JPH0271733A - Laryngeal disease diagnosing device - Google Patents

Abstract

PURPOSE:To make it possible to exactly determine whether the voice passage is normal or abnormal by successively performing tomography with the use of a scanner so as to detect the cross-sectional areas of the voice passage, and by calculating acoustic parameters of voices corrected by a voice collector. CONSTITUTION:A scanner 5 is operated so as to carry out a tomography, successively by n times at infinitesimal intervals DELTAx over the voice passage 17 of a body 11 to be tested. Then, thus obtained tomographic data are delivered to an image processing device 6 so as to reconstitute fault images and to calculate cross-sectional areas A1, A2, A3... of the voice passage at a high speed. In parallel with these steps, voices are collected by means of a voice collector 7, and then thus obtained voice signals are converted by a signal converter 8, and are then delivered to the image processing device 6 which therefore calculates acoustic parameters at a high speed. The position of the vocal bands as a sound source is obtained in accordance with thus calculated cross-sectional areas of the voice passage and voice signals. Then, the acoustic characteristic of thus obtained voice is compared with that of the normal voice of a normally healthy person so that the image processing device 6 determines whether the voice is normal or abnormal at a high speed. The result of the processing is stored in a memory device, and the image data are displayed on a display unit 10.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to a laryngeal disease diagnostic device for diagnosing laryngeal diseases such as laryngosis. The present invention relates to a laryngeal disease diagnostic device that can determine the transmission function of the vocal cords and correctly determine whether the vocal cords are normal or abnormal.

[Conventional technology]

Laryngopathy is one of the laryngeal diseases. Among laryngophores, vocal cord cancer that occurs in the vocal cords causes voice disorders when it progresses to a certain extent. Focusing on this, it has been proposed to extract the presence or absence of a voice disorder from voice signals collected from a subject and use it for diagnosis of cancer. Many such studies have been carried out, and it is said that by extracting several acoustic parameters originating from the vocal cords from speech, it is possible to distinguish between normal speech and impaired speech. The above acoustic parameters are (
1) The average value of the pitch frequency of the voice (2) The fluctuation of the pitch frequency (3) The noise component contained in the vocal cord sound source (4) The outline and fluctuation of the frequency spectrum of the voice are handled. Here, as a model of a speech generation system, a system as shown in FIG. 8 is generally considered. That is, exhaled air within a human body becomes a sound wave due to the vibration of the vocal cords 1, is articulated through a kind of acoustic tube called the vocal tract 2, and is then emitted as sound by being radiated from the mouth $3.

In order to detect laryngeal cancer early, mass screening can be considered as in the case of other cancers, but since the incidence of laryngeal cancer is low, it is possible to efficiently detect laryngeal cancer by using a machine to treat a large number of people in a short time. A medical examination is necessary. In order to implement such a medical examination, it is necessary to collect the sounds emitted by the subject, process the signals, and extract the acoustic parameters described above.

Therefore, in addition to a voice collector such as a microphone, a digital signal processing device consisting of, for example, an A/D converter, a memory, a computing unit, etc., is required as a dedicated device. Furthermore, general-purpose computers are not sufficient for processing audio data from a large number of people in a short period of time, and a dedicated high-speed computer is required.

[Problem to be solved by the invention]

However, the conventional way of screening for laryngeal cancer requires a digital signal processing device as a dedicated device as mentioned above, and a dedicated high-speed computer, which is expensive and a big burden. Ta. Further, in the digital signal processing device described above, the above-mentioned (1) to (
In order to extract the acoustic parameters of 4) from the voice, the 8th
The sound emitted from the lips 3 shown in the figure is collected by, for example, a microphone to determine the normality or abnormality of the vocal cords 1, which is the sound source. Since the acoustic resonance characteristics of No. 2 are modeled, the characteristics of each patient cannot be fully taken into account, and the determination of whether the vocal cords 1 as a sound source are normal or abnormal is inaccurate. Therefore, accurate diagnostic information for laryngeal cancer cannot be obtained, making diagnosis difficult.
Diagnostic efficiency was also reduced.

Therefore, an object of the present invention is to provide a laryngeal disease diagnostic device that can solve these problems.

[Means to solve the problem]

In order to achieve the above object, the laryngeal disease diagnostic apparatus according to the present invention includes an X-ray tube and an X-ray tube around the opening where the subject is placed.
A scanner, which is arranged to face a line detector and rotates both of the above around the subject to take continuous tomographic images of the throat at minute intervals, and a tomographic image of the throat is reconstructed from the imaging data obtained by this scanner. an image processing device, a voice collector that collects the sounds emitted by the subject, and a voice signal collected by the voice collector that converts the voice signals so that they can be processed by the image processing device and sends them to the image processing device. It is characterized by comprising a signal conversion device, a storage device that stores processing results from the image processing device, and a display device that displays image data from the image processing device.

(Operation) The laryngeal disease diagnostic device configured as described above uses a scanner to continuously take tomographic images of the throat of a subject at minute intervals.
A sound collector collects the sounds emitted by the subject, and a signal converter converts the collected sound signals so that they can be processed by an image processing device. It operates to reconstruct a tomographic image of , process the signals sent from the signal conversion device, store the processing results in the storage device, and display the image data on the display device. This measures the cross-sectional area of the vocal tract of the subject, determines the transfer function of the vocal tract, and accurately determines whether the vocal cords are normal or abnormal.

〔Example〕

Embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

FIG. 1 is a block diagram showing an embodiment of a laryngeal disease diagnostic apparatus according to the present invention. This laryngeal disease diagnostic device diagnoses laryngeal diseases such as laryngeal cancer using a CT scanner, and as shown in FIG.
, an audio collector 7 , a signal converter 8 , a storage device 9 , and a display device 10 .

The scanner 5 is used to continuously take tomographic images of the throat of the subject 11 at minute intervals, and an X-ray tube 13 and an X-ray detector 14 are arranged facing each other around an opening 12 in which the subject 11 is placed. 13 and 14 are rotated around the subject 11. The image processing device 6 inputs the imaging data obtained by the scanner 5 and reconstructs a tomographic image of the throat region of the subject 11 from the oral cavity 15 to the vocal cords 16 as shown in FIG. Signal conversion device i described below
! It inputs and processes the signal sent from A.
It consists of a /D converter, memory, arithmetic unit, etc. The voice collector 7 collects the voice emitted by the subject 11 and is composed of, for example, a microphone. The signal converter 8 converts the audio signal collected by the audio collector 7 so that it can be processed by the image processor 6, and sends it to the image processor 6, for example, an amplifier, a filter, and an A/D. It consists of converters, etc. The storage device 9 inputs and stores processing results such as acoustic parameters and determination results outputted from the image processing device 6, and is made of, for example, a magnetic tape or a magnetic disk. Furthermore, the display device 10 inputs the image data output from the image processing device 6 and displays the image of the subject 11 as shown in FIG.
It displays tomographic images of the throat and frequency spectrum analysis of voices, and consists of, for example, a television monitor.

And in the whole of such a device - XNjAc T
configuring the device.

Next, an outline of the operation of the laryngeal disease diagnostic apparatus configured as described above will be explained. First, the scanner 5 shown in FIG.
The oral cavity 15 of the subject 11 is driven as shown in FIG.
Tomographic images are taken successively n times at minute intervals of ΔX over the entire throat region from the to the vocal cords 16, that is, the vocal tract 17. Next, Δ of the vocal tract 17 obtained in this way
The n pieces of photographic data at intervals of X are input to the image processing device 6,
This image processing device 6 reconstructs a tomographic image of the throat of the subject 11 as shown in FIG. calculate. In parallel with this, the voice emitted from the subject 11 is collected by the voice collector 7, and the collected voice signal is converted by the signal converter 8 and sent to the image processing device 6. Next, the image processing device 6 uses its internal arithmetic unit to calculate the acoustic parameters (1) to (4) described above at high speed. Then, the position of the vocal cord 16 as the sound source is determined from the vocal tract cross-sectional area A 1+ A 2 H...+An calculated as above and the input audio signal.

Next, the acoustic characteristics of the voice of the subject 11 obtained in this way are compared with the acoustic characteristics of the normal voice of a healthy person recorded in advance by pattern matching, etc. The image processing device 6 quickly determines whether the image is normal or abnormal. The acoustic parameters and judgment results obtained by the image processing device 6 are stored in the storage device 9, and the tomographic image of the throat of the subject 11 and the frequency spectrum analysis of the voice are displayed on the screen of the display device 10. Is displayed.

Next, in the above operation, the vocal tract cross-sectional area A1.

A2. ...Calculating the vocal cords 16 as a sound source using tAn will be explained. As shown in FIG. 2, n vocal tract cross-sectional areas Al obtained at minute intervals ΔX.

The vocal tract 17 from Az, .

The sound waves generated by the vibration of the vocal cords 16 are transmitted to the vocal tract 17.
When propagating inside the acoustic tube, it can be assumed as a plane wave in a lossless acoustic tube. Therefore, from the lip 18 shown in FIG.
The propagation of the sound wave between the th section and the (m+1)th section is expanded and considered as shown in FIG.

In FIG. 4, it is assumed that the vocal tract cross-sectional area is constant within the minute interval ΔX, so the vocal tract cross-sectional area of the m-th section is denoted as A, and that of the (m+1)-th section is denoted as A@+t. The length of each section above is ΔX, and the sound wave is Δx /
Let Δt be the time required to propagate the length of 2. Then, the wave traveling from the vocal cords 16 to the lips 18 is defined as a forward wave,
The wave in the opposite direction is called a backward wave, and the mth interval and (m+1)th interval
Let t be the time at the boundary of the section. Now, a certain section n
The volume flow velocity of the forward wave at is expressed by the relationship fn(t),
If the volume flow velocity of the backward wave is expressed by the function bn(t), then the (mth)
The forward waves of the sound waves in the +1) section and the m-th section are.

Each is expressed as f net (-Δt) + fjt+Δt), and the backward waves are respectively -b, ÷□(t+Δ,).

-b, (is expressed as -Δt). And the (m+1)th
At the boundary between the section and the m-th section, the following equation holds based on the continuity condition that the volumetric flow velocities of the sound waves are equal to each other and the sound pressures are also equal.

fm”t(t-Δt)+bIl+1(L+Δt)=f,
(t+Δt)+b, (−Δt) ... (1) ρC
/ Am"1(f m"i(t-Δt) bs"t(
t+Δ1))=ρC/A11(f, (−Δt)−bs
(t+Δ1))...(2) In equation (2), ρ is the air density and C is the speed of sound.

Next, the boundary conditions at the vocal cords 16 are as shown in FIG. In Figure 5, the vocal tract cross-sectional area of the nth section is An
Let the cross-sectional area of the acoustic tube on the back side be A n +1. Moreover, if the volume flow velocity of the forward wave in the vocal cords 16 is expressed by the function i g (t), then the above equations (1) and (2)
Similarly to the formula, the following formula holds true.

1t(t−Δt)+bnet(t+
Δt)=f,,(t+Δt)+bn(t−Δt)
・(3)p C/ An+t(-b net(t+
Δ1))=p C/An(fn(t+Δt)−bn(
t-A1))...(4) Furthermore, the radiation characteristics of the sound waves from the lips 18 are as shown in FIG. In FIG. 6, the cross-sectional area of the vocal tract in the first section is designated as A1, and assuming that an appropriate acoustic tube is connected to the outside of the lips 18, its cross-sectional area is designated as Ao. At this time as well, the following equation holds true in the same manner as the above equations (1) and (2).

f x(t-Δt)+bt(t+Δt)=fo(t+Δt)
t)...(5) ρC/A*(ft(t-Δ) bl(t+Δ1)
)= p C/Ao(f o(t+Δ1))...(
6) Now, if the audio waves in each section are sampled at a period of τ=2Δt=ΔX/C, it can be expressed as follows. In addition, k=o, 1゜2,・
It is... This equation (7) will be simplified and expressed as follows.

Note that k=o, 1, 2. ...is...

Here, the reflection coefficient R1 at the boundary between the mth section and the (m+1)th section shown in FIG. 4 is further calculated by Z-transforming f, (k), and bn(k).

f, (k)-”F, (Z), bn(k)-B, (Z)
・Represented by (13), fn(k 1), bn(k 1
), 1ll(k-1) are Z-transformed, respectively, and defined as A a + A m '' 1.Then, the speech waves in each section are defined as τ=2Δt;
When sampling at a period of Δx/C, the above (1st
) to (6) become the following equations (10) to (12) using equations (7) to (9), respectively.

...(10) ...(11) The above equations (10) to (12) become the following equations (15) to (17), respectively.

...(15) ...(16) When the above equations (15) to (17) are illustrated in the form of a signal flow, it becomes as shown in FIG.

Here, equation (15) corresponds to block 7b in the figure, equation (16) corresponds to block 7a, and equation (17) corresponds to block 7a.
The formula corresponds to block 7c. From FIG. 7, if the output waveform Fo(Z) of the sampled speech wave is known, the reflection coefficients R,,R, can be determined by equation (9), so the input of the speech wave from the vocal cords 16 as the sound source is The waveform xg(z) is found.

Through the above operations, the vocal tract cross-sectional area AI obtained by continuous tomography using the scanner 5, A2. ...gAr.

and the audio separately collected by the audio collector 7,
The input waveform rg(z
) is determined and compared with a separately recorded normal vocal cord sound source, it is possible to correctly determine whether the vocal cords 16 of the subject 11 are normal or abnormal.

〔Effect of the invention〕

Since the present invention is configured as described above, the vocal tract cross-sectional area A obtained by continuous tomography using the scanner 5! ,Ax,・
. . , An and the sound separately collected by the sound collector 7, it is possible to obtain the input waveform of the sound wave from the vocal cords 16 as the sound source.

Therefore, the acoustic resonance characteristics of the vocal tract (2), which is the transmission characteristic of the voice generation system shown in FIG. You can sense the source of the sound. From this, by comparing with a separately recorded normal vocal cord sound source, it is possible to correctly determine whether the vocal cords 16 of the subject 11 are normal or abnormal. Therefore, for example, accurate diagnostic information on the laryngeal tube can be obtained, diagnosis can be made easier, and diagnostic efficiency can be improved.

[Brief explanation of the drawing]

Fig. 1 is a block diagram showing an embodiment of the laryngeal disease diagnostic device according to the present invention, Fig. 2 is an explanatory diagram showing tomographic images of the throat obtained by continuous tomography using a scanner, and Fig. 3 is a microscopic diagram. An explanatory diagram schematically showing a vocal tract approximated by a cascade of cylinders, Figure 4 shows the m-th section and (m+1)th section in Figure 3.
An enlarged explanatory diagram showing the continuity of the propagation of sound waves between sections, Fig. 5 is an explanatory diagram showing the boundary conditions in the vocal cords, Fig. 6 is an explanatory diagram showing the radiation characteristics of sound waves from the lips, and Fig. 7 8 is a signal flow diagram showing the flow of signals according to the acoustic tube model of the vocal tract, and FIG. 8 is an explanatory diagram showing the model of the voice generation system. 5...Scanner, 6...Image processing bag, 7...Audio collector, 8...Signal converter, 9...Storage device,
DESCRIPTION OF SYMBOLS 10... Display device, 11... Subject, 12... Opening part, 13... X-ray tube, 14... X-ray detector, 15
... Oral cavity, 16... Vocal cords, 17... Vocal tract, 18.
··lip. 1 vocal chord 3 figure high 4-concave A 52nd high school 6th ward A.

Claims (1)

[Claims] 1. A scanner in which an X-ray tube and an X-ray detector are arranged facing each other around an opening in which a subject is placed, and both are rotated around the subject to take continuous tomographic images of the throat at minute intervals. ,
An image processing device that reconstructs a tomographic image of the throat from imaging data obtained by the scanner, a voice collector that collects the sounds emitted by the subject, and an audio signal collected by the voice collector that is transmitted to the image processing device. a signal conversion device that converts the signal so that it can be processed by the image processing device and sends it to the image processing device; a storage device that stores the processing results from the image processing device; and a display device that displays the image data from the image processing device. A laryngeal disease diagnostic device comprising: