JPH02232030A - Ophthalmology diagnostic method - Google Patents

Abstract

PURPOSE:To obtain an exact correlation data value by counting a photoelectric detection signal as a photoelectric pulse every prescribed unit sampling time, storing a counted value to a memory over one measuring time, successively calculating a photon correlative function and a time correlation length every unit time for stored data after measurement is finished and comparing the time correlation length with a prescribed reference value. CONSTITUTION:A speckle signal counts the time sequential data of the photoelectric pulse corresponding to the intensity of light through a photon counting unit 52 every prescribed time and the data are successively stored to a memory 57. After the measurement is finished, the data are read from the memory 57 and the photon correlative function is calculated by a correlation device 53. Then, the calculated function is displayed and outputted to a CRT 55 or a printer 56 together with an analyzed and evaluated result. Namely, the pulse is counted every prescribed short sampling time DELTAt and the counted values are stored to the memory as n1, n2, n3-ni-nm. The DELTAt is set to a time to be counted while referring to a value taucmin in a measure considered as shortest out of a time correlation length tauc, which is obtained in the signal to be measured, and enough dividing the value.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to an ophthalmological diagnostic method, in particular, to irradiating a laser beam of a predetermined diameter to the fundus of the eye, and forming a laser speck on an observation surface by scattered reflected light from the biological tissue of the fundus. The fluctuation of the pattern is
This paper relates to an ophthalmological detection method in which the blood flow state of the fundus tissue is measured by detecting speckle light intensity changes through a micro circular detection aperture with a predetermined diameter and determining the photon correlation function. [Prior Art] As a method for measuring vascular blood flow in tissues such as the retina by irradiating the fundus of the eye with a laser beam, there is a method known as “InvestigativeOp.
hthal+mology. Vol1. ll. No.
ll, p. 936.November 1972. “Scien
ce, Vol1. l86. Nov. 29. p. 830
．． In 1974 and other publications, JP-A-55-75668. 55
-75669. Publication No. 55-75670, Japanese Unexamined Patent Publication No. 1983
-142885 (UK 13132/76.USP
4.166.695), JP-A-56-12503
Publication No. 3 (corresponding to British (GB) 79/37799), Japanese Patent Application Publication No. 118730/1983 IUsP 4,402.
601) The laser Doppler method shown in asp4.142.796 is known. However, the Doppler method is currently difficult to put into practical use due to the complexity and precision of the optical system, the complexity of handling, and the instability and uncertainty of the measurement results. In order to solve these problems, the applicant has developed a laser speckle method (for example, Japanese Patent Laid-Open No. 60-199430.60-2032) which has already been applied to skin blood flow measurement.
35. 60-203236 publication or Optic
sLetter Vol. 10, No. 3. 198
March 5th. p. A method has been proposed in which the blood flow state of the eye tissue is evaluated by applying the method (denoted as 104, etc.) to the eye region. These are published in Japanese Patent Application Laid-Open No. 62-275431 (
USP 4.743. l07, EPC2348691,
Japanese Unexamined Patent Publication No. 63-238843 (EPC 28424
8), Japanese Patent Application Laid-open No. 63-242220 (EPC 28)
5 314). In the methods described in these publications, for example, when measuring the fundus of the eye, an image is formed on an optical Fourier transformation surface, a Fraunhofer diffraction surface, or an imaging surface (or magnified imaging surface) conjugate with the fundus. The intensity change of the time-varying speckle pattern is extracted using a detection aperture. [Problems to be Solved by the Invention] When evaluating the blood flow of a specific blood vessel, it is sufficient to detect a speckle pattern on the image plane. Regarding this, the above-mentioned Japanese Patent Application Laid-Open No. 63-242220 (E
PC 285314). In this method, a detection aperture is set above a desired blood vessel image on an enlarged image plane, and a speckle signal is obtained by extracting temporal intensity fluctuations of the image plane speckle pattern there. Therefore, if the fundus image on the detection surface shifts laterally during measurement due to various factors such as eye movement, relative positional deviation between the device and the subject's eye, misalignment, vibration, etc., the blood vessel image being measured may differ from the detection aperture. Situations often occur in which the sensor moves out of position, approaches a blood vessel wall, or another blood vessel next to it enters the detection position. This is a major problem that constantly arises in patients with diseased eyes during clinical practice, especially due to insufficient fixation or fear on the part of the examinee. Therefore, when looking at the correlation function of the signal obtained from one measurement, various components are mixed, making it impossible to measure the correct vascular blood flow stably and with good reproducibility every time. Furthermore, it is extremely difficult and impractical to remove unnecessary components from the correlation function curve data once measured and extract the original vascular blood flow signal. In other words, what ingredients are unnecessary?
This is because it is unclear to what extent it is included and contributes to the final result. Also, if blinking or instantaneous noise-free signals are introduced during measurement, the shape of the measured correlation data may be distorted, making it impossible to evaluate correctly, and one measurement may be difficult due to the mere instantaneous signals. It was a waste and I had to start over again. Therefore, the present invention has been made to solve these problems, and an object of the present invention is to provide an ophthalmologic diagnosis method that can obtain reliable correlation data values and perform accurate ophthalmologic diagnosis. 10 Steps for Solving Problems J In order to solve such problems, in the present invention, the fundus is irradiated with a laser beam of a predetermined diameter, and a laser beam is formed on the observation surface by scattered reflected light from the fundus biological tissue. In an ophthalmological diagnostic method, photoelectric detection is used to photoelectrically detect fluctuations in a laser speckle pattern as speckle light intensity changes through a detection aperture, and measure the blood flow state of the fundus tissue based on the results of measuring the photon correlation function. Detection signals are counted as photoelectron pulses at each predetermined unit sampling time, and the counted value data is stored in a computer memory in chronological order over one measurement time, and after the measurement is completed, the stored data is stored every multiple unit times. Divide into photon correlation function and time correlation length sequentially for each unit time? :C is calculated and the measured data is classified by comparing the time correlation length with a predetermined reference value. [Operation] In the embodiment of the present invention, weak optical speckles from the fundus of the eye are detected as photoelectronic pulse signals, and the values obtained by counting the time-series pulses at each predetermined sampling time are stored in the memory, and after the measurement is completed, the data is re-transmitted. The signals are read out and evaluated, and based on preset criteria, it is determined whether the signal is from a blood vessel, a signal that has strayed into the surrounding tissue, or an ambiguous signal that cannot be said to be either. A final evaluation was made. This makes it possible to evaluate different signal components individually without mixing them, and even when the fundus moves, only the signals from the blood vessels can be extracted, allowing for accurate evaluation and effective use of data without wasting it. By this discrimination, it is possible to evaluate the blood flow state of the capillary region existing in the surrounding tissue and the blood flow state of the choroid layer, making it a very effective method. [Example] The present invention will be described in detail below based on the example shown in the drawings. The present invention targets the ocular region, particularly the fundus, and will be described below using an example in which fundus blood flow is measured using a fundus camera. FIG. 1 is a schematic diagram of the entire apparatus to which the method according to the present invention is applied. For example, red He-Ne (wavelength 632.
8n■) The laser beam from the laser light source I passes through a light intensity adjustment filter 2 for adjusting the light intensity via a condenser lens 1'. Furthermore, relay lens 3,
4 to the fundus illumination optical system of the fundus camera. Further, apertures 5 and 6 are installed between the relay lenses 3 and 4, and are used to select the size and shape of the laser beam irradiation area on the fundus of the eye. Also,
There is a shutter 7 at the exit of the laser light source 1, which opens and closes as necessary. As shown in Fig. 2, the laser beam guided by the relay lens 4 is reflected by a mirror 9 installed in a part of the annular opening 8a of the ring slit 8 in the fundus illumination optical system, and a light beam for fundus observation and photography is directed to the fundus. is guided onto the same optical path that is incident on the . Therefore, the laser beam is reflected by the perforated mirror I2 via the relay lenses 10 and l1, and once focused on the cornea 13a of the eye 13 through the objective lens 13', it is then diffused into the fundus 13b. reached,
Forms an irradiation area that is wider than the blood vessel diameter. This irradiation area is illuminated by an illumination optical system used as a fundus camera to facilitate observation. This observation optical system includes an observation light source 22, a condenser lens 23, and a condenser lens 25 arranged on the same optical axis as the photographing light source 24.
．． It consists of a filter 27 and a mirror 26. Since the laser beam is placed on the same optical path as this observation photographing light beam,
Laser light can be irradiated to a desired position on the fundus 13b by using the left and right, up and down swing mechanisms and fixation guidance mechanism of the fundus camera. Incidentally, since the filter 27 disposed between the condenser lens 25 and the mirror 26 is configured as a wavelength separation filter as shown in FIG. 3, the red component contained in the observation and shadow light is cut. Speckle light generated when the laser light is scattered by blood cells moving in the fundus blood vessels is received again by the objective lens 13', passes through the perforated mirror l2, and reaches the shadow lens 14 and the wavelength separation mirror l5. This wavelength separation mirror 1
Similar to the filter 27, the filter 5 has the spectral characteristics shown in FIG. Most of the speckle light (red) generated by this is reflected. This reflected light passes through the lens 16 once and then passes through the image plane 35.
is further imaged by the objective lens 19a of the microscope optical system 19.
and is magnified through the eyepiece lens 19b. The magnified image passes through the detection aperture 20, is collected again by the condenser lens 2l, and is detected by a photomultiplier tube (photomultiplier) 40. A shutter 40' is arranged in front of the photomultiplier tube 40, and an output signal from the shutter 40' obtained when the shutter is opened is inputted to a signal processing circuit 50. This signal processing circuit 50 includes an amplifier 51 as shown in FIG.
, a photon counting unit 52, a correlator 53, a microcomputer 54, a CRT 55, a printer 56, and a memory 57. On the other hand, the light that has passed through the wavelength separation mirror l5 passes through the relay lens 28, flip-up mirror 29, mirror 30, and reticle 3.
1. It is configured so that it can be observed through an eyepiece lens 33 and photographed with a photographic film 32. In the apparatus configured as described above, first, the power is turned on, a subject is set, the fundus 13b of the subject's eye 13 is observed through the observation optical systems 22 to 26, and the laser light source 1 is activated. At this time, the output level is set to the adjustment level using the light amount adjustment filter 2, the size and shape of the laser irradiation area is set using the apertures 5 and 6, the shutter 7 is opened, and the measurement position is set, and then the observation optical system 28 Check the speckle pattern through ~31. In this embodiment, in order to facilitate laser irradiation, the laser beam irradiation area at the measurement site of the fundus 13b is set to a wider area than the blood vessel, for example,
Since it is set to 3 msφ, there may naturally be cases where a plurality of relatively large blood vessels are included in addition to capillaries. Therefore, in order to measure the blood flow of a specific blood vessel, the blood vessel image to be measured is selected on the enlarged image plane, and the detection aperture 20 is set in the image. That is, a conjugate image of the fundus is formed on the imaging plane 35 in FIG. This is magnified by the objective lens 19a and eyepiece lens 19b of the microscope optical system 19, and a detection aperture 20 is placed on the surface of the magnified image to detect changes in speckle light intensity. The detected light is collected by a condenser lens 2l and converted into a signal by a photomultiplier tube 40 (shutter 40' is open). Photomultiplier tube 4 during measurement
The output from 0 is a speckle signal that fluctuates over time as blood cells move. The speckle signal is amplified by the amplifier 5l in the signal processing circuit 50, and then sent to the photon counting unit 52.
time-series data of photoelectron pulses corresponding to the light intensity is counted at regular time intervals through the 3D controller, and the counted value is stored in the red memory 57. ? After the M determination, the data is read from the memory 57, the photon phase leap number is determined by the phase shifter 53, and the data is displayed on the CRT 55 or printer 56 along with the analysis and evaluation results.
A series of these controls is performed by a microcomputer 54. As described above, in this embodiment, the detection aperture 20 is placed on the magnified image plane, so a desired blood vessel image to be measured in the laser irradiation area is selected, and the detection aperture 20 is placed within that blood vessel image. By adjusting the position of the detection aperture 20 or the fixation of the target eye l3, the blood flow in one specific blood vessel can be measured. As the detection aperture 20, a minute circular aperture such as a bottle hole can be used effectively. For example, as shown in FIG.
When the book 60 is obtained as an enlarged image, the image plane speckle 62 inside the blood vessel fluctuates through a pinhole 6l as shown in FIG. When the speckle crosses the detection aperture 20, the detected light intensity changes and a speckle signal can be obtained. The image plane speckles 62 that are actually observed exhibit a boiling motion characteristic of speckles from a living body due to multiple scattering effects and the like. In other words, even if the blood cells flow and move in a fixed direction, the image plane speckles 62 do not move in a fixed direction according to the flow like a simple image, which is a so-called translational motion, but the individual speckles move in a fixed direction. , without changing the location, randomly flashing light and dark over time,
It has become clear that the speckle pattern as a whole is a type of motion that constantly and randomly fluctuates in intensity. However, in this case as well, the bottle hole 6l
There is no difference in the fact that irregular changes in speckle blinking can be obtained as a speckle signal. If the blood flow is faster, the speed at which the image surface speckle 62 repeatedly flashes between light and dark on the enlarged image will change faster, and the time change of the speckle signal will become faster, so the signal will have more high-frequency components. The signal processing circuit 50 calculates the autocorrelation function of the signal, and evaluates the degree of attenuation based on the correlation time. In that case, for example, if the delay time at which the correlation value becomes l/e (or l/2, etc.) is the correlation time -cc as shown in FIG. The speeds are linearly related. Since the fluctuation speed of the image plane speckle 62 directly reflects the blood flow velocity, the blood flow velocity V is determined from 1/τC.
can be evaluated from the relationship shown in Figure 8. In the signal processing circuit 50, the time-series pulse signal obtained from the photon counting unit 52 is processed as shown in FIG.
is converted into a pulse train signal with a density proportional to the intensity of the speckle signal. Storing this information one by one in memory requires a very large amount of memory, and requires a processing system with high-speed time response, which is impractical. Therefore, in this embodiment, pulses are counted as shown in FIG. 9(C) every predetermined short sampling time Δt, and the counted value is calculated as nl. n2. n3. =-ni・-n Store in memory as m. Therefore, m during one measurement time T
Sampling is performed twice, and m pieces of data are stored. In other words, T=mΔt. It is best to set Δt to a time that can be sufficiently divided and measured, with reference to the value τcmin, which is considered to be the shortest of the time correlation lengths C of the signal to be measured. For example, τcmin=2
If 0 71 sec, Δt≦0.5 ~ 1 u
Set to the value of sec. This is because Δt is the minimum time unit and determines the time resolution of the measurement. Next, Figure 10(a). As shown in (b), one measurement time T
When m pieces of data are stored with = mΔt, T f! :m
/ h equal parts, and each divided time T'=hΔt is divided into unit time Ul. U2. Let U3=Uw/h, Fig. IO (C
), the correlation function is found for each unit in sequence.
At this time, the same delay time Δτ may be set from U1 to Us/h, or the optimum Δτ may be set individually. In this way, the time resolution T' that divides I/h constant T into m/h pieces is
With hΔt, one measurement can be evaluated in a time-sharing manner. However, if T"=hΔt is too short, the correlation function of each unit will not converge sufficiently, so Δt<<T" is necessary.
Also, if the number of divisions is small and <T' is close to T, there is no meaning in time division, so at least T'<T/10 (m/h
210) conditions are preferable. At this time, the number of decompositions m/h can be changed repeatedly, and the optimal number of decompositions can be searched for. If there is clearly abnormal data such as blinking during the process, the unit containing it can be excluded from the evaluation, which is very practical clinically. The correlation function of each unit is the first! When obtained as shown in Figure O (C), the correlation time t Cl. after each smoothing. 1: C2. r.
Find c3-z cm/h. In general, in the case of image plane detection, in the case of signals from blood vessels, the speckles move quickly and τC becomes short, as shown in Figure 11, and in the case of signals from surrounding tissues, the speckles move as shown in Figure 12. late, τC
becomes longer. The threshold value of C for the blood vessel signal is set in advance as V, and the threshold value of C for the surrounding tissue signal is set as t. These data have been statistically examined from multiple human eye data to obtain approximate values. Also, if necessary, set value τV
．． t can be changed. Therefore, the τC (tecl.tec2...-τC■/h) of each unit is sequentially determined by τV and tet, and the units are divided into groups. That is, (i) τC〈
τv. (iilτV≦teC≦τt. (iiiil
Divide into three parts: tt and tc. If the average value of τC is calculated for each group after classification, the average ici in (i) is the value of the blood vessel signal, the average value in (ii) is the value of uncertain data, and the average value in (iii) is the value of the blood vessel signal. This means that, on average, ciii can be evaluated separately from the surrounding tissue signal value. In this way, even if the measurement point deviates from the blood vessel during measurement, or even if it accidentally returns to the blood vessel, we collect and evaluate only the data from when the blood vessel was being measured. Results can be evaluated well, and clinical measurements can be performed efficiently without wasting anything. Furthermore, since the data near the blood vessel wall is divided into the group (ii), it is possible to avoid evaluating the measurement data of slow blood vessels near the blood vessel wall in the group (i). On the other hand, if you want to include such a component in an overly comprehensive evaluation as (i), you can adjust the threshold value τV to a longer value. This is very practical because even data that is clearly abnormal due to blinking or noise among the measured data will not be included in (i) by the above grouping, and can be evaluated after being excluded. The data at this time (iii) may be used to evaluate the blood flow state of the capillary region and choroid.
Overall, almost no data is wasted, and evaluation is possible even if there is movement, so re-measurement is not necessary. Also the first
In Figure 3, T. Instead of calculating the average value of C, all the data of n in the units belonging to (il are joined together to form one The correlation function may be calculated and smoothed to obtain τci.
Perform the same process for (ii) and (iii) and
ii. Find τciii. Furthermore, the T Cl. for each unit obtained in FIG. c2
-... is plotted against the time axis scale as shown in FIG. 14, and the threshold value l/V. 1/τt is also displayed as a line, l/tc>1/τV11/ is a blood vessel signal, and l/τV21/tc>1
/T: By displaying t as an uncertain signal and l/τt≧l/τc as a surrounding tissue signal, for example, the 14th
It is possible to obtain a characteristic diagram similar to the one shown in the figure. Regarding the example in the same figure, it can be inferred that it detached from the blood vessel midway through and returned again. Therefore, the time-series changes shown by the measurement data can be clearly seen, which is very useful for evaluation. Also, a certain data 75 in FIG. 14 is a blood vessel signal, but if this becomes a surrounding tissue signal in a blot like data 75', this data 75 (75''
) is l/τC, ≧1 7 t: v
Therefore, it is difficult to think that only unit U2 was dislodged from the blood vessel, and it can be evaluated that correct measurement was not possible due to blinking and other unnecessary noise. In other words, it is also very useful to be able to judge and consider from Figure 14 whether or not the time-series changes in data are systematic. Also, as a special case of the above example, let ZV=zt.
rc<zv=it and zc>T. It is also possible to evaluate the results by dividing them into two groups, v = τt. In this case, there is no group of uncertain signals and the signal is divided into either (i) or (iii), which simplifies signal processing, but it also has the disadvantage that data near the boundary is likely to cause judgment errors. be. However, this can be ignored if the number of data is large enough. Furthermore, the data in Figure 14 is of great value as data that shows part of the state of eye movement. In other words, it is possible to get an idea of how fast the object is moving, and it is also possible to compare the movement when it leaves the blood vessel and when it returns, which is thought to provide extremely useful information academically. [Effects of the Invention] As explained above, according to the present invention, a photoelectric detection signal is counted as a photoelectron pulse at each predetermined unit sampling time, and the counted value data is chronologically stored in a computer over one measurement period. The data is stored in a memory, and after the measurement is completed, the stored data is divided into multiple unit times, the photon correlation function and the time correlation length τC are sequentially determined for each unit time, and the time correlation length is compared with a predetermined reference value. By doing this, the measured data is classified, so after the measurement is completed, the data is read out again and according to the classification, for example, is it a signal from a blood vessel or a signal when it is missed in the surrounding tissue?
It is possible to easily evaluate whether the signal is ambiguous or unclear, making accurate and reliable ophthalmologic diagnosis possible.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of an ophthalmological apparatus in which the method of the present invention is used, FIG. 2 is a block diagram of a ring slit, and FIG.
The figure is an explanatory diagram showing the spectral characteristics of the filter.
FIG. 5 is a block diagram showing the configuration of the signal processing device, and FIG. 5 is an explanatory diagram showing image plane speckle observed in the plane of the detection aperture. FIG. Figure 7 is a characteristic diagram showing correlation values with respect to delay time, and Figure 8 is a characteristic diagram showing correlation values with respect to delay time.
Characteristic diagrams showing the relationship between speed and correlation value, Figures 9(a) to (
C) is a waveform diagram explaining the sampling of the speckle signal, Figures 10(a) to (e) are signal waveform diagrams explaining the operation of dividing the measurement time and calculating the correlation function, and Figure 11. Figure 12 is a characteristic diagram showing the correlation function between blood vessels and surrounding tissues, Figure 13 is an explanatory diagram showing how to calculate the average of time correlation lengths divided into groups, and Figure 14 is a diagram showing the time correlation function. This is a characteristic diagram showing the reciprocal of length over time. 20...Detection aperture 3 5--Image plane 40...Photomultiplier tube 50...Signal processing device 53...Correlator 54...Microcomputer Fig. 7 *M Hyphasic mJt#, E, if * ile Fig. 8 S Δ'-7 Fushi 椹 {I 1 1 cedar t Fig. 9 Correlation tt, r, i Reimu 11 Fig. 62 Familiarization L accepted parasitic diagram Figure 12 Figure 13

Claims (1)

[Scope of Claims] 1) A laser beam of a predetermined diameter is irradiated to the fundus of the eye, and fluctuations in the laser speckle pattern formed on the observation surface by scattered reflected light from the biological tissues of the fundus are detected as speckle light through a detection aperture. In an ophthalmological diagnostic method that measures the blood flow state of fundus tissue based on the results of photoelectric detection as intensity changes and measurement of the photon correlation function, the photoelectric detection signal is counted as photoelectron pulses at each predetermined unit sampling time, The counted value data is stored in the computer memory in time series over one measurement time, and after the measurement is completed, the stored data is divided into multiple units of time and the photon correlation function and time correlation are sequentially calculated for each unit time. An ophthalmological diagnostic method, characterized in that the measured data is classified by determining the length τc and comparing the time correlation length with a predetermined reference value. 2) The ophthalmological diagnosis method according to claim 1, characterized in that one or two of the reference values are provided to classify the measurement data into two or three types. 3) A blood vessel signal, a surrounding tissue signal, or an unknown uncertain signal is obtained according to the value of each reciprocal of the time correlation length obtained sequentially.
Ophthalmological diagnostic methods described in Section. 4) The ophthalmological diagnosis method according to any one of claims 1 to 3, characterized in that the average value of the time correlation length is calculated for each group of the classified measurement data. 5) The ophthalmological diagnosis method according to any one of claims 1 to 4, characterized in that one photon correlation function is determined for each group of the classified measurement data.