JPS5944241A - Apparatus for measuring pulse of embryo - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a fetal heartbeat detection device, and particularly to a device for detecting the pulsatility of the cardiovascular system in order to verify the survival of a fetus in early pregnancy. In particular, the present invention applies an ultrasonic imaging device, obtains an ultrasonic image captured in real time at high speed in the form of digital video data via an electric signal, and detects a moving object between the frames. We also aim to detect υ periodic beats by identifying the regularity of their movements. Such a system and device is useful for detecting and identifying fetal heartbeats by observing the ultrasound image with the naked eye and discovering its pulsatile components. Alternatively, it has the feature that it can be configured so that the heart rate rate can be determined immediately.

Traditionally, when examining a woman who may be pregnant,
In other words, it is considered that there is no better method than the ultrasonic method as a method that allows testing results to be obtained on the spot. Biochemical methods such as the HGC test are highly sensitive but depend on other factors)
One of the points of resistance is that it may give a false positive result, and samples must still be collected from the test subject. As for ultrasound, there was a time when CW Doppler equipment was touted as a fetal detector, but these days, pulse-echo imaging equipment, especially real-time B-scan equipment, has become the mainstay. This is because, although the detection principle is different, it is CW Doppler.

For both pulse echo methods, the directional gain and the ultimate S/N under the ultrasound irradiation amount with the same energy density
However, while there is not much difference in detectability, the practical detection limit (the number of weeks of pregnancy that should be detected in most cases) is 8.5 to 9 weeks for CW Doppler and 6 weeks for electronic linear B-scan. It's wide open. The reason for this is that the CW Doppler method can only be used by ``bringing the fetal heart into the sensitive area in front of the probe using a blind eye while being able to hear,'' and therefore, it is difficult to use a device with too sharp directivity. The highest sensitivity cannot be achieved (
The latter has the disadvantage that the directivity gain cannot be maximized), whereas the latter can bring the directivity gain close to the theoretical maximum value (this is necessary to improve resolution),
Fortunately, even if the noise figure of the system is not the best, the output of the device can appeal well to human vision.
``ID is much more advantageous for searching for a fetus. These differences in conditions lead to the differences in the field of practical re-optometry mentioned above.

By the way, once a fetal heartbeat has been detected and identified on a B-Cucco-1 image, in most cases, the same transducer can be used in M mode (T
Recording is also performed in M mode).

Many devices have a mechanism to do this either in a switching manner or simultaneously, and also have a mechanism to adjust the sound rays in M mode so that they cross the fetal heart. 1. Reasons for recording in M mode This is because this cannot be used as a record that can objectively prove the pulsation of the fetal heart in question, and even if the real-time B-scope image was taken as a methyl photograph, it cannot be used as such evidence. There is also a method of converting the image into a TV format video signal in some way and storing the result in a VTR, but this method is too exaggerated and is not practical.

Regarding such an M-mode ultrasound system, even if it is implemented in real time, it is very difficult to discover the fetal heart by itself. Therefore, discovery and identification are only possible using the real-time B scope. In order to leave evidence behind, there is a division of labor between recording in M mode. However, unlike the M-mode of adults, M-mode information of fetuses, especially those of a young fetus, is rarely used beyond evidence that the fetus has pulsatility or is alive.

Therefore, the present invention was made in view of these points,
The purpose is to semi-automate the step 1 of discovery 1 above, and on the other hand, to semi-automatically maintain the fetal heart without losing sight of the fetal heart from the ultrasound image containing the fetal heart once captured. The method is to calculate the fetal heart rate from the pulsatility information obtained using a more advanced method.
The purpose is to display these values and record them continuously to draw a fetal heart rate diagram.

FIG. 1 is a block diagram showing an embodiment of the present invention, in which the upper portion surrounded by a dotted line is a known electronic real-time linear B-scan type ultrasound imaging device, in particular, Still image retention (Freeze/
It is not necessary (although it may be somewhat advantageous to have one) to have a frame memory for frames (frames) or for scan conversion to TV format. It should be noted that the parts indicated by dotted lines in the figure are shown very schematically, and the gist of the present invention is not restricted thereby. All that is required here is that the scan be performed regularly (at a fairly constant frame rate) so that its frame and line synchronization signals can be extracted. That is, block 2 at the bottom of Figure 1
0-1 are supplied with a video signal and both synchronization signals. This video signal does not have to be obtained from the output of the video amplifier 17, but may be obtained from the detector 16 or from the high frequency (echo itself) signal before detection, as long as it has the same meaning, that is, the echo intensity. On the other hand, it is also possible to perform A/D conversion within the upper block. Furthermore, if there is a scan conversion system in the upper block, it is also possible to implement it by using the line rate and frame rate converted to the standard TV system, rather than the line rate and frame rate based on the unique transmission and reception rate of the ultrasound system. . There is no essential difference in either case.

Therefore, the lower block 20 in FIG. 1 will be explained.

This is a kind of video MTI system, which extracts uncorrelated components between frames to obtain an image of a moving object. For this purpose, a video MT system is established in which a frame memory (buffer memory) for a certain number of frames is provided and a digital filter (sample value filter) is used between mutually corresponding pixels. This type of technique has already been known and applied in radar and other applications for a long time, so there are many similar techniques that can be used here. For example, if uniform noise is not suppressed, it will be included in the MTr output as an uncorrelated component, so it is better to apply some smoothing (not only in the X+Y direction within a frame but also between frames). Or, it is better to perform non-linear processing (zero clip, etc.) on the difference value (I won't go into those details here).

The frame rate of a typical electronic linear B-scan is approximately 10 to 30 frames per second. In contrast, the heart rate of a fetus in the early stages is 180 to 2 JOBPM (beats per minute).
even then, only a few frames can be effectively imaged during each cardiac cycle. This may seem like a difficulty, but on the contrary, the amount of data that must be processed to test for periodicity can be reduced, which can be advantageous in some cases. What should be noted is that in this case, what is processed is not a complex waveform such as fetal heart sounds or fetal Doppler signals, but rather a basic periodic signal similar to that representing the reciprocating stroke of the fetal heart itself. (similar signals include waveforms of pulse waves, blood pressure, blood flow, etc.).

Anyway, video MTI processing is done between frames.

This requires a dedicated high-speed microprocessor closely coupled to each frame memory and another frame memory (output buffer) to hold the processing results. M
If you continuously read out the location (address) of a place in the TI output buffer that exhibits particularly characteristic pulsatility, you will get the above 1.
A video MTI system for such a purpose is constructed by a video MTI system in which a fundamental period signal is obtained and the period can be immediately measured, and thus the heart rate can be measured (calculated). It is not necessary to perform the scanning over the entire screen being imaged by the scanning device. All you have to do is bring what you want to see to the center of the screen IA, set a moderately sized attention area IAI around that area, and perform processing within that area (as shown in FIG. 2).

In this case, it is not to say that images outside this area are of no use at all; on the contrary, they have the effect of providing great advice in visually determining whether or not the target region has been accurately captured. That is, the operator specifies the position and size of the region of interest while looking at the B-mode echogram displayed below. Q, adjust the position and angle of the entire probe so that the part of interest is within this area of interest within the scanning sound field. In the latter case, this attention area may be relatively fixed on the screen as a function. However, whenever the function of the caliper is used, it is preferable to be able to set this area of interest with some degree of freedom, and to be able to move at least the center position.

Therefore, once such a region of interest is set and it appears that the target pulsatile echo source is located within it, image data of this region is generated for each frame. -12
Temporarily store a dataset with a dimensional structure. This situation is shown in Figure 3. In other words, there will be dozens of such two-dimensional data sets from several seconds in the past.

Problem 1 is reduced to searching for an element with periodicity in the direction perpendicular to the surface from among the data set DS (or data grain) of one large number of sheets. Therefore, the processing that constitutes the inter-plane difference or higher-order change extraction filter is performed through the same pixel address of these data planes. This situation is shown in Figure 4. The filter used (this As the impulse response of a so-called paired sample point sequence filter (or what is commonly called a digital filter), the simplest impulse response may be a differential 1 substituted with a so-called 1 difference as shown in Figure 5 (a); It is also possible to use a method like the one shown in the same figure (b), which is a quadratic extension of The heart rate of the fetus is already as high as K 1ao BpM (beats per minute) by the time it is observed using the CW Doppler method (it then decreases as the baby grows, and by the time of delivery it is around 140 BPM). In an even younger fetus (also called an embryo), 5
~120 immediately after starting heartbeat around 6 weeks
It is generally known that there is only about BPM.

This is thought to be due to the fact that the heart (myocardial muscle) is formed, but only its spontaneous beating (ventricular condition) is observed, and the control system is not configured, as is the usual course of development. In other words, it is thought that the degree of completeness increases as the atrioventricular nodal condition or sinus node condition is obtained, just at 8, 5 to 9 weeks, when it can be detected by the CW Doppler method.

Therefore, one cycle 1 of the digital filter in the time axis direction that matches the fetal heartbeat cycle is 120 to 1sa.
]) It is preferable to have a variable range within a range that satisfies the PM, or one that broadly covers this range and has a short impulse response (at most, the shape shown in Figure 5 (c)).

In the process of performing filter processing on the time axis for each pixel address, the resulting energy (i.e., the square of the amplitude) is determined, and in order to find a noteworthy peak regarding that value, pixel address P1 is set. If you can find it, you can assume that you have found a fetal heartbeat at that location.

Such a pixel address (or even a small area)
When the amplitude after the filter is scaled on the time axis, it becomes as shown in FIG. If you know the value of this cycle and find its reciprocal, you can find the fetal heart rate. Since there is no need to worry about the instantaneous variability of heart rate values at this stage, it may be possible to measure the period by summing up the period of several beats. Of course, the period may also be determined by autocorrelation, Fourier transform, spectrum analysis, etc.

In addition, when using the autocorrelation method, it is actually easier to determine the heartbeat cycle using the method shown in FIG. Regarding the method of recognizing heart rate period or heart rate value using the autocorrelation method, US put 5,991,565 or Patent No. 1 by the same inventors
No. 065159 may be applied as the basic technical idea.

As a result of these procedures, frames (data plane) in which moving targets are selectively extracted by differential or higher-order filtering (that is, this is nothing but the so-called MTI K in the video domain) are processed moment by moment,
It can be displayed next to the original B-mode image, or it can be displayed in a different color superimposed on the original image.

Furthermore, once the heart rate is determined, it is preferable to display it numerically somewhere on the screen.

If the extracted area of interest 1 (Fig. 2) is extremely large, the number of pixels will be large and it will take time to process, but this can be reduced to the practical limit as much as possible (for example, 16 x 16 pixels).
If it is small, it can be followed in real time! It can be considered as a thing/. This is 8×8~+O in terms of actual tomographic plane.
It corresponds to about x10mm.

In reality, the so-called G, S, (Ge 5tation
When the image of Sue) is seen, this area of interest can be set to surround it. In practical terms, if one frame (data plane) for several past pictures is stored in the memory, a video MTI with processing contents as shown in Figures 5 (a) and (b) can be constructed, so this method can be used. The simplest system based on this would be as long as it has a buffer for processing work, a frame for accommodating the MTI processing result, and a means for reading it out moment by moment to a display means (coexisting in one system). (i.e., it can run in real time) if it can be stored in the CPU.

However, the modification method described below makes it possible to significantly simplify the apparatus, although it causes some troublesome operation and a decrease in the probability of detection. That is, the first modification is to further reduce the number of pixels involved even when applying the video MTI system to the selectively extracted region of interest (FIG. 2). The cell is divided into sections as shown in the figure, and MTI is performed sequentially on each section, so that the cell with the highest pulsatility (the amplitude of the filter output) is examined more precisely. However, when this method is performed in real time, it is rather difficult to use because the probe must be held so as to obtain a constant screen.

Therefore, it is useful as a measure to improve the processing speed when data for several seconds is first taken into the working memory and then processed later. This method is especially useful when the processing is not done with a dedicated device, such as an external desktop computer, since it is difficult to simultaneously perform data transfer and work below MTI processing.

If you want to describe the processing content in an ink interpreter level language and execute it immediately, you will be concerned about the slowness of the interpreter (5 to 200 times slower than assembler), so first write pixels of video values with periodicity. We should consider the amount of labor and processing required to make one discovery. The processing procedure after successful discovery can be considered to be the same as in the previous embodiment.

However, if further labor saving is desired, the following method is possible. In other words, attach at most one pixel to the MTI, and place it at the exact location of the caliper or cursor mark (of course, you can use another hand), or by adding some pixels around it as you like. The values extracted using the two-dimensional evaluation function are obtained for each frame repetition period and are applied to the sample motion filter,
Alternatively, D/A conversion is performed before or after sample and hold, and the fundamental wave component of the periodicity of the heartbeat is extracted using an analog filter as an analog value. This method is to find it using the time method. This method is characterized by the fact that it is only necessary to process the signal waveform of one channel, whether the sample sequence of pixel values is obtained and the subsequent processing is done using a digital method or an analog method. It is advantageous that the processing can utilize general pulse wave heart rate metering, electrocardiogram, heart sound, or Doppler heart rate metering techniques.

The shape of the cursor or caliper mark to display the extraction position of the desired sample value may be something like a well-known crosshair; however, when the crosshair is brought to the target location, the image of the target object is Since they are often crushed and cannot be seen, it is better to use a square or round mark, or a spread crosshair with the area near the intersection removed (Figure 8).

It should be noted that this pixel(s) extraction method can also be implemented in a purely analog imaging system without 1d frame memory. However, there is still another issue,
The opening side of the heartbeat as described here! Strictly speaking, it is virtually impossible to maintain a constant relative relationship between the probe and the living body as if scanning a complete plane (it is well known that this is close).

If the hand holding the probe is unsteady or the body moves due to breathing, false periodic echo images may appear. In order to reduce this as much as possible, a so-called 1-flat type 1 probe (FIG. 9) may be used. It is worn over the abdominal wall (using a belt or bandage) as in a normal fetal monitoring device. This method is suitable for long-term continuous monitoring and monitoring of fetal movements and other factors.

In either case, once the heart rate value is determined moment by moment, the value is digitally written out next to the ultrasound image.
Furthermore, it is used by creating a heart rate diagram and displaying it on the same screen (multiple) (Figures 10 (a) and (b)).

As explained above, according to the present invention, B mode or B mode
By using an ultrasonic imaging device that operates simultaneously in mode and M mode, it is possible to automatically detect elements having periodic pulsation components in an image, and to determine the period thereof and display it.

Therefore, if the target element is the fetal heartbeat, its period or heartbeat value can be determined and the heartbeat value can be easily displayed numerically. Continuous monitoring is also possible,
It is also easy to include a fetal heart rate diagram. In these cases, it is of course possible to automatically display areas with pulsations.

[Brief explanation of the drawing]

Fig. 1 is a block diagram showing an embodiment of the present invention, Fig. 2 is a diagram explaining a region of interest processed by a video MTI system, and Fig. 3 shows the storage state of image data for each frame of the region of interest. Fig. 4 is a data brain diagram for explaining the process of forming a filter for extracting differences between surfaces or higher-order changes; Fig. 5 is a diagram showing the impulse response of the filter used; Fig. 6. Figure 7 is a diagram showing how the fetal heart rate is determined by filter processing, Figure 7 is an explanatory diagram when the region of interest is divided into sections, Figure 8 is a diagram showing the shape of the cursor or caliper mark, and Figure 9 is FIG. 10 is a block diagram of a flat probe suitable for the device of the present invention, and is a diagram showing an example of display. 11... Probe, 15... Receiving amplifier, 16...
Detector, 17... Video amplifier, 19... CRT display, 21... AD converter, 23... Image buffer memory, 24... Two-dimensional MTI filter,
26...MTI video amplifier, 27...Extraction circuit, 28...Rate meter. /knee

Claims (2)

[Claims] (1) A process of capturing a real-time B-mode ultrasound image, and a process of extracting inter-frame differences or higher-order changes regarding a point of interest in screen content and its vicinity at the stage of a video signal of the ultrasound image; In this process, the periodicity of the change in the local area that is recognized to have a periodicity is examined, and as a result, the M-mode ultrasound is applied to the sound ray that includes the local area that is recognized to have a periodic change. Fetal heart rate measurement system using a B-mode ultrasound imaging device
3, (2) The fetal heart rate measuring device according to claim 1, wherein the B-mode ultrasound imaging device is capable of displaying M mode and B mode simultaneously. (51) The fetal heart rate measuring device according to claim 1, wherein the B-mode ultrasound imaging device is capable of displaying a fetal heart rate value or a single column graph of the fetal heart rate.