for imaging from the subject in synchronization with the synchronization signal;
An image generation unit that generates an image of the subject from the collected data for imaging;
An imaging apparatus comprising: The imaging apparatus is an MRI apparatus;
The imaging apparatus according to claim 11 . High-frequency emphasis on at least one first ECG signal acquired from the subject to generate a second ECG signal;
A template corresponding to a waveform that is generated based on the second ECG signal and subjected to high-frequency emphasis processing on a specific waveform to be detected is compared with the second ECG signal to obtain an evaluation value. Calculate
Detecting the detection target waveform based on the evaluation value;
ECG waveform detection method. On the computer,
High frequency enhancement of at least one first ECG signal acquired from a subject to generate a second ECG signal;
A template corresponding to a waveform that is generated based on the second ECG signal and subjected to high-frequency emphasis processing on a specific waveform to be detected is compared with the second ECG signal to obtain an evaluation value. A calculating step;
Detecting the detection target waveform based on the evaluation value;
ECG waveform detection program to execute.

Images