JP2017164224A - Calibration method and ocular movement examination system adopting the same - Google Patents

Abstract

A calibration method capable of detecting eye movement with high accuracy by reducing the burden on a subject and eliminating error factors based on individual differences and the like, and an eye movement inspection system employing the calibration method.
A front-view image capturing step of step S1 for capturing a front-view image of an eyeball, an angle pixel coefficient calculation step of step S2 for calculating an angle pixel coefficient, and a rotated image of step S3 for capturing a rotated image of an eyeball. A photographing step and a moving angle calculating step of step S4 for calculating the moving angle.
[Selection] Figure 1

Description

  The present invention relates to a calibration method used in an eye movement inspection method for inspecting a subject's eye movement and an eye movement inspection system employing the calibration method, and in particular, eliminates the need for gazing at an index multiple times. The present invention relates to a calibration method for detecting eye movement with high accuracy and an eye movement inspection system employing the same.

Conventionally, in an eye movement inspection method for inspecting a subject's eye movement, a calibration method using an index plate, a wall, or an index such as a laser beam incorporated in an infrared Frenzell camera has been employed.
This calibration method will be described. For example, in the case of using an infrared Frenzel camera, first, an index is placed in front of the subject to be watched, and an infrared image of the eyeball at that time is taken. Subsequently, the indicators are arranged by being rotated by a certain angle θ around the subject's head, respectively, upward, downward, right side, and left side, and are respectively watched without moving the head. And the infrared image of the eyeball in each angle (theta) is image | photographed.
Thereafter, by comparing the infrared image taken at the front and the infrared image at the angle θ, for example, the number N of pixels proportional to the distance X in which the pupil center is actually moved in the horizontal direction is measured. In the human body, there is an individual difference in the distance from the camera constituting the infrared Frenzel camera to the center of the iris. Therefore, the enlargement ratio of the image [(number of pixels N × length of one pixel) / The distance X] varies. Therefore, the measured pixel count N cannot be immediately used as a distance index of eye movement. Therefore, an angle α per pixel in the horizontal direction is obtained by dividing the angle θ by the number N of pixels, and this angle α is used as a calibration coefficient.
Specifically, for example, an infrared image of the eyeball is taken when the subject is gazing at the index moved to an arbitrary horizontal position. The number M of pixels on which the pupil center has moved is measured on this image, and the angle of rotation of the eyeball is obtained from the product of the number M of pixels and the angle α per pixel described above, and the presence or absence of nystagmus can be diagnosed. .
However, in such a calibration method, the work of gazing at an index that moves to a plurality of positions has a problem that it is both cumbersome and time consuming for both the subject and the examiner. Furthermore, since the subject cannot accurately gaze at the moved index, there is a problem that the accuracy of the angle α per pixel is not good, and the accuracy of the angle of rotation of the eyeball is not good.
Therefore, in recent years, technologies have been developed that can accurately evaluate eye movements by creating and analyzing time series data of eye movements and correcting the measured pupil size. The invention is disclosed.

Patent Document 1 discloses an invention relating to an nystagmus inspection device that objectively evaluates the presence or absence of nystagmus, a program used therefor, and the like, under the name of “nystagmus inspection device and program and recording medium used therefor”.
Hereinafter, the invention disclosed in Patent Document 1 will be described. The invention disclosed in Patent Document 1 includes a video generation unit that generates a video including a target for inducing reciprocation of the eyeball of the subject, and a video presentation unit that presents the video generated by the video generation unit to the subject. Rhythm fluctuation analysis from the eyeball measuring means for measuring the movement of the eyeball when the subject follows the visual target presented by the video presenting means, and the time series data of the eyeball movement measured by the eyeball measuring means The nystagmus examination apparatus is provided with a calculation means for performing the above and a display means for displaying the analysis result by the calculation means as diagnosis support data.
According to the nystagmus examination apparatus having such a feature, the analysis result obtained by analyzing the rhythm fluctuation of the time series data of the movement of the eyeball is displayed on the display means as the diagnosis support data. The presence or absence can be objectively evaluated.

Next, Patent Document 2 discloses an invention relating to a measurement system that measures the size of the pupil and the position of the line of sight in a non-invasive manner under the name of “measurement system using ellipse approximation resistant to noise”.
The invention disclosed in Patent Document 2 is based on an accurate ellipse approximation that is robust to noise, in which a pupil imaged by a fixed focus camera is approximated by an ellipse by image processing, and the size of the pupil can be continuously measured from the approximated ellipse. It is a measurement system.
According to the measurement system having such a feature, the length on the image is proportional to the length of the real world by using the fixed focus camera. The pupil size can be measured by obtaining the proportionality constant in advance. The radius of the pupil can be obtained by approximating the pupil with an ellipse by image processing and multiplying the major axis of the ellipse by the proportional constant described above. Furthermore, the line-of-sight position can be measured from the center coordinates of the ellipse obtained by approximating the pupil with an ellipse.

JP 2003-250763 A JP-A-8-145644

  However, in the invention disclosed in Patent Document 1, a test is performed by the subject to follow the visual target of the video presented by the video presentation unit a plurality of times. Therefore, although the complexity of the gaze work on the inspector side is reduced, the burden on the subject side is not much reduced.

  Next, in the invention disclosed in Patent Document 2, the center coordinates of the obtained ellipse (that is, the center coordinates of the pupil) are measured over time, so that it can be applied to the diagnosis of nystagmus. It is done. However, the proportionality constant for obtaining the size of the ellipse is obtained from the ratio of the actual length of the pupil and the length on the computer screen by photographing a ruler with a camera. On the other hand, as described above, there is an individual difference in the distance from the camera to the center of the pupil. Therefore, calculating the pupil size using a constant proportional constant as described above causes an error due to the individual difference. It is considered that it is not appropriate. That is, there is a possibility that the accuracy of the center coordinates of the pupil based on the size of the pupil obtained by this proportionality constant is not good.

  The present invention has been made in response to such a conventional situation. The burden on the subject is reduced by eliminating the need to pay attention to the index multiple times, and error factors based on individual differences are eliminated. An object of the present invention is to provide a calibration method capable of detecting eye movement with high accuracy and an eye movement inspection system employing the calibration method.

  To achieve the above object, the first invention is a calibration method used in an eye movement inspection method for examining eye movement by analyzing image data of a subject's eye, and specifies eye movement. A front-view image capturing process for capturing a front-view image of the eyeball in front view in association with a feature point included in an iris that rotates with a rotation angle around the eye rotation point, The angle pixel coefficient calculation step for calculating the angle pixel coefficient based on the distance pixel coefficient obtained by dividing the diameter by the number of pixels reflecting the diameter of the iris on the visual image, and the iris A rotation image capturing process for capturing a rotation image of the eyeball when the rotation is performed around the eyeball rotation point with a moving angle, an angle pixel coefficient, and a feature point on the front view image A moving angle calculation step of calculating a moving angle by calculating a product of the number of moving pixels corresponding to the displacement of the feature point when moving from the position to the position on the rotation image. To do.

In the invention having such a configuration, the phrase “in association with the feature points included in the iris” in the front-view image photographing process is to enable the feature points themselves and their positions to be extracted after photographing. This feature point is selected in advance from anatomical features that do not vary in shape due to the surrounding environment and are easy to recognize, such as one point on the center of the pupil or the periphery of the iris. It is.
Further, “front view” is a concept including a state in which the line of sight is slightly shifted from the front in the vertical and horizontal directions.

  Next, in the angle pixel coefficient calculation step, the distance pixel coefficient is obtained by dividing this diameter by the number of pixels reflecting the diameter of the iris on the front view image. Further, the angle pixel coefficient is calculated based on the distance pixel coefficient. This angle pixel coefficient is employed as an index for calibration in the present invention.

  Further, in the rotation image capturing process, when the examiner instructs the examiner to gaze at an arbitrary direction, the eyeball of the subject rotates with a moving angle, so that the iris is also different from the case where the eyeball is viewed in front. Move to position. The image of the eyeball when the iris is moved is photographed from the same direction as when photographed in the front-view image photographing process in association with the feature point.

Finally, in the moving angle calculation step, the moving pixel number is converted into a moving angle by the product of the angle pixel coefficient and the moving pixel number.
Note that the displacement of the feature point is the displacement of the distance. In addition, the number of moving pixels is a value including an error based on the individual difference of the human body as described above because it is measured using the front view image and the rotated image. On the other hand, the moving angle is a value indicating how much the eyeball is rotated around the eyeball rotation point, and thus the above error is completely eliminated.

  Next, a second invention is characterized in that, in the first invention, the displacement is a displacement along at least one of a horizontal direction and a vertical direction. In the invention having such a configuration, in addition to the action of the first invention, the displacement is obtained independently in the horizontal direction and the vertical direction.

Subsequently, the third invention is an eye movement inspection system adopting the first or second invention, and a front-view image photographing means for photographing a front-view image, and an angle pixel coefficient calculation for calculating an angle pixel coefficient. And a rotation image photographing means for photographing a rotation image, and a movement angle calculation means for calculating a movement angle.
In the invention having such a configuration, the first or second invention is executed, and the movement angle is automatically calculated.

According to the first invention, in the front-view image photographing process or the movement angle calculation process, the movement angle can be easily set without requiring a program for performing complicated coordinate calculation as in the invention disclosed in Patent Document 2. It can be calculated.
Further, the moving angle can be obtained from the number of moving pixels measured using the front view image and the rotated image. That is, by paying attention to the feature points on the captured planar image, it is possible to obtain the movement angle that is the rotation angle of the true eyeball from which the error due to the individual difference is eliminated.
Therefore, according to the first invention, it is possible to obtain the movement angle with high accuracy while using a simple calibration method. In addition, the movement angle can be measured over time by performing a plurality of movement angle calculation steps.
Furthermore, according to 1st invention, since the test | inspection of a test subject's eye tracking over multiple times like the invention disclosed by patent document 1 is unnecessary, the test subject's burden and the tester side Any of the burdens can be eliminated.

  According to the second invention, in addition to the effects of the first invention, since the displacement is obtained independently in the horizontal direction and the vertical direction, the movement of the eyeball by changing the line of sight is changed for each direction. I can grasp.

  According to the third invention, the same effects as those of the first or second invention are obtained.

6 is a process diagram of a calibration method according to Embodiment 1. FIG. FIG. 5 is a longitudinal sectional view of an eyeball for explaining the operation of the calibration method according to the first embodiment. 6 is a schematic diagram for explaining angle pixel coefficients according to Embodiment 1. FIG. (A) is the image which image | photographed the test subject's eyeball from the front, (b) is a schematic diagram for demonstrating the movement angle which concerns on Example 1. FIG. (A) And (b) is the schematic diagram and graph for demonstrating the precision of the movement angle which concerns on Example 1, respectively. It is a block diagram of the eye movement test | inspection system which concerns on Example 2. FIG.

A calibration method according to an embodiment of the present invention will be described in detail with reference to FIGS. FIG. 1 is a process diagram of the calibration method according to the first embodiment. FIG. 2 is a vertical cross-sectional view of the eyeball for explaining the operation of the calibration method according to the first embodiment.
As shown in FIGS. 1 and 2, the calibration method 1 according to the first embodiment is a calibration method used in an eye movement inspection method that inspects the movement of the eye 50 by analyzing image data of the eyeball 50 of the subject. It is.
The calibration method 1 according to the first embodiment includes a front-view image capturing process in step S1, an angle pixel coefficient calculating process in step S2, a rotated image capturing process in step S3, and a movement angle calculating process in step S4. Prepare. Hereinafter, each of these steps will be described in detail.

First, the iris in the front view image capturing process in step S1, an index for identifying the movement of the eyeball 50, rotation angle theta about the eyeball rotation point 51 (theta _V is the vertical rotation angle) of rotation with a A front view image of the eyeball 50 in front view is captured by the front view image capturing means 4 (see FIG. 6) in association with the feature point 2 (indicated by X in the figure) included in the reference numeral 52. The feature point 2 is more accurately the center of the pupil 54.
The vertical rotation angle θ _V takes a positive value when rotating toward the upper eyelid 53a around the eyeball rotation point 51 with the horizontal axis X as a reference (θ _V = 0 radians), and the lower eyelid 53b side Takes a negative value when turning to. The horizontal rotation angle θ _H (not shown) is also a positive value when rotating to the right toward the eyeball 50 around the eyeball rotation point 51 with the horizontal axis X as a reference (θ _H = 0 radians). Take a negative value when turning to the left.

As will be described later, the front-view image photographing means 4 is a Frenzel camera 8 and is connected to a computer 9 (see FIG. 6) via a connecting means. The computer 9 is input with digital data of a front-view image output from the front-view image photographing means 4, and uses a recording unit 9b that stores the digital data, a display unit 9c that displays a front-view image, digital data, and the like. The control unit 9a includes a calculation unit 9e that performs all the calculation processes.
Accordingly, the iris 52 and the feature point 2 are clearly visible on the front view image, and the diameter d of the iris 52 is controlled by the control unit 9a from the digital data of the front view image stored in the computer 9. Can be extracted by expressing the pixel number P _X in the horizontal direction or the pixel number P _Y in the vertical direction. The position of the feature point 2 can also be extracted as coordinates by the control unit 9a. In addition, the size of the diameter d and the position of the feature point 2 are automatically extracted and determined by using the software in the calculation unit 9e of the computer 9, and the diameter d is determined from a keyboard 9d and a mouse 9f described later. It may be determined by manually inputting the position of the iris 52 and the position of the feature point 2 corresponding to both ends.

Next, the angle pixel coefficient calculation step in step S2 will be described in detail with reference to FIGS. FIG. 3 is a schematic diagram for explaining angle pixel coefficients according to the first embodiment. 1 and 2 are denoted by the same reference numerals in FIG. 3 and description thereof is omitted.
In the angle pixel coefficient calculation in step S2, as shown in Expression (1), first, the diameter d is divided by the pixel number p ₁ reflecting the diameter d of the iris 52 on the front view image, and the distance pixel coefficient β (See FIG. 3). The pixel number p ₁ is the extracted horizontal pixel number P _X or vertical pixel number P _Y. Therefore, the distance pixel coefficient β is a distance (mm / pixel) per pixel on the front view image.
The diameter d is a value obtained by multiplying the actual diameter d ₀ of the iris 52 (see FIG. 4) by the magnification k (k> 1) of the front-view image. This is because the iris 52 is actually located away from the eyeball turning point 51 at a turning radius D ₀ (see FIG. 4, more specifically, D ₀ is the distance from the eyeball turning point 51 to the center of the pupil 54). This is because it is projected on the front-view image at a distance D (D = kD ₀ ) away from the eyeball rotation point 51.
As the diameter d ₀ , an average value of a plurality of adults is used. Specifically, the diameter d ₀ is 11.50 (mm), and for an adult, the individual difference is slight.

The relationship between the distance pixel coefficient β and the angle pixel coefficient α ₁ (see FIG. 3) is given by Expression (2). Therefore, Expression (3) is established from Expressions (1) and (2).
That is, since the angle pixel coefficient α ₁ is a rotation angle θ that forms a distance of one pixel (that is, the distance pixel coefficient β) on the front view image, the unit of the angle pixel coefficient α ₁ is (radians / Pixel).
As the rotation radius D _0, the average value of a plurality of adults used. Specifically, the turning radius D ₀ is 9.77 (mm), and the individual difference is slight in the case of an adult as in the case of the diameter d ₀ .

Subsequently, the rotation image photographing process in step S3 will be described in detail with reference to FIG. 4A is an image obtained by photographing the eyeball of the subject from the front, and FIG. 4B is a schematic diagram for explaining the movement angle according to the first embodiment. The constituent elements shown in FIGS. 1 to 3 are denoted by the same reference numerals in FIG. 4 and the description thereof is omitted.
As shown in FIG. 4A, in the rotation image capturing process of step S3, when the iris 52 is rotated with the movement angle θ ₁ along the horizontal direction H around the eye rotation point 51 in association with the feature point 2. A rotating image of the eyeball 50 is taken. This rotated image is similarly photographed using the front-view image photographing means 4 (see FIG. 6) when the front-view image is photographed. Then, the digital data of the rotated image is output from the front-view image capturing unit 4 (that is, the rotated image capturing unit 6, see FIG. 6) to the computer.
Furthermore, it is also possible to repeat the rotation image photographing process in step S3 at regular intervals. This indicates, for example, the case where the rotated image photographing means 6 is a video camera.

As shown in FIG. 4B, when the iris 52 is rotated along the horizontal direction H with the movement angle θ ₁ , the feature point 2 (x mark) is changed from the position on the front view image to the position ( The number of moving pixels P ₁ corresponding to the displacement X ₁ of the feature point 2 when moving to the mark ● or ◇ is extracted from the digital data of the convolution image. This extraction is executed by the calculation unit 9e of the computer 9 as described above.
The displacement X1 is a displacement along at least _one of the horizontal direction H and the vertical direction V. Therefore, for example, when the iris 52 rotates along the vertical direction V with the movement angle θ ₁ , the displacement occurs along the vertical direction V (not shown). Even when the iris 52 is simultaneously rotated along the horizontal direction H and the vertical direction V with the movement angle θ ₁ , the displacement along the horizontal direction H and the vertical direction V is recognized independently by the computer 9. The
In the figure, β ₀ is a distance pixel coefficient on the iris 52, and β is a distance pixel coefficient on the front view image.

Finally, the movement angle calculation step of step S4 calculates the movement angle θ ₁ (radian) by calculating the product of the angle pixel coefficient α ₁ and the number of movement pixels P ₁ as shown in equation (4). . This calculation is also executed by the calculation unit 9e of the computer 9.
Further, the rotation imaging process of step S3, if repeated every predetermined time, the movement angle calculating step in step S4 is also repeated in the same manner, the temporal change of the moving angle theta ₁ is determined.

  By substituting equation (3) into equation (4), equation (5) is derived.

Thus, the movement angle θ ₁ is obtained from the known diameter d ₀ , the known rotation radius D ₀ , the number of pixels p ₁ extracted from the front view image, and the number of movement pixels P ₁ extracted from the rotation image. .
Further, Equation (4) is the angle pixel coefficient alpha ₁ is used if it is in a certain range are desirable. The reason for this will be described below with reference to FIGS. FIG. 5A and FIG. 5B are a schematic diagram and a graph, respectively, for explaining the accuracy of the movement angle according to the first embodiment. 1 to 4 are denoted by the same reference numerals in FIG. 5 and description thereof is omitted.
More precisely, since the distance pixel coefficient β ₀ on the iris 52 is represented by the distance pixel coefficient β ₀ ^* of Expression (6), the angle pixel coefficient α ₁ is represented by Expression (7).

The case where the expression (2) may be used instead of the expression (6) is a case where the angle pixel coefficient α ₁ is sufficiently small. This is because the following equation (8) is established.

Thus, the range in which the angle pixel coefficient α ₁ is within the range is described with reference to FIG.
As shown in FIG. 5A, α _1m is a rotation angle obtained by multiplying the angle pixel coefficient α ₁ by m, and α _m1 = m · α ₁ (m = 1, 2, 3,...). . β _m and β _n are distances corresponding to α _{n1 obtained} by multiplying α _m1 and α ₁ by n, respectively (n = m + 1 = 2, 3, 4,...).
In FIG. 5B, the horizontal axis represents | α _1m | (0 <| α _1m | ≦ π / 8), α ₁₁ = π / 180 (radian), β ₀ ^* = D ₀ sin α ₁₁ , β _n = D ₀ sin α _1n and β _m = D ₀ sin α _{1 m} , and the vertical axis represents the error rate R = [β ₀ ^* − (β _n −β _m )] / (β _n −β _m ) represented by the equation (9). It is the result.

Equation (9) represents the error rate of β ₀ ^* with respect to the true values (β ₂ -β ₀ ^* ), (β ₃ -β ₂ ), (β ₄ -β ₃ ).
As can be seen from FIG. 5B, if the allowable range of the error rate R is, for example, 5% or less, 0 <| α _1m | ≦ 0.31 (radian), that is, 0 degree <| α _1m | ≦ 18. Within the range of degrees, equation (2) may be used. Further, if this allowable range is, for example, 10% or less, within the range of 0 <| α _1m | ≦ 0.43 (radian), that is, 0 degree <| α _1m | ≦ 25 degrees, the expression (2) May be used.
Further, since both the turning radius D ₀ and the moving pixel number P ₁ are irrelevant to using the equation (2), only the error of the distance pixel coefficient β based on the equation (2) propagates to the moving angle θ. Therefore, in the former case, the expression (4) may be used within the range of 0 degree <| θ _1m | ≦ 18 degrees. Actually, since the maximum rotation angles | θ _H | and | θ _V | of the eyeball 50 are about 20 degrees, the error rate R is 5% or less.

As described above, according to the calibration method 1 according to the first embodiment, the movement angle θ ₁ can be easily calculated based on the equations (1) to (5). This movement angle θ ₁ is not affected by the enlargement factor k, as can be seen from equation (5). Therefore, the error is eliminated by individual differences of the human body as described above, can be obtained accurately moving angle theta _1.
In addition, since the angle pixel coefficient α ₁ is determined by using the constant diameter d ₀ and the turning radius D ₀ with almost no individual difference, the accuracy of the moving angle θ ₁ can be improved.
Furthermore, according to the calibration method 1, by performing the movement angle calculating step of convolutions image capturing step and step S4 of multiple step S3, because it can measure the movement angle theta ₁ over time, the eye 50 It is possible to reliably detect temporal changes in movement, that is, nystagmus.
Conventionally, a doctor can directly detect the presence or absence of nystagmus, for example, by directly observing the image of the camera. However, according to the calibration method 1, the reciprocation of the eyeball is performed in the horizontal direction H and the vertical direction. Since the direction V can be detected quantitatively, it can contribute to differential diagnosis of a disease causing nystagmus.
Furthermore, according to the calibration method 1, since the subject does not need to follow the gaze index with the eyes multiple times as in the invention disclosed in Patent Document 1, the burden on the subject side and the examiner side Any of the burdens can be eliminated.

Next, an eye movement inspection system that employs the calibration method according to the first embodiment and that is an eye movement inspection system according to the second embodiment of the present invention will be described with reference to FIG. FIG. 6 is a configuration diagram of the eye movement inspection system according to the second embodiment. The components shown in FIGS. 1 to 5 are denoted by the same reference numerals in FIG. 6 and the description thereof is omitted.
As illustrated in FIG. 6, the eye movement inspection system 3 according to the second embodiment is an eye movement inspection system that employs the calibration method 1 according to the first embodiment.
Eye movement test system 3 according to the second embodiment, the front view image capturing unit 4 for taking a front view image, the angle pixel coefficient calculating means 5 for calculating an angle pixel coefficients alpha _1, convoluted image photographing for photographing a convoluted image Means 6, movement angle calculation means 7 for calculating the movement angle θ ₁ , and connection means 10 are provided. Hereinafter, each of these means will be described in detail.

First, the front-view image photographing means 4 is specifically a Frenzel camera 8. The Frenzel camera 8 includes a video camera 8a and a recording unit 8b in which the output is recorded, and a display unit 8c.
More specifically, the video camera 8a is a CCD camera that can capture reflected light reflected by the eyeball 50 from the visible light incident on the eyeball 50 as a still image or a moving image. The front-view image taken by the video camera 8a is output as digital data to the recording unit 8b and is recorded by the recording unit 8b. The recorded digital data is displayed as an image of the iris 52 by the display unit 8c which is a liquid crystal panel.
The recording unit 8 b is connected to the control unit 9 a configuring the computer 9 via the connection unit 10 so as to be communicable.

Next, the angle pixel coefficient calculation means 5 is a calculation unit 9e provided in the control unit 9a. The calculation unit 9e calculates the distance pixel coefficient β and the angle pixel coefficient α _{1 according} to the calibration method 1 using the digital data of the front view image output from the recording unit 8b of the Frenzel camera 8.

  Further, the rotated image photographing means 6 is the same Frenzel camera 8 as the front-view image photographing means 4. That is, the front-view image photographing unit 4 is also a rotated image photographing unit 6.

The movement angle calculation unit 7 is the same calculation unit 9e as the angle pixel coefficient calculation unit 5. Therefore, the arithmetic unit 9e, using the digital data of the rotation image, follow the calibration method 1, it calculates the movement angle theta _1.

In addition to the control unit 9a, the computer 9 includes a recording unit 9b, a display unit 9c, a keyboard 9d, and a mouse 9f that are communicably connected to the control unit 9a.
Among these, the recording unit 9b records the digital data of the front view image and the rotated image, the angle pixel coefficient α ₁ , the movement angle θ ₁ , the position of the feature point 2 and the like in association with each other as appropriate. The recorded digital data and the like can be freely read out by the controller 9a.

Further, the display unit 9c is a display, and read the front view image and rotation image from the control unit 9a, an angle pixel coefficient alpha ₁ and can display a moving angle theta _1. Furthermore, when the movement angle θ ₁ changes with time, the display unit 9 c can also display the movement angle θ ₁ as a continuous waveform in a graph.
Furthermore, the keyboard 9d and the mouse 9f can input subject information and the position of the feature point 2 to the control unit 9a. The input subject information and the like are associated with the front view image, the rotated image, and the like, the angle pixel coefficient α ₁ and the movement angle θ ₁ are calculated by the calculation unit 9e, and recorded together with the calculation result and the like by the recording unit 9b. .
The connection means 10 connects the Frenzel camera 8 and the computer 9 so that they can communicate with each other, and is a conversion device for converting the digital data format of the Frenzel camera 8 into a format that can be processed by the computer 9. .

As described above, according to the eye movement inspection system 3, the calibration method 1 according to the first embodiment can be performed with a simple configuration such as the Frenzel camera 8 and the computer 9. θ ₁ can be automatically and accurately calculated.

The calibration method 1 according to the present invention and the eye movement inspection system 3 employing the calibration method 1 are not limited to those shown in the present embodiment. For example, a general digital camera or a Frenzel infrared camera may be used for the front-view image photographing means 4. In addition, when a Frenzel infrared camera is used, an infrared video camera is used instead of the video camera 8a.
Further, the Frenzel camera 8 and the computer 9 may have configurations other than those shown in the embodiments. For example, the recording unit 8b of the Frenzel camera 8 may not be provided. In this case, the connection means 10 connects the video camera 8a and the control unit 9a so that they can communicate with each other.

  INDUSTRIAL APPLICABILITY The present invention can be used as a calibration method used in an eye movement inspection method for inspecting eye movement and an eye movement inspection system employing the calibration method.

  DESCRIPTION OF SYMBOLS 1 ... Calibration method 2 ... Feature point 3 ... Eye movement test | inspection system 4 ... Front-view image imaging means 5 ... Angle pixel coefficient calculating means 6 ... Rotation image imaging means 7 ... Movement angle calculating means 8 ... Frenzel camera 8a ... Video camera 8b ... Recording unit 8c ... Display unit 9 ... Computer 9a ... Control unit 9b ... Recording unit 9c ... Display unit 9d ... Keyboard 9e ... Calculation unit 9f ... Mouse 10 ... Connection means 50 ... Eyeball 51 ... Eye rotation point 52 ... Iris 53a ... Up Eyelid 53b ... Lower eyelid 54 ... Pupil

Claims (3)

A calibration method used in an eye movement examination method for examining eye movement by analyzing image data of a subject's eye,
It is an index for specifying the movement of the eyeball, and captures a front view image of the eyeball when viewed from the front in association with a feature point included in an iris that rotates around the eyeball rotation point with a rotation angle. Front-view image shooting process,
An angle pixel coefficient calculation step of calculating an angle pixel coefficient based on a distance pixel coefficient obtained by dividing the diameter by the number of pixels reflecting the diameter of the iris on the front view image;
In association with the feature point, a rotation image photographing step of photographing a rotation image of the eyeball when the iris rotates with a movement angle around the eyeball rotation point;
The product of the angle pixel coefficient and the number of moving pixels corresponding to the displacement of the feature point when the feature point moves from a position on the front view image to a position on the convolution image, And a movement angle calculation step of calculating the movement angle.   The calibration method according to claim 1, wherein the displacement is a displacement along at least one of a horizontal direction and a vertical direction. An eye movement inspection system employing the calibration method according to claim 1 or claim 2,
Front-view image capturing means for capturing the front-view image;
Angle pixel coefficient calculating means for calculating the angle pixel coefficient;
A convolution image photographing means for photographing the convolution image;
And a movement angle calculation means for calculating the movement angle.

Images