WO2018164361A1

WO2018164361A1 - Nystagmus video test device and method using infrared camera

Info

Publication number: WO2018164361A1
Application number: PCT/KR2018/000114
Authority: WO
Inventors: 남윤영; 공영선
Original assignee: 순천향대학교 산학협력단
Priority date: 2017-03-06
Filing date: 2018-01-03
Publication date: 2018-09-13
Also published as: KR101905716B1; KR20180101816A; CN110325104A; CN110325104B

Abstract

The present invention relates to a nystagmus video test device and method and, more specifically, to a nystagmus video test device and method using an infrared camera, the device and the method rotating a patient to the left and right so as to influence a vestibular organ of the patient, and analyzing, on the basis of the rotation condition of the patient, an infrared pupil image captured by the infrared camera, so as to perform a nystagmus test on the patient.

Description

Apparatus and method for checking video nystagmus using infrared camera

The present invention relates to a video nystagmus testing apparatus and method, and more particularly, to rotate the patient from side to side to affect the vestibular organ of the patient and analyze the infrared pupil image taken by the infrared camera based on the rotation conditions of the patient. The present invention relates to a video nystagmus test apparatus and method using an infrared camera to perform a nystagmus test on a patient.

Dizziness is so common that there are statistics that one in five adults experienced more than once a year.

Dizziness usually occurs when an abnormality occurs in the ear (peripheral), brain (central), heart, eyes, and the like.

One of the abilities of a person is the ability to change the field of view according to the head's fluctuations when a head shake occurs. In our bodies, Vestibular-Ocular Reflex (VOR) is produced to prevent the field of view from changing with head shaking.

Vestibular reflection helps to fix the eye by moving the eye in the opposite direction of the head's rotation.

Compensatory Eye Movement is induced by sensing angular acceleration in the semicircular vestibular system and called nystagmus.

Vestibular dysfunction, or nystagmus, is a disorder in which the target escapes from the center and causes shaking of the field of vision.

Conventionally, electro-nystagmography (ENG), sclera search coil system (SSCS), and the like have been used to detect eye movement signals for diagnosing such nystagmus.

The electroencephalography method consists of attaching an electrode to the skin around the eye and amplifying the depolarization generated during eye movement with an electrical signal to graph the amplitude of the eye movement over time.

However, since electrospinography is not a method of capturing signals directly from the external eye muscles, it does not provide accurate eye position information. ㅇ There was a problem in that the two eye movements could not be distinguished due to overlapping electric signals from the lower rectus muscle.

In addition, electrostable recording method is sensitive to changes in electrode attachment state or changes in skin resistance due to sweat, etc. according to the movement of the patient, and thus requires frequent calibration of the device. There was a problem that the difficulty in the examination in patients not good.

The sclera search coil system is a method of three-dimensionally graphing the intensity change of the magnetic field of the muscle generated during eye movement after contacting the sclera with a contact lens-type device with a wire embedded therein.

The sclera search coil system has the advantage of receiving more accurate signals and capturing minute eye movements compared to electrospinography.

However, since the sclera search coil system uses contact lenses, there is a problem that the use of the sclera search coil is limited in children who are less cooperative.

Therefore, an object of the present invention is to rotate the patient from side to side affects the vestibular organ of the patient and based on the rotation conditions of the infrared camera to analyze the infrared pupil images taken by the infrared camera to perform a nystagmus test for the patient An apparatus and method for testing video nystagmus are provided.

The video nystagmus test apparatus using an infrared camera according to the present invention for achieving the above object comprises: a rotating chair which is configured in the form of a chair in which the patient sits and rotates left and right under control; An infrared imaging unit configured to output an infrared eye image by infrared photographing an eye area including at least one of two eyes of the patient; And controlling rotation of the swivel chair according to preset rotation control information for eye examination, detecting the pupil by binarizing the infrared eye image, detecting a movement characteristic of the pupil, and detecting based on the rotation control information. It is characterized in that it comprises a nystagmus inspection unit for performing the nystagmus test by analyzing the pupil movement characteristics and output the result.

The nystagmus inspection unit, the rotation control unit for controlling the rotation of the rotary chair according to the rotation control information preset for the nystagmus test; An infrared image acquisition unit configured to control and drive the infrared image capture unit, and to acquire and output the infrared eye image through the infrared image capture unit; An image processor configured to detect the pupil after binarizing the obtained infrared eye image and output image data of the detected pupil; A pupil motion detector for receiving and analyzing the image data and outputting pupil motion characteristic information on pupil movement; And an nystagmus determination unit that performs an nystagmus test and analyzes nystagmus by analyzing the detected pupil movement characteristic information based on the rotation control information.

The image processing unit may include a region of interest designation unit configured to output an infrared front eye part image including only the front eye part from the infrared eye image; An image converter configured to output the binarized preocular part image by binarizing the infrared preocular part image; An edge detector for detecting an outline of the anterior part of the eye from the binarized anterior eye part image and outputting an anterior eye image; And a circle detector for detecting a pupil from the anterior image of the anterior part, detecting a center point of the pupil, and outputting image data of the pupil according to the movement to the detected center point.

The image converting unit may include: a first image converting unit converting the infrared whole eyepiece image into a gray eyepiece image of a gray channel and outputting the gray eyepiece image; And a second image converter configured to binarize the gray anterior eye part image based on a preset reference value and output the binarized front eye part image.

The image processing unit may further include a filtering unit which removes noise by applying erosion and expansion by performing a morphology operation on the binarized frontal eye image output from the image conversion unit.

The circle detector receives an anterior eye segment outline image for at least two or more binarized anterior eye segment images converted from a plurality of different reference values, and detects a circle from each anterior eye segment outline image. An optimal circle detector for selecting and outputting an image of an anterior part of the eye where the circle corresponds to the pupil; And calculating a center point of the circle for the selected anterior eyelid image, and outputting the pupil image data by estimating the distance and angle of the actual pupil movement in pixel units in consideration of the three-dimensional structure and curve of the pupil. And an image calibration unit.

The pupil motion detection unit may include: a first interpolation unit which receives the image data and applies a linear spline algorithm to interpolate a signal missing due to blinking to output a first interpolated image signal; A differentiator for obtaining a slope from the first interpolated image signal and outputting a differential image signal obtained by squaring the slope; A fast phase remover for outputting a slow phase signal by removing the fast phase signal from the first interpolated video signal by applying the differential image signal to the first interpolated video signal; A second interpolation unit for outputting the slow phase signal after interpolation by applying a linear spline algorithm; A filtering unit for filtering and outputting the interpolated slow phase signal; And a feature extractor for extracting gain, asymmetry and phase characteristics from the slow phase signal to output the pupil motion characteristic information.

According to the present invention, there is provided a video nystagmus inspection method using an infrared camera, the rotation control step of controlling the rotation of the swivel chair according to the preset rotation control information for nystagmus inspection; An infrared image acquisition step of controlling and driving an infrared image photographing unit and then obtaining and outputting an infrared eye image through the infrared photographing unit; An image processing step of detecting a pupil after binarizing the obtained infrared eye image and outputting image data of the detected pupil; A pupil movement detection step of receiving and analyzing the image data and outputting pupil movement characteristic information on pupil movement; And nystagmus determination step of analyzing the detected pupil movement characteristic information based on the rotation control information, performing nystagmus test, and determining and outputting nystagmus.

The image processing step may include: an ROI designation step of outputting an infrared front eye part image including only the front eye part from the infrared eye image; An image conversion step of binarizing the infrared eyepiece image and outputting a binarized eyeball image; An edge detection step of detecting an outline of an anterior eye part from the binarized anterior eye part image and outputting an anterior eye image; And a circle detection step of detecting a pupil from the frontal image of the front part, detecting a center point of the pupil, and outputting image data of the pupil according to the movement to the detected center point.

The image converting step may include: a first image converting step of converting and outputting the infrared whole eyepiece image to a gray eyepiece image of a gray channel; And a second image conversion step of outputting the binarized preocular part image by binarizing the gray anterior eye part image based on a preset reference value.

The image processing step may further include a filtering step of removing noise by applying erosion and swelling by performing a morphology operation on the binarized frontal eye image output from the image converter.

The circle detection step may be input from an anterior eye region outline image detection unit for at least two or more binarized anterior eye segment images converted based on a plurality of changed reference values, and detect a circle from each anterior eye segment outline image. An optimal circle detection step of selecting and outputting an image of an anterior part of the eye where the circle corresponds to the pupil; And calculating a center point of the circle for the selected anterior eyelid image, and outputting the pupil image data by estimating the distance and angle of the actual pupil movement in pixel units in consideration of the three-dimensional structure and curve of the pupil. And an image calibration step.

The pupil motion detecting step may include: a first interpolation step of receiving the image data and applying a linear spline algorithm to interpolate a signal missing due to blinking to output a first interpolated image signal; A derivative step of obtaining a slope from the first interpolated image signal and outputting a differential image signal obtained by squaring the slope; A fast phase removing step of applying a differential video signal to the first interpolated video signal and outputting a slow phase signal from which a fast phase signal is removed from the first interpolated video signal; A second interpolation step of interpolating and outputting the slow phase signal by applying a linear spline algorithm; A filtering step of filtering and outputting the interpolated slow phase signal; And extracting gain, asymmetry and phase characteristics from the slow phase signal to output the pupil motion characteristic information.

According to the present invention, the pupil movement is photographed by irradiating infrared rays, the pupil region is separated by using the contrast between the pupil and the iris in the photographed image frame, and the boundary coordinates and the ellipsoid approximation of the pupil are applied to set the center coordinates of the pupil. Since it can be displayed for each time zone, it has the effect of recording the three-dimensional movements of the eye (horizontal, vertical, and line) in real time.

In addition, the present invention can record the eye movements in detail for each time zone in real time has the effect of performing a more accurate nystagmus test.

1 is a diagram showing the configuration of a video nystagmus test apparatus using an infrared camera according to the present invention.

2 is a diagram illustrating a detailed configuration of a nystagmus testing unit of a video nystagmus test apparatus according to the present invention.

3 is a view showing the detailed configuration of the image processing unit of the nystagmus test unit according to the present invention.

4 is a diagram illustrating an eye image for each stage processed by an image processor according to the present invention.

5 is a diagram illustrating a binarization image and a circle detection image according to different reference values at the time of image binarization applied according to the present invention.

6 is a view showing the detailed configuration of the pupil motion detection unit of the nystagmus test unit according to the present invention.

7 is a diagram illustrating a horizontal pupil signal for horizontal pupil data according to an embodiment of the present invention.

8 is a view showing signal waveforms for each step of image processing in the pupil motion detector according to the present invention.

9 is a view illustrating a signal processing process of a slow phase signal in the pupil motion detector according to the present invention.

10 is a view showing a gain curve according to the rotational stimulation frequency of the present invention.

11 is a view showing a phase curve according to the rotational stimulation frequency of the present invention.

12 is a flowchart illustrating a video nystagmus inspection method using an infrared camera according to the present invention.

13 is a flowchart illustrating an image processing method of a video nystagmus testing method according to the present invention.

14 is a flowchart illustrating a pupil motion characteristic detection method of a video nystagmus test method according to the present invention.

Hereinafter, with reference to the accompanying drawings will be described the configuration and operation of the nystagmus test apparatus using an infrared camera according to the present invention, and describes the nystagmus test method in the device.

Referring to FIG. 1, a video nystagmus testing apparatus using an infrared camera of the present invention includes a display unit 100, an infrared imaging unit 200, a swivel chair 300, and a nystagmus testing unit 400, according to an embodiment. The input unit 500 may further include.

Under the control of the nystagmus tester 400, the display unit 100 displays information generated during the nystagmus test process, the nystagmus test result, and the like as text, graphics, still images, and videos.

The infrared imaging unit 200 captures at least one or more eye regions of two eyes of a patient who is an eye test in a state where infrared light is irradiated through the infrared LED and the infrared light is irradiated through the infrared LED. It may include an infrared camera (not shown) for output.

The infrared image capturing unit 200 may be configured in the form of a camera to be placed in front of the patient's eyes, or in the form of virtual reality (VR) glasses combined with the display unit 100 to be worn on the patient's head. It may be configured.

In addition, the infrared imaging unit 200 may further include an acceleration sensor (not shown) that measures and outputs an acceleration according to at least one left and right rotation of the patient and the patient's head when the infrared imaging unit 200 is configured to be worn on the head. Could be

Rotating chair 300 is configured in the form of a chair that the patient can sit and rotates the chair by adjusting the direction and speed under the control of the nystagmus test unit 400.

The nystagmus tester 400 controls the rotation of the swivel chair 300 according to the preset rotation control information for the nystagmus test, detects the pupil by binarizing the infrared eye image input from the infrared imaging unit 200. In addition, the motion characteristics of the detected pupil are detected, the pupil movement characteristic is analyzed based on the rotation control information, and the nystagmus test is performed and the result is output.

The input unit 500 may include at least one key, button, or switch for controlling the operation of the nystagmus tester, and outputs an input signal corresponding to the operated key, button, or switch to the nystagmus tester 400.

The input unit 500 may be a keyboard, a mouse, a touch pad, or the like, and accordingly to the present invention, a threshold value for binarization of an image may be directly input or selected from an administrator (or a “doctor”).

Referring to FIG. 2, the nystagmus checker 400 includes a rotation controller 410, an infrared image acquisition unit 420, an image processor 430, a pupil motion detector 440, and an nystagmus determination unit 450.

The rotation controller 410 controls the rotation direction and the rotation speed of the swivel chair 300 according to the rotation control information preset for the nystagmus test.

The rotation controller 410 sequentially or randomly selects one of the frequencies 0.01, 0.02, 0.04, 0.08, 0.16, 0.32, and 0.64 Hz, or two or more of the frequencies at a sinusoidal speed to determine the rotation speed of the rotating chair 300. You will be able to control.

In addition, the rotation controller 410 provides the nystagmus determination unit 450 with information on the direction and speed currently being controlled.

The infrared image acquisition unit 420 drives the infrared imaging unit 200 to obtain an infrared eye image through the infrared imaging unit 200 and outputs the infrared eye image to the nystagmus inspection unit 400.

The image processor 430 detects the pupil after binarizing the infrared eye image acquired through the infrared image acquisition unit 420, and outputs the detected image data of the pupil to the pupil motion detector 440.

The pupil motion detector 440 receives the image data and analyzes the image data, and outputs pupil movement characteristic information on the movement of the pupil to the nystagmus determination unit 450.

The nystagmus determination unit 450 receives the rotation control information from the rotation control unit 410, analyzes the pupil movement characteristic information detected based on the rotation control information, performs an nystagmus test, and determines whether or not nystagmus is determined. It is displayed through the display unit 100.

3 is a view showing the detailed configuration of the image processing unit of the nystagmus test unit according to the present invention, Figure 4 is a view showing the eye image step-by-step processed by the image processing unit according to the present invention, Figure 5 is an image applied according to the present invention A diagram illustrating a binarization image and a circle detection image according to different reference values during binarization. Hereinafter, the detailed configuration and operation of the image processor 430 will be described with reference to FIGS. 3 to 5.

The image processor 430 includes an ROI designator 431, an image converter 432, a filter 435, an outline detector 436, and a circle detector 437.

The ROI designator 431 outputs to the image converting unit 432 the infrared front eye region image 502 obtained by extracting only the front eye portion from the infrared eye image 501 of FIG. 4 input from the infrared image acquisition unit 420. . The eyepiece portion may be selected by an administrator through the input unit 500 or may be configured to be automatically detected by eye pattern detection.

The image converter 432 binarizes the infrared ray eyepiece image obtained through the ROI designator 431 and outputs the binarized eyeball image 503.

The image converter 432 binarizes the first image converter 433 for converting the infrared eyepiece image to a gray eyepiece image of a gray channel and outputs the gray eyepiece image based on a preset reference value. The second image converter 434 outputs the binarized eyepiece image 503.

The binarization is performed by converting a pixel value of a gray front eye image converted into a gray channel exceeding a preset reference value to 255, and converting the following values to zero. The preset reference value may be one, or may be two or more. In case of two or more reference values, the second image converter 434 performs binarization corresponding to each reference value and binarization pre-image part 504 for each reference value. Will print

The filtering unit 435 transmits the binarized eyeball image 504 from which the noise included in the binarized eyepiece image 503 input from the second image converter 434 is removed to the outline detector 436.

The filtering unit 435 removes noise by applying erosion and dilate by morphology (or "morph") operation. Erosion is the replacement of the minimum pixel value among the neighboring pixels with the current pixel value, and may cause small noise (small chunks) to disappear. In addition, the expansion replaces the maximum pixel value among the neighboring pixels with the current pixel value, and serves to disappear the small holes in the object. That is, the erosion operation removes the noise, the expansion operation removes the empty space in the object, and restores the reduced object size.

The outline detector 436 detects the outline of the anterior eye part from the binarized anterior eye part image 504 filtered by the filtering part 435, and transmits the front eye part outline image 505 to the original detector 437.

The circle detector 437 detects a circle corresponding to the pupil from the voltage portion outline image 505 input from the outline detector 436, and outputs the detected circle, that is, the image data of the pupil according to the movement of the pupil.

When the second binarization unit 434 outputs the binarization predecessor image 504 for a plurality of reference values, the original detector 437 may convert at least two or more binarization conversions based on a plurality of changed reference values. An anterior eye image (505) for the ocular image (503) is received from the outline detector (436), a circle (506) is detected from each anterior eye image (505), and the detected circle (506) is applied to the pupil. An optimal circle detector 438 that selects and outputs a matching anterior eye image 505 and calculates a center point of a circle for the selected anterior eye image, and considers the calculated center point in consideration of the three-dimensional structure and curve of the pupil. The image calibrator 439 may output an image data of the pupil by estimating the distance and angle of the actual pupil in pixels.

That is, the optimal circle detector 438 acquires binarized frontal eye contour images 505-1, 505-2, 505-3, and 505-4 by applying different reference values as shown in FIG. 5. When the circle 506 is detected from the eyeball outline images, the circle will be detected as shown in FIGS. 511-1 to 511-4 of FIG. 5. In FIG. 5, 505-1 is a case where 262 is applied as a reference value, 505-2 is a case where 270 is applied as a reference value, 505-3 is a case where 275 is applied as a reference value, and 505-4 is a case where 285 is applied as a reference value. It is shown.

Therefore, the optimal circle detector 438 selects and outputs an anterior eye region outline image 505-3 for 511-3 in which the pupil and the detected circle 506 coincide among the 511-1 to 511-4.

In addition, since the image calibrator 439 needs an angle per pixel value in order to obtain a speed of horizontal eye movement, the image calibration unit 439 needs to convert an eye movement unit to an angle in an image. Preferably, the angle is 0.49 ms per pixel.

6 is a view showing a detailed configuration of the pupil motion detection unit of the nystagmus test unit according to the present invention, Figure 7 is a view showing a horizontal pupil signal for the horizontal pupil data according to an embodiment of the present invention, Figure 8 is a present invention 9 is a view illustrating a signal waveform for each step of image processing in a pupil motion detector according to an embodiment of the present invention, and FIG. 9 is a view for explaining a signal processing process of a slow phase signal in a pupil motion detector according to the present invention, and FIG. FIG. 11 is a diagram illustrating a gain curve according to a stimulus frequency, and FIG. 11 is a diagram illustrating a phase curve according to a rotational stimulation frequency of the present invention. Hereinafter, a description will be given with reference to FIGS. 6 to 11.

The pupil motion detector 440 includes a first interpolator 441, a differentiator 442, a fast phase remover 443, a second interpolator 444, a filter 445, a slow phase signal detector 446, and Feature extraction unit 447 is included.

The first interpolator 441 receives the image data from the image processor 430 and applies a linear spline algorithm to interpolate the missing signal due to blinking to output the first interpolated image signal.

For example, in the horizontal pupil signal of the horizontal pupil movement data among the image data, a frame in which no pupil is detected due to blinking may be set to a negative value.

The first interpolator 441 applies a linear spline algorithm to complement a frame set to a negative value.

In addition, the first interpolator 441 converts an irregular sampling rate (27-29 frames / second) into a 30 Hz signal by applying a third-order spline algorithm, which is an interpolation method.

In conclusion, the first interpolator 441 interpolates a pupil movement signal including a horizontal pupil signal such as 701 of FIG. 7 corresponding to the image data and a pupil movement signal including the interpolated horizontal pupil signal of 702 of FIG. 7. Output as image data.

The differentiator 442 obtains a slope by differentiating the interpolated first interpolated image signal as shown in 801 of FIG. 8, and outputs a signal as shown in 802 of FIG. 8, and squares the slope to obtain a slope of FIG. 8. Outputs a differential video signal such as

The fast phase remover 443 applies the differential image signal to the first interpolated image signal, and thus, a fast phase signal corresponding to the dotted rectangle as shown at 804 of FIG. 8 in the first interpolated image signal illustrated at 801 of FIG. 8. Outputs a slow phase signal such as 901 of FIG.

The second interpolator 444 interpolates the slow phase signal by applying a linear spline algorithm and then outputs a slow phase signal as shown in 902 of FIG. 9.

The filtering unit 445 filters the interpolated slow phase signal and outputs a slow phase signal such as 903 of FIG. 9. A low pass filter is preferably applied to the filtering unit 445, and the cutoff frequency of the low pass filter may be set by Equation 1 below.

[Equation 1]

Where k stands for a constant and n stands for each cycle number.

The feature extractor 446 extracts gain, asymmetry, and phase characteristics from the slow phase signal and outputs the pupil motion characteristic information.

Gain is the ratio of the amplitude of eye movements to the amplitude of head stimuli. It is calculated by dividing the maximum velocity of slow phase eye movements by the maximum velocity of the head (chair). This can be represented by Equation 2 below.

[Equation 2]

For example, a patient with weak bilateral vestibularity will have a reduced gain at low frequencies of rotation.

The gain can be represented by a theoretical log gain curve according to the rotational stimulation frequency, as shown in FIG. 10, linearly increasing with increasing frequency at rotations below 0.05 Hz, constant from about 1 to 100 Hz, It can be seen that at higher rotational frequencies, the frequency decreases as it increases. The decrease in gain can be seen at all frequency stimuli, but only at constant stimulus frequencies. In the case of partial pruning, gain is usually reduced abnormally at low frequencies (<0.1 Hz), normally at high frequencies (> 0.1 Hz), and at low speed stimulus with fixed rotational frequency and maximum speed It is shown to decrease at normal but high velocity stimuli.

Symmetry / Asymmetry is shown by comparing the equivalent slow phase elements on the left and right sides when rotating left and right for each frequency. Asymmetry may be calculated by the following equation (3).

[Equation 3]

Here, b1 and b2 represent the maximum velocity of the slow phase eye movement when rotating left and right, respectively.

Normally, the normal symmetry range is 15-20%.

Phase is the time relationship between the onset of head movement and the reflex eye reaction. If the head and eyes move in opposite directions at exactly the same speed, the phase is assumed to be 180 Hz. A phase lead appears when the movement of the head causes an eye response reflection, and a phase lag appears when the correction eye movement proceeds. Abnormally long leads and short leads can be considered peripheral and cerebellar lesions, respectively.

The phase angle θ is calculated by the following equation.

[Equation 4]

Where f represents the frequency of the eye movement signal, and Δt represents the time difference between the maximum velocity of slow phase eye movement and head (chair) movement.

The phase can be represented by a phase curve according to the rotational stimulation frequency as shown in Fig. 11. If perfect electrostatic reflection occurs, the phase is zero Hz.

Referring to FIG. 12, when the nystagmus test start event occurs, the nystagmus tester 400 controls the rotation direction and the speed of the swivel chair 300 according to the preset rotation control information (S111). The nystagmus test start event may occur when an input signal corresponding to the nystagmus test start command is input through the input unit 500. The rotation control information may include information such as the direction, speed, time to rotate the rotating chair 300.

When the control of the swivel chair 300 is performed, the nystagmus tester 400 drives the infrared imaging unit 200 to start capturing the patient's eyes to acquire an infrared eye image (S113).

When the infrared eye image starts to be acquired, the nystagmus tester 400 performs image processing to detect the pupil by binarizing the input infrared eye image to generate image data including the movement information of the pupil (S115).

When the image data is generated, the nystagmus tester 400 estimates pupil movement information from the image data (S117), and analyzes pupil movement information based on the rotation control information to determine nystagmus (S119).

Referring to FIG. 13, the image processing method 430 detects only an anterior eye region, which is a region of interest, from an infrared eye image, and outputs an infrared anterior eye region image including only the detected anterior eye region (S211). ).

The image processor 430 converts the infrared eyepiece image to a gray eyepiece image, which is an image of a gray channel (S213).

When the gray eyepiece image is converted into an image, the image processor 430 compares the gray eyepiece image with a preset reference value, converts a pixel value larger than the reference value to 255, and converts a pixel value smaller than the reference value to 0 into a binary conversion process. In operation S215, the binarization whole eye part image is output.

When the binarized preocular part image is generated, the image processor 430 filters and outputs the binarized preocular part image (S217). The contour of the anterior eye part is detected from the filtered binarized anterior eye part image, and an output of the anterior eye part outline image is output (S219).

When the anterior eye segment is detected, the image processor 430 detects a circle corresponding to the pupil from the anterior eye segment image and detects a center point of the circle (S221), and generates and outputs image data of the pupil according to the movement to the detected center point. (S223).

Referring to FIG. 14, the pupil motion detector 440 receives the image data output from the image processor 430 and applies a linear spline algorithm to interpolate a missing signal due to eye blink to interpolate a first interpolated image signal. Output it (S311).

When the first interpolation is performed, the pupil motion detector 440 obtains an inclination from the first interpolation image signal and outputs a differential image signal obtained by squaring the inclination (S313).

When the differential video signal is generated, the pupil motion detector 440 removes the fast phase signal from the first interpolated video signal by applying the differential video signal to the first interpolated video signal, and removes the fast phase signal from the first interpolation. A slow phase signal which is a video signal is output (S315).

When the interpolated slow phase signal is input, the pupil motion detector 440 interpolates in a second order by applying a linear spline algorithm and outputs the second interpolated signal (S317).

The second interpolated slow phase signal is filtered using a low pass filter (S319), and motion feature information including gain, asymmetry, and phase information is extracted from the low pass filtered slow phase signal and output to the nystagmus determination unit 450. do.

On the other hand, the present invention is not limited to the above-described typical preferred embodiment, but can be carried out in various ways without departing from the gist of the present invention, various modifications, alterations, substitutions or additions in the art Anyone who has this can easily understand it. If the implementation by such improvement, change, replacement or addition falls within the scope of the appended claims, the technical idea should also be regarded as belonging to the present invention.

[Description of the code]

100: display unit 200: video recording unit

300: swivel chair 400: nystagmus tester

410: rotation control unit 420: infrared image acquisition unit

430: Image processor 431: Region of interest designator

432: Image converter 433: First image converter

434: second image converter 435: filter

436: outline detector 437: circle detector

438: optimal circle detection unit 439: image calibration unit

440: pupil movement detector 441: first interpolator

442: Differentiator 443: Fast Phase Control

444: second interpolator 445: filtering unit

446: feature extraction unit

450: nystagmus

Claims

Rotating chair consisting of a chair sitting patient is rotated to the left and right under control;

An infrared imaging unit configured to output an infrared eye image by infrared photographing an eye area including at least one of two eyes of the patient; And

Control the rotation of the swivel chair according to the preset rotation control information for the eye examination, detect the pupil by binarizing the infrared eye image, detects the movement characteristics of the pupil, detected based on the rotation control information A video nystagmus testing apparatus using an infrared camera, characterized in that it comprises a nystagmus inspection unit for analyzing the pupil movement characteristics to perform nystagmus inspection and output the results.
The method of claim 1,

The nystagmus test unit,

Rotation control unit for controlling the rotation of the swivel chair in accordance with the preset rotation control information for the eye test;

An infrared image acquisition unit configured to control and drive the infrared image capture unit, and to acquire and output the infrared eye image through the infrared image capture unit;

An image processor configured to detect the pupil after binarizing the obtained infrared eye image and output image data of the detected pupil;

A pupil motion detector for receiving and analyzing the image data and outputting pupil motion characteristic information on pupil movement; And

And a nystagmus determination unit for performing nystagmus test by analyzing the detected pupil movement characteristic information based on the rotation control information and determining whether nystagmus is present.
The method of claim 2,

The image processor,

A region of interest designation unit configured to output an infrared anterior eye region image including only an anterior eye region from the infrared eye image;

An image converter configured to output the binarized preocular part image by binarizing the infrared preocular part image;

An edge detector for detecting an outline of the anterior part of the eye from the binarized anterior eye part image and outputting an anterior eye image; And

And a circle detector for detecting a pupil from the anterior image of the anterior part, detecting a center point of the pupil, and outputting image data of the pupil according to the movement to the detected center point.
The method of claim 3,

The image conversion unit,

A first image converter converting the infrared eyepiece image into a gray eyepiece image of a gray channel; And

And a second image converter configured to binarize the gray anterior eye part image based on a preset reference value and output the binarized front eye part image.
The method of claim 3,

The image processor,

And a filtering unit which removes noise by applying erosion and swelling by performing a morphology operation on the binarized frontal eye image output from the image converter.
The method of claim 3,

The circle detector,

Receive an anterior eye segment image for at least two or more binarized anterior eye segment images based on a plurality of changed reference values from an edge detector, detect a circle from each anterior eye segment image, and match the detected circle to the pupil An optimal circle detector for selecting and outputting an image of a frontal part; And

An image of calculating the center point of the circle for the selected anterior eye image, and outputting the image data of the pupil by estimating the distance and angle of movement of the actual pupil in pixels based on the three-dimensional structure and curve of the pupil. Video nysight inspection device using an infrared camera, characterized in that it comprises a calibration unit.
The method of claim 2,

The pupil motion detector,

A first interpolation unit receiving the image data and applying a linear spline algorithm to interpolate a signal missing due to blinking to output a first interpolated image signal;

A differentiator for obtaining a slope from the first interpolated image signal and outputting a differential image signal obtained by squaring the slope;

A fast phase remover for outputting a slow phase signal by removing the fast phase signal from the first interpolated video signal by applying the differential image signal to the first interpolated video signal;

A second interpolation unit for outputting the slow phase signal after interpolation by applying a linear spline algorithm;

A filtering unit for filtering and outputting the interpolated slow phase signal; And

And a feature extractor for extracting gain, asymmetry, and phase characteristics from the slow phase signal to output the pupil motion characteristic information.
A rotation control step of controlling rotation of the swivel chair according to rotation control information preset for the nystagmus test;

An infrared image acquisition step of controlling and driving an infrared image photographing unit and then obtaining and outputting an infrared eye image through the infrared photographing unit;

An image processing step of detecting a pupil after binarizing the obtained infrared eye image and outputting image data of the detected pupil;

A pupil movement detection step of receiving and analyzing the image data and outputting pupil movement characteristic information on pupil movement; And

And a nystagmus determination step of analyzing the detected pupil movement characteristic information based on the rotation control information, performing nystagmus test, and determining and outputting nystagmus.
The method of claim 8,

The image processing step,

A region of interest designation step of outputting an infrared anterior eye region image including only an anterior eye region from the infrared eye image;

An image conversion step of binarizing the infrared eyepiece image and outputting a binarized eyeball image;

An edge detection step of detecting an outline of an anterior eye part from the binarized anterior eye part image and outputting an anterior eye image; And

And detecting a pupil from the anterior image of the anterior part of the eye, detecting a center point of the pupil, and outputting the image data of the pupil according to the movement to the detected center point.
The method of claim 9,

The image conversion step,

A first image conversion step of converting and outputting the infrared eyepiece image to a gray eyepiece image of a gray channel; And

And a second image converting step of outputting the binarized preocular part image by binarizing the gray anterior eye part image based on a preset reference value.
The method of claim 9,

The image processing step,

And a filtering step of removing noise by applying erosion and swelling by performing a morphology operation on the binarized frontal eye image output from the image converter.
The method of claim 9,

The circle detection step,

An input from an anterior eye region image edge detector for at least two or more binarized anterior eye segment images converted based on a plurality of changed reference values, a circle is detected from each anterior eye segment image, and the detected circles coincide with the pupil. An optimal circle detection step of selecting and outputting an anterior eye image; And

An image of calculating the center point of the circle for the selected anterior eye image, and outputting the image data of the pupil by estimating the distance and angle of movement of the actual pupil in pixels based on the three-dimensional structure and curve of the pupil. A video nystagmus test method using an infrared camera comprising a calibration step.
The method of claim 8,

The pupil movement detecting step,

A first interpolation step of receiving the image data and applying a linear spline algorithm to interpolate a signal missing due to blinking to output a first interpolation image signal;

A derivative step of obtaining a slope from the first interpolated image signal and outputting a differential image signal obtained by squaring the slope;

A fast phase removing step of applying a differential video signal to the first interpolated video signal and outputting a slow phase signal from which a fast phase signal is removed from the first interpolated video signal;

A second interpolation step of interpolating and outputting the slow phase signal by applying a linear spline algorithm;

A filtering step of filtering and outputting the interpolated slow phase signal; And

And a feature extraction step of extracting gain, asymmetry, and phase characteristics from the slow phase signal to output the pupil movement characteristic information.