NL2031115B1

NL2031115B1 - Method and system for determining a pupillary distance for a person.

Info

Publication number: NL2031115B1
Application number: NL2031115A
Authority: NL
Inventors: Waleson Jouke-Thiemo; Ismailov Fehim; Pieter Leendert Wisse Robert
Original assignee: Easee Health B V
Priority date: 2022-03-01
Filing date: 2022-03-01
Publication date: 2023-09-07

Abstract

A method for determining a pupillary distance for a person, comprising the steps of: - instructing the person to position oneselves within the field of view of a camera; - capturing an image of at least a portion of each iris of the person; - detecting, on the image, for each iris, a border between the sc|era and the iris; - on the image, using the detected border, obtaining an iris centre position for each eye; - on the image, determining a relative distance between the iris centre positions of the individual eyes; - on the image, based on said detected border, determining a relative iris diameter for at least one iris; - correlating said relative iris diameter with a pre-defined absolute diameter; and - based on said correlation and the determined relative distance between the iris centre positions of the two individual eyes, obtaining the absolute pupillary distance.

Description

Title: Method and system for determining a pupillary distance for a person.

BACKGROUND OF THE INVENTION

The present disclosure relates to determining a pupillary distance (PD) for a person. In particular, the present disclosure relates to determining a pupillary distance for a person in the context of performing an eye test in an “online” format, the eye test being executed by the user on his/her own computer device, at a location of his/her convenience, without the help of an instructor / supervisor / examiner and with the minimal use of any hardware equipment (except for the computer device on which the test is performed), in particular without the use of any hardware equipment that needs to be send to the user / bought by the user especially / exclusively for the purpose of being able to carry out the test.

The pupillary distance, as is known in the art, relates to the distance between the centre of the left pupil and the centre of the right pupil. This distance is typically included in a prescription of glasses, lenses of the glasses preferably being cut in such a way that the lenses are the sharpest right in front of the pupil centres.

US 2020/0054209 A1 relates to a system and method for measuring the pupillary distance. According to US 2020/0054209 A1 a method is implemented on a computing device having at least one processor, storage, and a communication platform connected to a network. The method includes obtaining a head image of a user with one or more dimension indicators, determining an eye region in the head image of the user captured with a camera and determining a pupillary distance of the user based on the one or more dimension indicators and the determined eye region.

According to US 2020/0054209 A1 the dimension indicator e.g. corresponds to especially designed glasses frames, e.g. of paper, e.g. to be cut out by the test person in advance of the test, the glasses frames including a ruler at the top of the frame. Specially designed software can recognize both the centres of the pupils in the captured head image and the ruler. From these data points, knowing the scale of the ruler, the pupillary distance can be obtained.

The method explained in US 2020/0054209 A1 suffers from a few disadvantages. First of all, the method makes it cumbersome for a user to carry out an eye test after purchasing it / makes the bar for purchasing the eye test higher. A glasses frame including a ruler at the top of the frame is not a piece of equipment that is found in most households and it should be send to the user in order to perform the test of US 2020/0054209 A1. This would mean that one purchases the test, gets send the glasses and only then can carry out the test. This results in a delay of at least about 12 hours and likely longer, which is highly inconvenient to the user. Alternatively, the user may be able to print out the glasses, but also printers are more and more disappearing as standard household items. Even so, when the glasses can be printed they have to be cut out subsequently, possibly leading to errors and for sure leading to a delay in starting the eye test of at least 10 — 15 minutes (which, to put in in the right perspective, is about 0.5 — 1 times the total time the test takes). This is all highly undesirable.

An additional disadvantage of the method explained in US 2020/0054209 A1 is that the paper frame, when seen from the perspective of the camera, can block a portion of the eyes of the person, so that the centre of the pupil can be determined with less accuracy.

Accordingly, objects of the present disclosure are to improve the method as explained in US 2020/0054209 A1 by allowing it to be run semi-instantaneously with as less waiting/processing time as possible and with an as clear picture of the eye as possible.

SUMMARY OF THE INVENTION

Therefore, a first aspect of the present disclosure relates to a computer- implemented method for determining a pupillary distance for a person, wherein use is made of a camera configured for capturing images, the method comprising the steps of: - instructing the person to position oneselves within the field of view of the camera; - capturing, with the camera, an image of at least a portion of a face of the person, said face portion comprising at least a portion of each iris of the person; - detecting, on the image captured with the camera, for each iris, a border between the eye white (sclera) and the iris; - on the image captured by the camera, obtaining an iris centre position for each eye;

- on the image captured by the camera, determining a relative distance between the iris centre positions of the individual eyes and/or the nose bridge; - on the image captured by the camera, based on said detected border, determining a relative iris diameter for at least one iris; - correlating said relative iris diameter with a pre-defined absolute diameter; and based on said correlation and the determined relative distance between the iris centre positions of the two individual eyes and/or the nose bridge, obtaining the absolute pupillary distance. Advantageously, the present method can be performed while a person is sitting in front of the camera / computer screen, without the need for any equipment besides the camera and the computer itself. It is noted that most consumers are nowadays in the possession of e.g. a laptop with a built-in camera and/or a desktop computer with a built-in or separately connected webcam camera.

As the pupillary distance is a part of the prescription of the user, the eye test effectively is already started when the present method is carried out, such that there is no delay at all in carrying out the eye test. Advantageously, accurate results may be obtained with the present method, especially if the method is performed for subsequent images and the results are compared with each other until a stable value for the PD is obtained. Advantageously, in most cases the PD may be obtained within maximum 60 seconds.

In accordance with the disclosure, the method is implemented on a computer.

Within the context of the present disclosure, in principle any device including a processor is deemed a computer. The most notable examples may include a laptop, desktop, tablet and mobile smartphone, but possibly smartwatches and other wearable device may be suitable for carrying out the present method as well.

In accordance with the disclosure, the pupillary distance (PD) is determined for a person. In this document, said person may alternatively be referred to as ‘user’, ‘test subject’ or ‘patient’, all terms referring to the same: the person taking the eye test. As stated in the above, preferably the PD is obtained by the person him/herself, without the assistance of a second person and without the use of dedicated equipment.

In accordance with the disclosure, use is made of a camera. It should be noted that, when an eye test is offered for sale on the Internet, it may be bought by many different persons having vastly different hardware set-ups. As such, when offering an eye test for sale over the Internet, it is typically not known in advance which hardware setup a user has. This leads to the situation that it is not known what the parameters / specifications of the camera with which the images will be captured are. In particular, a difficulty when selling eye tests over the Internet is that when it is e.g. not known what the focal length of the camera is, any relative measurement / calculation performed on relative distances on an image recorded by the camera cannot be converted to an absolute, real-life distance. Accordingly, as will be explained in more detail in the below, some kind of calibration or correlation must be performed.

In accordance with the present disclosure, the person of whom the PD is to be determined is instructed to position him/herself in front of the camera, in the field of view of the camera. For example the field of view of the camera may be displayed on a computer screen associated with the camera, in real time, so that the user can check whether his/her face is visible by the camera. Preferably the person is at a stable distance of in between 10 cm and 1 meter from the camera, more preferably less than 60 or even more preferably about 30 cm from the camera. In general, the closer the person is to the camera, the higher the resolution of the image captured with the camera and the more accurate the method. However, the person should not be so close that e.g. only one eye is visible by the camera, as then the PD cannot be determined. Preferably, the person is instructed to look straight towards/into the camera, the person preferably being positioned right in front of the camera.

In accordance with the present disclosure, the camera captures an image of at least a portion of the face of the person. It is thus not necessary that the entire face of the person is visible to successfully carry out the method. When only a part of the face is visible, namely the part including a portion of both irises, the method can be carried out successfully. Moreover, as will be understood from the below, it is not even necessary that the eyes are entirely visible, only a part of the eyes may need to be captured by the camera in order to successfully carry out the method. In particular, preferably the right part of the left iris and the left part of the right iris need to be on the image captured by the camera for the method to be carried out. It is not necessary, but it may be preferred, that the entire iris is visible. Also the pupil centres themselves need not be visible on the image captured by the camera.

Seeing only a part of the iris may be sufficient to accurately estimate / extrapolate the layout of the remainder of the eye. The exact percentage of the iris that needs to 5 be visible depends on many factors, including the quality of the camera, the lighting conditions in the room and the training of the software, but less than 50% will generally be enough. In accordance with the disclosure, preferably both irises of the patient are visible on the same image recorded by the camera.

In accordance with the present disclosure, for both irises on the image a border between the sclera (= eye white) and the iris is detected. When the iris is fully visible this may lead to the detection of a circumferential border line; when the iris is not fully visible this may lead to the detection of a curved line that is not closed.

However, from the part of the iris that is visible (i.e. the part of the border that can be detected) as well as from prior knowledge about the shape of an iris, a closed circumferential border line may be extrapolated. For example, one may assume that the iris is a perfectly round circle, or be represented by a mathematical formula deviating somewhat from the round circle. It is however not required per se that a circumferential border line of the iris is obtained at all during the method.

In accordance with the present disclosure, the centre of each iris is obtained.

It may be convenient to determine the position of the centre of each iris from the previously-detected border between the sclera and the iris, wherein it may be assumed that the position of the centre of the iris corresponds to the position of the centre of the pupil. It is noted that whereas the border between the eye white and the iris can be determined rather reliably, it may be much more difficult to determine the border between the iris and the pupil, especially for users having dark (brown) eyes.

Also, the size of the pupil fluctuates quite rapidly with altering light conditions, so that in consecutive measurements the border between the iris and the pupil may be at different positions whereas the border between the iris and the eye white will remain at roughly the same position in subsequent images, assuming the head and eyes of the user are not moved. Hence, determining the centre of the pupils from the border between the iris and the eye white results in accurate and stable results.

An alternative working embodiment may however rely on a determination of a border between the iris and the pupil, and obtain the centre position of the pupil,

corresponding to the centre of the iris, from said latter border. In the context of the present document, it is assumed that the iris centre position and the pupil centre position relate to the same position; the two expressions being used as synonyms.

In accordance with the present disclosure, when the centre positions of both irises are identified on the captured image, the relative distance between them can be determined. The word “relative” is used here to indicate that the distance is a distance as determined from an image, and does not correspond to a “real life” distance. For example, the relative distance between the centre positions of the irises can be computed in pixels and/or in % image width. As an alternative, or in addition, to obtaining the relative distance between the iris centre positions, the centre position of the nose bridge may be obtained after which the relative distance between the centre of the left iris and the nose bridge as well as the relative distance between the centre of the right iris and the nose bridge may be obtained.

In accordance with the present disclosure, from the border between the iris and the eye white, previously obtained, also the diameter of the iris, on the image, may be obtained. This iris diameter is known to one skilled in the art as the ‘white-to- white’ distance. Also this distance is relative, as it is determined from the image captured by the camera, and can e.g. be expressed in pixels and/or % image width.

It has been found by the present inventors that, in absolute numbers / in the physical world, this eye white to white distance is highly constant both during the day and among the human population, irrespective of factors such as age, gender, race and others. Hence, when the relative iris diameter as captured by the camera is known, e.g. in pixels, this can be correlated to the absolute, real life, iris diameter without needing to perform any measurements on the patient. From this correlation, one can e.g. determine the scaling on the image as captured by the camera, e.g. in terms of number of pixels per real life millimetre.

In accordance with the present disclosure, having determined the scaling between the relative and the absolute iris diameter, the same scaling factor can be used to determine the absolute PD from the relative distance between the iris centre positions as determined from the image. As is conventional in the art of eye prescriptions, the PD may be calculated in two different formats: a ‘single’ PD value, corresponding to the distance between the centres of the irises/pupils, and/or a UR’

PD value, containing two values, of on the one hand the distance between the centre of the left eye pupil/iris and the centre of the nose bridge and on the other hand the distance between the centre of the right pupil/iris and the centre of the nose bridge.

According to the method presented herein one or both of these PD values may be obtained.

It is noted that the steps of the method as described in the above, in particular the steps related to determining relative distances, are preferably performed on a single image, so that the scaling between the “real” dimensions and the “relative” dimensions are the same for the entire image. It may be possible that thereafter another, second, image is captured and that the method is performed a second time to determine a second value for the absolute PD.

It is furthermore noted that it is not essential that the steps of the method are carried out in the order as described in the above, although some steps must obviously be performed before other steps can be performed. For example, before the relative distance between the iris centre positions of the individual irises can be obtained from an image, the image first has to be captured by the camera. However, as one skilled in the art will understand, it may be possible to determine the relative iris diameter before the relative distance between the iris centres is determined.

In an embodiment of the present disclosure, the method may include the step of determining the relative iris diameter for each of the left and the right eye, and additionally the step of computing a difference between the relative iris diameter of the left eye and the relative iris diameter of the right eye, wherein the step of obtaining the absolute pupillary distance is only performed if said difference is below a pre-defined threshold. In absolute values, the left iris and the right iris will have virtually the same diameter. When the diameter as computed from the image varies too much between the left and the right iris, this may e.g. indicate that the user has looked into the camera at an angle instead of in a straight line. This could lead to inaccuracies when the PD is calculated. Therefore, any such images may be neglected in the PD calculations. For example, a too large variation may be a variation of more than 1% or more than 3%.

In an embodiment of the present disclosure, the method may additionally include the step of instructing the person to look into the camera while capturing the image. It is found by the inventors that when a person is instructed to look into the camera, the eyes focus on a certain point by rotating towards each other somewhat, and the PD is less then when the same person is looking into the far field. In several scenarios, this effect can be used to the advantage of the method.

One scenario is when the prescription of the user turns out to contain a positive dioptre, i.e. a correction for near-field vision. Namely, when using the prescribed glasses a hyperopic person will focus his/her eyes at a near point, resulting in a lower (effective) PD. In such cases it may be advantageous to cut the lenses taking the smaller, near-field, PD value into account instead of the larger, far- field, PD. Preferably, in this scenario, the person is sitting relatively close to the camera when an image of a person is captured, at about the same distance as the person would be to a book / a phone while reading from it and wearing the near-field corrective glasses.

Another scenario in which the above effect can be used to the advantage of the method is when the distance between the patient and the camera is known. In that case, when the patient looks into the camera, the amount of eye rotation can be calculated and from the obtained “near field PD” also the “far field PD” can be determined by accounting for said rotation.

Accordingly, in an embodiment of the disclosure the method may additionally include the step of instructing the user to position themselves at a pre-defined distance from the camera. For example, the user may be instructed to measure a distance between him/herself and the camera, e.g. by placing an A4-sized paper, in the length direction, with one side against the head/eyes of him/herself and with the other side against the camera so that about 29,7 cm In embodiments it may be possible to further check whether the user is indeed at the correct distance, e.g. with methods known for determining a distance between a user and a camera, but it may be sufficient to assume that the user is indeed at the instructed position with a high enough degree of accuracy.

In an embodiment of the disclosure, the camera is integrated with or associated with a computer screen, and the method may additionally include the step of displaying an object on the computer screen, preferably near the camera position.

As users tend to be distracted quite fast and as it is relatively uncomfortable for a user to look straight into a camera of a computer setup, to help the user in looking semi into the camera an object may be displayed on a computer screen, the display position of the object preferably corresponding to a position near the camera, e.g. near the top centre of the screen. In such cases, the user can simply be instructed to look at the displayed object while the camera records an image of (the eyes of) the user. An additional advantage of displaying an object on the computer screen is that the chance that a user looks straight (i.e. at a 180 degree angle) towards this object — and thus: into the camera — is larger than when no an object is displayed.

In an embodiment of the disclosure, the method may additionally include the step of altering a shape, position, and/or colour of the object displayed on the computer screen, while repeatedly performing the method. As will be described in more detail in the below, the method may benefit from it being executed repeatedly for a number of images, e.g. for between 10 — 200 images, the images e.g. being 0.001 — 0.1 seconds apart from each other. Depending on the exact setup chosen, this will require users to look at the camera for a relatively long period of time. As attention of users is easily lost, it may be advantageous to alter the displayed object, so that the user is focussed at the altered object and the eyes of the user remain in view of the camera while performing the method several times.

In an embodiment of the disclosure the method may additionally include the step of instructing the person to look at a position behind the camera, preferable more than 3 meters behind the camera, while capturing the image. This way, the “far field” PD may be obtained, the PD when the eyes are looking straight ahead..

It is possible for the method to only obtain the “far field” PD, while instructing the user to look behind the camera, to obtain both the “near field” PD and the “far field” PD, and it is possible to only obtain the “near field” PD. Myopic patients may benefit from their lenses being cut for their “far field” PD. When the pupillary distance is determined before other parts of the eye test are carried out, it is recommended to obtain both PD’s.

In an embodiment of the disclosure, the method may be performed multiple times, by capturing multiple, individually unique, images, the images captured at subsequent time instances. In such an embodiment the value of the absolute pupillary distance may be determined while taking account of the previously determined absolute pupillary distance(s), for each second and subsequent image.

As such, a “running average” of the PD may be computed, in which the individual measurements over consecutive images are averaged out. Therefore, first the absolute PD is e.g. calculated for the latest image, after which the latest-calculated

PD is compared with the previous evaluations and an average thereof is computed.

As known, to one in the respective art, there is a great many different number of ways in which an average of a sequence of numbers can be determined.

When determining the PD in such a way, the average PD value may be computed until it doesn’t vary for a pre-determined number of frames. For example, the PD may be calculated with an accuracy of 0.1mm, until it remains constant for at least 5, such as at least 10, e.g. at least 15 images. For example, 1 — 60 images may be taken per second, the timeframe between two consecutive images preferably not exceeding the computation time that is needed to perform the method once.

In embodiments, when it is determined that the running average of the PD value fluctuates too much, it may be possible to “reset” the running average by deleting / ignoring all or a portion of the previously obtained results and re-starting the calculation, in order to converge to a deemed correct value of the PD faster.

In an embodiment of the disclosure, the method may additionally include the step of instructing the person to take off ones glasses before performing the method.

It was found that if the glasses are kept on during the capturing of the image, inaccurate results for the PD are obtained. This most likely results from light rays captured by the camera being broken by the lenses of the glasses and the camera viewing a distorted image of the face of the person. This negative effect is mitigated when the person takes off his or her glasses.

In an embodiment of the disclosure the method includes the additional step of asking the person to input their eye colour. Depending on the eye colour, it may be harder or more easy to distinguish between the sclera (eye white) and the iris. When the eye colour is inputted to a border recognition program, this may be made more easily.

In an embodiment of the disclosure the method includes the additional step of instructing the person to remove any objects in between their eyes and the camera, to obtain the clearest image of the eyes of the patient as possible.

A second aspect of the present disclosure relates to a system for determining a pupillary distance for a person, the system comprising:

- a camera configured for capturing images of at least a portion of a face of a person positioned before the camera, said face portion comprising at least a portion of each iris of the person; and - a processor arranged in communication with the camera, the processor configured for: - receiving, from the camera, a captured image; - on the image, for each iris, detecting a border between the eye white (sclera) and the iris; - onthe image, using the detected border, obtaining an iris centre position for each eye; - on the image, determining a relative distance between the iris centre positions of the two individual eyes; - on the image, based on the detected border, determining a relative iris diameter for at least one iris; - correlating said relative iris diameter with a pre-defined absolute diameter; and - based on said correlation and the determined relative distance between the iris centre positions of the two individual eyes, obtaining the absolute pupillary distance.

Advantages and effects achieved with the system according to the second aspect are the same as the advantages and effects achieved with the method according to the first aspect of the present disclosure. It goes without saying that optional embodiments described in the above with reference to the method, may equally advantageously be applied to the system according to the second aspect of the disclosure.

These and other aspects of the present invention are now elucidated in the below, with reference to the attached figures. In these figures, like or same features are indicated with the same reference numbers.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 schematically shows a person in front of a computer screen, a camera associated with the camera screen, the computer screen displaying what is seen by the camera;

Figure 2 schematically — in high resolution — shows a first possible image captured by the camera as shown in Figure 1; and

Figure 3 schematically — in low resolution — shows a second possible image captured by the camera as shown in Figure 1.

DETAILED DESCRIPTION OF THE DRAWINGS

Shown in Figure 1 is a person 2 in front of a computer screen 5. Associated with the computer screen 5 is a camera 3, the field of view of the camera 3 being displayed on the computer screen 5. In the field of view of the camera 3 is a part 21 of the body of the person 2, in particular a part 21 of the face. Included in the part 21 of the face are, here, both eyes 32, 33. Note that the person 2 is not wearing any glasses, even though they may be worn in daily life. Also note that there are no objects in between the camera 3 and the eyes 32, 33 of the person 33.

In accordance with the present method it may be preferred that the person 2 looks straight into the camera 3 (as shown). Accordingly, the method may include a step of instructing the person 2 to look straight into the camera 3. To help the person 2 look (almost) into the camera 3, an object 6 may be displayed on the screen 5, preferably near the position where the camera 3 is integrated with the screen 5. To help the person 2 look at the camera for a longer period, it may be possible to e.g. alter the shape, colour or position of the object 6 while repeatedly performing the method as described in the below.

Alternatively, the person 2 may be instructed to look at a position behind the camera 3.

For example, the person 2 may be instructed to position herself at a pre- defined distance from the camera 3 / screen 5, e.g. about 30 centimetres, while an image of the person 2 is being captured by the camera 3. For example, this distance may be obtained by measuring a distance between the eyes 32, 33 and the screen 5, e.g. with a ruler or the like. Object of the method is to determine the distance 1 between the centres of the pupils of the person 2, known in the art as the pupillary distance, PD. The PD is here indicated as a single value, from the centre of the left pupil to the centre of the right pupil. The PD may alternatively and/or additionally be expressed as a ‘right side PD’, corresponding to the distance 1a between the centre of the right pupil and the centre of the nose bridge and a ‘right side PD’, corresponding to the distance 1b between the centre of the left pupil and the centre of the nose bridge. When summing the ‘right side PD’ and the ‘left side PD’, the single PD value should be obtained. For that purpose, an image captured by the camera 3 is analysed. An example of such a captured image is shown in Figures 2 and 3.

Figures 2 and 3, then, show an image 4 as captured by the camera 3 when the person 2 is situated in front of the camera 3, as in Figure 1. In figure 2 an image 4 is displayed that is taken with a camera 3 having a high resolution (many pixels, each pixel capturing a tiny portion of the face), whereas in figure 3 an image 4 is displayed that is taken with a camera having a low resolution (few pixels, each pixel capturing a larger portion of the face). It is noted that the image 4 is a virtual one, and that the scaling of the image 4 may significantly differ from the real life dimensions. However, the scaling is not a priori known as the parameters of the camera 3 may not be known a priori. Hence, determining the PD of the person from the image 4 is not possible by e.g. counting the number of pixels in between the centres of the pupils.

Instead, according to the present method a border 24, 25 between each iris 22, 23 and the eye white 26 is detected on the image 4. Note that in the shown image 4 the iris 22, 23 is not fully visible, so that only a part of the border 24, 25 can be detected. However, this is sufficient for the objective of the present method. An even smaller portion of the iris 22, 23 may be visible and the method may still be carried out successfully.

Having obtained the border 24, 25 between the iris 22, 23 and the eye white 26, the centre positions 27, 28 of each iris 22, 23 as well as the relative diameter 30, 31 of the irises 22, 23, as determined on the image 4, can be obtained. For example, a particular pixel may be selected as the centre position 27, 28 of the iris 22, 23, and the relative diameter 30, 31 of the irises 22, 23 can be expressed in number of pixels.

From the centre positions 27, 28 of each of the irises 22, 23, a relative distance 29, as measured on the image 4, between the individual centre positions

27, 28 can be obtained. Alternatively and/or additionally, the method may include a step of determining a centre position of the nose bridge 35 and obtaining a relative distance 29a between the centre position of the right iris 27 and the centre position of the nose bridge 35 as well as a relative distance 28b between the centre position of the left iris 28 and the centre position of the nose bridge 35. This relative distance 29, 29a, 29b can again e.g. be expressed in number of pixels.

As it is found by the present inventors that the real life value of the iris diameter is relatively constant among the human population, irrespective of age, gender, race, time of day and other parameters, from the relative iris diameter 30, 31 as determined for the image 4, a correlation between the relative distances on the image 4 and physical, real world distances may be made. For example, knowing the “true” value of the iris diameter, and having obtained the “relative” iris diameter on the image 4, a correlation can be made, for the particular image 4 that is analysed, between “true” and “relative” dimensions. In particular, it may be possible to determine the “width” of one pixel, in the physical space or the number of pixels per captured length unit.

When for the same captured image the iris diameter 30, 31 and the distance 29, 29a, 29b between the iris centre positions 27, 28 and/or the nose bridge 35 have been determined, from the above-mentioned correlation also the distance 29, 29a, 29b can be converted to a “true” or “physical” dimension, the (absolute) PD.

It is to be understood that if this method is carried out once, for one captured image, there may be a relatively large error in the calculation. Accordingly, it is preferred if the method is executed several times, each time for a newly captured image 4, while an average of all computed PD values is recorded and outputted. For example, the method may be executed for a fixed number of times (e.g. in the order of several tens of times, e.g about a hundred time) or the method may be executed as often as is needed to obtain a stable average PD value over all images.

It is to be understood that the method becomes more accurate when the person is positioned straight in front of the camera 3 (irrespective of whether the person looks into the camera or behind the camera). Accordingly, the method may compute the relative iris diameter 30, 31 for each of the left and the right eye, and compute a difference between the relative iris diameters 30, 31. When the difference is larger than a pre-defined threshold, this is an indication that the person is not looking straight into the camera, but rather at an angle.

Accordingly, such an image may be disregarded in determining the PD.

Claims

CONCLUSIONS

1. A computer-implemented method for determining an interpupillary distance (1, 1a, 1b) for a person (2), using a camera (3) designed for capturing images (4), wherein the method includes the steps of: - instructing the person (2} to position themselves within the field of view of the camera (3); - capturing, with the camera (3), an image (4) of at least a part of a face (21) of the person (2), wherein the facial part (21) includes at least part of each iris (22, 23) of the person (2); - detecting, on the image (4) captured with the camera (3), for each iris (22, 23), of a boundary line (24, 25) between the white of the eye (sclera) (26) and the iris (22, 23); - on the image (4) captured by the camera (3), obtaining an iris center position (27, 28) for each eye (32, 33); - on the image (4) captured by the camera (3), determining a relative distance (29, 29a, 29b ) between the iris center positions (27, 28) of the individual eyes (32, 33) and/or the nasal bridge (35); - on the image (4) captured by the camera (3), based on the detected boundary line (24, 25), determining a relative iris diameter (30, 31) for at least one iris (22, 23); - correlating the relative iris diameter (30, 31) with a predefined absolute diameter; and - based on the correlation and the determined relative distance (29, 29a, 29b) between the iris center positions (27, 28) of the two individual eyes (32, 33) and/or the nose bridge (35), obtaining the absolute pupillary distance (1, 1a, 1b).

The method of claim 1, wherein the method includes the step of determining the relative iris diameter (30, 31) for each of the left and right eyes, and wherein the method additionally includes the step of calculating a difference between the relative iris diameter (30) of the left eye (32) and the relative iris diameter (31) of the right eye (33), where the step of obtaining the absolute pupillary distance (1, 1a, 1b) is only performed if the difference is below a predefined threshold value.

The method according to claim 1 or 2, wherein the method additionally comprises the step of instructing the person (2) to look into the camera (3) while the image (4) is being captured.

The method of claim 3, wherein the method additionally comprises the step of instructing the person (2) to position themselves at a predefined distance from the camera (3).

The method according to claim 3 or 4, wherein the camera (3) is integrated with or associated with a computer screen (5), and wherein the method additionally comprises the step of displaying an object (6) on the computer screen (5). ), preferably near the camera position.

The method according to claim 5, wherein the method additionally comprises the step of adjusting a shape, position and/or color of the object (6) displayed on the computer screen (5) while repeatedly executing the method.

The method according to any of the preceding claims, wherein the method additionally comprises the step of instructing the person (2) to look at a position behind the camera (3), preferably more than 3 meters behind the camera ( 3), while the image (4) is captured.

The method according to any one of the preceding claims, wherein the steps of determining a boundary line (24, 25) of an iris (22, 23), obtaining an iris center position (27, 28), determining a relative distance (29, 29a, 29b), and determining a relative iris diameter (30, 31) are all performed based on the same captured image (4).

The method according to any of the preceding claims, wherein the method is performed multiple times, by capturing multiple, unique, images (4), wherein the images (4) are captured at successive times, and where, for every second and next image (4), the value of the absolute interpupillary distance (1, 1a, 1b) is determined based on the currently evaluated image (4) as well as the previously determined absolute interpupillary distance(s) (1, 1a, 1b).

The method according to any of the preceding claims, wherein the method additionally comprises the step of instructing the person (2) to remove the glasses before performing the method.

The method according to any one of the preceding claims, wherein the method additionally comprises the step of asking the person (2) to input their eye color.

The method of any preceding claim, wherein the method additionally includes the step of instructing the person (2) to remove any object between their eyes (32, 33) and the camera (3).

The method according to any one of the preceding claims, wherein the absolute pupillary distance (1, 1a, 1b) is expressed as a right eye-pupil distance (1a) and a left eye-pupil distance (1b).

14. A system for determining an interpupillary distance (1, 1a, 1b) for a person (2), wherein the system comprises: - a camera (3) adapted to capture images (4) of at least part of a face (21) of a person (2) positioned in front of the camera (3), wherein the facial portion (21) includes at least a portion of each iris (22, 23) of the person (2); and - a processor arranged in communication with the camera (3), wherein the processor is adapted to: - receive, from the camera (3), a recorded image (4); - on the image (4), for each iris (22, 23), detecting a boundary line (24, 25) between the white of the eye (sclera) (26) and the iris (22, 23); - on the image (4), obtaining an iris center position (27, 28) for each eye (32, 33); - on the image (4), determining a relative distance (29, 29a, 29b) between the iris center positions (27, 28) of the two individual eyes (32, 33) and/or the nose bridge (35); - on the image (4), based on the detected boundary line (24, 25), determining a relative iris diameter (30, 31) for at least one iris (22, 23); - correlating the relative iris diameter (30, 31) with a predefined absolute diameter; and - based on the correlation and the determined relative distance (29, 29a, 29b) between the iris center positions (27, 28) of the two individual eyes (32, 33) and/or the nose bridge (35), obtaining the absolute pupillary distance (1, 1a, 1b).