JP2536410B2 - ECG analyzer - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an electrocardiogram analyzer which can accurately detect and analyze segment points even when only a limb-lead electrocardiogram, which is easy to measure, is used among standard 12-lead electrocardiograms.

[0002] The electrocardiogram is mainly used in clinical situations such as discovery and diagnosis of heart diseases and monitoring of pathological conditions. Further, since the effects of stress and fatigue on the autonomic nervous system appear in heart rate, heart rate fluctuation, etc., they are also used in fields such as ergonomics and industrial medicine.

Therefore, the present invention can be applied to all fields using electrocardiogram, such as medicine and ergonomics.

[0004]

2. Description of the Related Art FIG. 2 exemplifies each wave and division point that characterize an electrocardiogram. In the automatic analysis of electrocardiogram, the division points in one heartbeat, that is, the starting position, the ending position of each wave of PQRSTU, the peak position, the electric potential, etc. at each are detected, and the diagnosis support information is output by the analysis based on the knowledge of clinical physiology . Here, since the processing after the detection of the division points is based on the knowledge of clinical physiology, technically, the accuracy of the detection of the division points is a key point. In the conventional electrocardiogram analysis device, the chest lead data having a large electric potential out of the standard 12 leads (6 leads in the chest and 6 leads from the extremities) is used to improve the recognition rate of segment points and the positioning accuracy. .

[0005]

In the conventional electrocardiogram segment point detection, the chest lead data is mainly used, noise is removed by smoothing, and the second-order differential or second-order difference is applied to detect the segment points. ing.

[0006] In the chest lead, waveform data having a large electric potential and a distinctive feature can be obtained, but it is necessary to attach electrodes to the chest and maintain a rest and supine posture, and the subject must be restrained. For this reason, the psychological burden becomes large, which is a non-negligible fluctuation factor in electrocardiogram measurement for the purpose of stress measurement. On the other hand, in limb guidance, measurement is simple,
The potential of the obtained waveform data is small. Therefore, the start point and the end point of the gently changing T wave become unclear due to the smoothing, and the feature enhancement effect due to the second derivative or the second difference is hard to appear. . This impairs the reliability of the analysis result.

[0007] If there is an electrocardiogram analyzer that can detect segment points with the same or better accuracy than the case of using chest lead data with only limb lead data, not only for medical use but also for stress measurement, etc. It is also useful for targeted electrocardiographic measurements.

An object of the present invention is to provide an electrocardiogram analysis apparatus which can detect segment points with the same degree of accuracy as in the case of using chest lead electrocardiogram data only with limb lead electrocardiogram data.

Another object of the present invention is to eliminate the need to attach electrodes to the chest when using the positional relationship of the division points for diagnosis, for example, when examining arrhythmias or excitatory conduction abnormalities, and the burden on the subject. An object of the present invention is to provide an electrocardiogram analysis device that reduces

Another object of the present invention is to reduce the constraint during measurement when the psychological burden during measurement is likely to affect the data, such as stress measurement, and the measurement itself becomes a cause of fluctuation. An object of the present invention is to provide an electrocardiogram analysis device capable of suppressing this.

[0011]

The electrocardiogram analysis apparatus of the present invention stores the electrocardiogram time series data required for calculating the feature amount for detecting the division points in the apparatus for detecting the division points of the electrocardiogram and analyzing the electrocardiogram. The input data storage unit, the waveform enhancement unit that calculates the time-series data of the enhanced waveform that emphasizes the features of the electrocardiogram waveform, the flexion calculation unit that calculates the flexion used for peak detection, and the data on the baseline. The base line discriminant amount calculation unit for calculating the base line discriminant amount for discriminating whether or not, the emphasized waveform time series data output by the waveform emphasizing unit, the flexion degree output by the flexion degree calculation unit, and the base line discriminant amount. A characteristic amount storage unit that receives and stores the baseline discriminant amount output by the calculation unit, and searches for the emphasized waveform time-series data, the degree of bending, and the baseline discriminant amount stored in the characteristic amount storage unit, and divides the points. Detect the ward A division point detection unit that outputs division point information, such as the position of a point and a potential at the division point, and an analysis unit that receives and analyzes the division point information output by the division point detection unit and outputs diagnostic support information. It is characterized by having.

[0012]

In the device for detecting the electrocardiogram segmentation points and analyzing the electrocardiogram, since it has the input data storage unit, it is possible to store the electrocardiogram time-series data required for the feature amount calculation for segmentation point detection. , Since it has the waveform enhancement unit, it is possible to calculate the enhanced waveform time-series data that emphasizes the characteristics of the electrocardiographic waveform, and since it has the flexibility calculation unit, it is possible to calculate the flexibility used for peak detection.
Since the base line discriminant amount calculation unit is provided, the base line discriminant amount for discriminating whether or not the data is on the base line portion can be calculated. The enhanced waveform time-series data, the bending degree output by the bending degree calculation unit, and the baseline discriminant amount output by the baseline discriminant amount calculation unit can be received and stored, and a segment point detection unit is included. From the above, the enhanced waveform time-series data, the degree of bending, and the baseline discriminant amount stored in the feature amount storage unit are searched to detect a partition point, and the partition point information such as the position of the partition point and the potential at the partition point is output. Since it has an analysis unit, it can receive and analyze the division point information output by the division point detection unit, and output the diagnosis support information.

From the above, even with only the limb-lead electrocardiogram data having a small electric potential, the segment points can be accurately detected and analyzed,
Diagnosis support information can be output.

[0014]

Embodiments of the present invention will be described with reference to the drawings.

FIG. 1 shows the basic construction of an embodiment of the electrocardiogram analysis apparatus of the present invention.

The electrocardiogram analysis apparatus of the present invention includes an input data storage unit 1 for storing electrocardiogram time series data 100 required for calculating feature quantities for detecting segment points, and an emphasized waveform time series emphasizing the characteristics of the electrocardiogram waveform. A waveform enhancement unit 2 that calculates data 101, a bending degree calculation unit 3 that calculates a bending degree 102 used for peak detection, and a baseline that calculates a baseline determination amount 103 for determining whether the data is on a baseline portion. The discriminant amount calculation unit 4, the emphasized waveform time-series data 101 output by the waveform emphasizing unit, the bending degree 102 output by the bending degree calculating unit, and the baseline discrimination amount 103 output by the baseline discriminating amount calculation unit are received. The feature quantity storage unit 5 stored in advance, the emphasized waveform time series data 101 stored in the feature quantity storage unit 5, the bending degree 102 and the baseline discriminant amount 103 are searched to detect a division point. Potential or the like in the position and classification point of demarcation points, the division point detection unit 6 which outputs the segment point information 200,
The analysis unit 7 receives and analyzes the division point information 200 output by the division point detection unit 6 and outputs the diagnosis support information 300.

The electrocardiogram time series data 100 is input to the input data storage unit 1. The input data storage unit 1 has a first-in first-out (FIFO) structure, and stores a to the latest data among the electrocardiogram time-series data 100 obtained one after another. The latest data of the time series data 100 of the electrocardiogram is set as x _p, and the electrocardiographic data stored in the input data storage unit 1 is set as X _i (i = 1, 2, ...,
When _{a), X 1 = x p} - (a-1), X 2 = x p - (a-1) +
1, ..., X _a = x _p . As the number a of data to be stored, a number sufficient for the processes of the waveform enhancing unit 2, the bending degree calculating unit 3, the baseline discriminating amount calculating unit 4, and the dividing point detecting unit 6 described later is specified in advance. .

The waveform emphasizing unit 2 is the input data storage unit 1.
From the electrocardiogram time series data 100 stored in, the b pieces of data necessary for the waveform emphasis calculation are read out, and the emphasized waveform time series data 101 is calculated. As the emphasized waveform data 101, for example, a mean square value that reflects a sudden change in power is used,

[0019]

Calculate b is the number of X _i used to calculate the root mean square value and is an odd number of 3 or more. C ₁ is a constant, and is added to prevent the waveform from being reduced when X < _i ² <X _i when 0 <X _i <1. C _{2 is} also a constant,
It is for returning the shift of the waveform due to the addition of C ₁ . If C ₂ = −C ₁ , the value of X _i = 0 will be stored.

The bending degree calculating unit 3 receives the emphasized waveform time series data 101 output from the waveform enhancing unit 2 and calculates the bending degree 102. The bending degree 102 is, for example, 2
Utilizing floor _{difference, R k = R j - m} = 2Y j - (Y j - m + Y j + m) (j = m + 1, m + 2, ···, a- (b-1) -m;
k = 1, 2, ..., A- (b-1) -2m) is calculated. Here, m is the width of the second-order difference M (an odd number of 3 or more)
Is a number such that m = (M−1) / 2. The tortuosity 102 is obtained from the time-series data 101 of the emphasized waveform, so that the peak sharply reflects even when the potential is small as in the limb lead electrocardiogram.

The baseline discriminant amount calculation unit 4 reads the required number of electrocardiogram time series data 100 stored in the input data storage unit 1 and calculates the baseline discriminant amount 103. In the electrocardiogram, the fact that there is a difference in the variance of the short section between the base line portion and each PQRSTU wave portion is used, for example,

[0023]

Is calculated as the base line discrimination amount 103. Here, v _h is the variance of the electrocardiogram time series data within the short section. TH _{V b} is a baseline variance threshold that serves as a reference for discriminating between the peak and the baseline in the electrocardiogram waveform. TH _{V p} is X
It is a peak variance threshold value that serves as a reference for determining whether _i is in the peak portion of the electrocardiogram waveform. w is the number of electrocardiogram time series data used for calculation of v _h , and is an odd number of 3 or more.
The variance v _h in the short interval is

[0025]

It can be obtained as follows. Where w
Is the number of data contained in the section, μ _h is the average value of the electrocardiogram data in the short section,

[0027]

It can be expressed as Appropriate values are designated in advance for the baseline variance threshold TH _{V b} and the peak variance threshold TH _{V p} , or they are designated by the user.

[0029] Here, the obtained B _h is, B _h = 1 if _{X h + (w - 1)} / 2 is that in the peak portion, B _h
If = 0, it means that Xh _{+ (w-1) / 2} is in the base line portion.

The characteristic quantities storage unit 5 stores the first-in first-out (FI
It has a FO) structure, receives the emphasized waveform time-series data 101, the degree of bending 102, and the baseline discriminant amount 103, and stores a predetermined number of the latest data for them.

The division point detecting section 6 detects division points (see FIG. 2) in the electrocardiogram. The time interval of each division point has a normal range based on physiological findings. Normally, in detecting the segment points of the electrocardiogram, the reference R-wave peak position is first detected. For other division points, it is most efficient to search and determine an appropriate range from the relative positional relationship between the normal range of each division point and the R wave peak position.
Also in the present embodiment, the procedure will be described from the detection of the R-wave peak position by setting the segment point search range in the same procedure as the conventional method.

For the R-wave peak detection, the degree of curvature R _k of the characteristic quantities stored in the characteristic quantity storage unit 5 is used.

First, the latest value of the bending degree R _kn is compared with a predetermined threshold value TH _p, and if R _kn ≧ TH _p , R _k is searched for in the direction of decreasing k. R _kn ≧ within a predetermined search range
When TH _p and R _k have peaks, i corresponding to k = k _{R p} at that time is calculated as the electrocardiogram R wave peak position i.
_{Let R p} .

After the R-wave peak position is determined, the search range of each division point is searched and the division points are sequentially determined. The search and determination of the Q wave start point and the T wave end point will be described below as an example.

To detect the Q wave starting point, after detecting the peak position of the R wave of the electrocardiogram, from h = h _{R p} corresponding to k _{R p} ,
The search is performed while comparing the value of B _h with a predetermined threshold value TH _{Q b} in the direction of decreasing. The B _h ≦ TH _{Q b} and since h = h _{Q b} and Q wave start point, the i corresponding to h _{Q b,} and electrocardiogram Q wave start point position i _{Q b.}

When the T wave end point is detected, the T wave end point search is started from the R wave peak position i _{R p} by RT (the later in time). Here, RT is a value based on physiological knowledge, and is set so that the search start point of the T wave end point falls within the T wave portion. In the direction of increasing _h , the value of B _h is compared with a predetermined threshold value TH _{T e} for searching, and B
_h ≦ TH _{T e} and became the h = h _{T e} and T wave end point, h
Let i corresponding to _{T e} be the T wave end point position i _{T e} of the electrocardiogram.

With respect to the other division points as well, within the search range set for each division point, by comparing the bending degree R _k for the peak and the threshold value for each peak, the boundary between the base line portion and each wave is determined. The base line discriminant amount B _h is compared with a threshold to search and determine.

By reading the measured value at each division point position from the input data storage unit 1, the peak position and the measured value at that time are the same as the position of each wave start point like i _{R p} and X _{i R p.} The measured value at that time is i _{Q b} and Xi _{Q b} ,
The position of each wave end point and the measured value at that time are i _{T e} and X
_{It is calculated} like _{i T e} . The division point detection unit 6 outputs a combination of these as division point information 200.

The analysis unit 7 receives the division point information 200 output from the division point detection unit 6, and determines, for example, whether the heartbeat interval is normal from the R wave peak interval or the P wave peak and R wave peak interval. Whether or not there is a ventricle block is analyzed, and diagnostic support information 300 such as the appearance of arrhythmia, excitatory conduction abnormality in the heart, etc. is output.

Actually, 4529 manufactured by NEC Sanei Co., Ltd.
1270A and 1 manufactured by NEC Sanei Co., Ltd., which are induced using an electrocardiographic electrode and an induction cord such as 8 and 47348.
The ADX-98E manufactured by Canopus, in which the electrocardiogram amplified by an electrocardiogram amplifier such as 253A is incorporated into a personal computer such as PC-9800 series manufactured by NEC Corporation.
By performing A / D conversion with an A / D converter such as the above, electrocardiogram time series data can be obtained. With respect to the obtained electrocardiogram time series data, the above-described processing is operated as a program for a personal computer such as PC-9800 series manufactured by NEC Corporation, and the diagnostic support information 300 is CR.
By outputting to T or the like, the electrocardiogram analyzer of the present invention is realized.

As an actual example of the division point detection processing described above, FIG. 3 shows an example in which the R wave peak, the Q wave start point, and the T wave end point are detected in the actual II-lead electrocardiogram data. In this detection example, electrocardiogram time series data sampled at 300 Hz is used, and the above-mentioned Xi, Yj, R
For k and Bh, a = 301, b = 11, C ₁ = 1,
C ₂ = -1, m = 5, w = 11, TH _{V b} = 0.000
3, TH _{V p} = 0.2, TH _p = 0.5 mV, TH _{Q b}
= 0.9, TH _{T e} = 0.3, RT = 200 msec.

[0042]

EFFECTS OF THE INVENTION By using the present invention, segmented points can be detected with the same degree of accuracy as in the case of using thoracic lead ECG data even with only limb lead ECG data, and ECG analysis can be performed.

Further, when utilizing the positional relationship of the division points for diagnosis, for example, when examining arrhythmia or excitatory conduction abnormality, it is possible to eliminate the need to attach electrodes to the chest and reduce the burden on the subject. .

Further, when there is a high possibility that the psychological burden at the time of measurement such as stress measurement will affect the data, the constraint at the time of measurement can be reduced and the measurement itself can be suppressed from becoming a cause of fluctuation. .

[Brief description of drawings]

FIG. 1 is a diagram showing a basic configuration of an embodiment of an electrocardiogram analysis apparatus of the present invention.

FIG. 2 is a diagram showing division points of an electrocardiogram (see Japanese Medical Association: ABC of Electrocardiogram, Kyowa Planning Communication, 1992).

FIG. 3 is a diagram exemplifying a result of detecting an electrocardiogram segment point using the electrocardiogram analysis apparatus of the present invention.

[Explanation of symbols]

 1 Input data storage unit 2 Waveform enhancement unit 3 Flexibility calculation unit 4 Baseline discriminant amount calculation unit 5 Characteristic amount storage unit 6 Classification point detection unit 7 Analysis unit 100 Electrocardiogram time series data 101 Enhanced waveform time series data 102 Flexibility 103 Baseline Discrimination amount 200 Classification point information 300 Diagnosis support information

Claims (1)

(57) [Claims] 1. An electrocardiogram analyzer for detecting electrocardiogram segment points and analyzing an electrocardiogram, wherein electrocardiogram time-series data is used as input data, and electrocardiogram time-series data required to calculate a feature amount for segment point detection is stored. In addition to the input data storage unit to output, the waveform emphasis unit for calculating and outputting emphasized waveform time series data emphasizing the characteristics of the electrocardiogram waveform from the electrocardiogram time series data output by the input data storage unit, and the waveform emphasis unit A tortuosity calculation unit that calculates the tortuosity used for peak detection from the emphasized waveform time-series data that is output, and an electrocardiogram time-series data that the input data storage unit outputs to determine whether or not the data is on the baseline portion. A baseline discriminant amount calculating unit that calculates and outputs a baseline discriminant amount, emphasized waveform time series data output by the waveform enhancing unit, and a bend output by the bend degree calculating unit. Degree, the characteristic discriminant storage unit for receiving and storing the baseline discriminant amount output from the baseline discriminant amount calculation unit, the emphasized waveform time-series data, the bending degree, and the baseline discriminant stored in the characteristic discriminant storage unit. Detects the segment point by searching the amount, and receives and analyzes the segment point detection unit that outputs the segment point information such as the position of the segment point and the potential at the segment point, and the segment point information output by the segment point detection unit. And an analysis unit that outputs diagnostic support information.