Methods for assessing tear film stability

Abstract

A method for measuring the stability of a patient's tear film by means of an apparatus comprising: a backlit translucent plate having a test pattern positioned in front of at least one eye (101) of a patient (100); at least one digital camera connected to a computing system having image processing and analysis means, the camera lens being directed towards the patient's eye so as to photograph the reflection of the test pattern on the patient's eye, the method comprising, from a start time (t0) consisting of an eyelid blink: capturing a series of images (200); detecting a series of micro-movement areas (510) in the images and calculating and storing data (520) on the position, surface area and number of micro-movement areas of the tear film in each image using the image processing and analysis means; and calculating (530) and storing (540) a measurement of the amplitude of the micro-movement areas as a function of time based on the continuously calculated position, surface area and number of micro-movement areas of the tear film.

Description

Technical Field

[0001] The present invention relates to the field of detection of dry eye syndrome. Background Art

[0002] Tests used to assess dry eye syndrome conventionally employ tests involving fluorescence. The physician places a drop of fluorescein into the patient's eye and observes the fluorescence using a slit lamp. The physician determines the time between the eyelid blinking and the appearance of the first non-fluorescent area. This time is a measure of the tear film's "breakup time," or TBUT. This method allows for the measurement of the first breakup time, regardless of whether the breakup is due to a local defect, such as a speck of dust, or a true dry eye problem. The time measured is approximately 6 to 10 seconds.

[0003] Another evaluation method for measuring the so-called "non-invasive break-up time" or NIBUT is a non-invasive method for measuring the break-up time. Tests using this method use, in particular, the reflection of a test pattern from the cornea, which has a pattern consisting of concentric circles and radial lines.

[0004] When the film breaks, the lines and / or circles of the reflected pattern are distorted or areas of the pattern are missing.

[0005] Images of the reflected pattern are acquired sequentially by a camera and processed by a computerized processing system which analyses the changes in the pattern over time in order to detect deformations of the pattern and to infer therefrom the time of occurrence of the membrane rupture zone.

[0006] Technical issues

[0007] Currently available devices that utilize circular patterns and radial lines cause several problems:

[0008] Due to the shape of the test chart and the extent of the cornea covered by the disk segments, the spatial resolution at the edges of the field of view is inherently worse than at its center.

[0009] Furthermore, at the center of the field of view where the radial lines converge, it is difficult to determine the deformation of the pattern.

[0010] To increase the spatial resolution, the density of the test pattern must be increased. Furthermore, the algorithms used to detect the deformation of the circle (which requires finding the most likely circle in the image), whether a point in the image belongs to the possible circle to be evaluated, and the distance from a given point to the circle to be calculated are all complex, and the processing and computing time required to detect the cracked area based on the image is considerable.

[0011] It will thus be clear that, on the one hand, devices employing conventional patterns of concentric circles and radial lines suffer from a lack of spatial resolution uniformity, and that, on the other hand, increasing the resolution at the edges of the field of view can have a detrimental impact on computation time.

[0012] Due to the low resolution of this method, the measurements can produce inaccurate and overestimated rupture times.

[0013] Of particular note, measurements obtained using fluorescence and noninvasive methods generally do not correlate well: a 2-fold difference between TBUT and NIBUT measurements is often observed, with the fluorescence method returning a shorter time. This suggests that the device used to measure NIBUT lacks sensitivity or resolution.

[0014] Likewise, NIBUT values measured using different available devices do not correlate well.

[0015] In the diagnosis of dry eye syndrome, neither breakup time measurement (obtained by fluorescein) nor NIBUT measurement correlates with patient-reported distress.

[0016] The aforementioned methods do not reveal the overall status of the tear film, which is not useful for physicians who require higher measurement accuracy and repeatability for diagnosis. Summary of the Invention

[0017] The present invention improves this situation and provides a method for measuring the stability of the tear film on a patient's cornea based on detecting modifications of this film using a device comprising a pattern composed of light and dark horizontal or vertical lines, and provides observation of the reflection of this pattern from said cornea.

[0018] To this end, in the context of the present patent application, micro-movements in the tear film are sought via analysis of images of the lines of a test pattern.

[0019] Local variations in the total thickness of the tear film are called micromovements. This variation induces a slope in the surface of the tear film, which is evidenced by local modifications of the images of the lines of the test chart and in particular by modifications or local deformations of the width of the images of these lines.

[0020] More specifically, the present patent application provides a method for measuring the stability of a patient's tear film with the aid of a device, the device comprising: a backlit translucent plate provided with a test pattern, the test pattern being positioned in front of at least one eye of the patient; and at least one digital camera connected to a computing system having means for processing and analyzing images, the objective lens of the camera being directed toward the patient's eye in order to capture reflections of the pattern of the test pattern from the patient's eye, characterized in that the test pattern comprises a pattern consisting of a series of alternating light and dark lines reflected from the patient's eye, the means for processing and analyzing images being configured to detect reflections of the pattern of the test pattern from the patient's eye. The method comprises the following steps: detecting deformations of the light or dark lines of the pattern of the test chart reflected by the eye and identifying tear film micromovements revealed by these deformations by comparing the positions of image points on the edges of the lines relative to the estimated line edges of the lines, the method comprising: continuously capturing images from the start time t0 of an eyelid blink; continuously detecting areas of micromovement in the images, and continuously calculating and storing data on the positions and number of areas of micromovement of the tear film in each image by means of the means for processing and analyzing images, and calculating a measure of the average amplitude of the areas of micromovement as a function of time based on the continuously calculated positions and numbers of areas of micromovement of the tear film.

[0021] The advantage of this measurement is that it provides an idea of the thickness variations of the tear film (film too thin or locally broken) throughout the measurement sequence.

[0022] The features described in the following paragraphs may be optionally implemented. The features may be implemented independently of each other or in combination with each other:

[0023] The method may comprise capturing (200) an image every 0.1 to 0.5 seconds and preferably every 0.3 seconds.

[0024] Said step of continuously capturing images advantageously ends with the first occurrence of one of the following events: expiration of the time delay or detection of the next eyelid blink.

[0025] The method may comprise tracking one or more eyes of the patient by means of an iris tracking method in order to relocate the area of micro-movements of the tear film detected and stored relative to the analyzed eye.

[0026] The estimated line edges may be calculated by polynomial regression.

[0027] The method may include determining a group of points representing the area of micro-movement observed in the image of the line by calculating the absolute values of distances to the polynomial along an axis perpendicular to the general direction of the line of pixels P1 to Pn at the edge of the image line separated by a distance dP in pixels greater than a threshold from the polynomial, and storing the distances dP1 to dPn for representative points P1 to Pn.

[0028] The method may include: for each image, for the representative point group, calculating the equivalent area Sp = dP × pixel width × pixel height for each pixel P1 to Pn, and calculating the sum ΣSp of the equivalent areas SP1 to SPn from the image point to its polynomial on the entire image and the normalized sum NΣSp of the distances obtained by dividing the sum ΣSp by the total length of the lines found in the image.

[0029] The method may comprise calculating an overall score consisting of the sum A of the normalized sums NΣSp over a series of images from the starting time t0 to a given time t.

[0030] The length of the measurement may be chosen depending on the patient population, appearing to be a length of about 6 seconds sufficient to detect dry eye problems while avoiding the risk of blinking.

[0031] The method may include storing the locations of the areas of micro-movement and generating a map of the locations of the micro-movements image by image.

[0032] The method may include calculating and storing an appearance time of the micro-movement point of the tear film and / or an appearance rate of the micro-movement point of the tear film.

[0033] The method may include storing all or some of the calculated data in a database for monitoring the patient and comparing the data over multiple examinations.

[0034] According to another aspect, a computer program is provided comprising instructions for carrying out all or some of the methods such as defined herein when this program is executed by a processor.

[0035] According to another aspect of the present invention, there is provided a computer-readable non-volatile storage medium having such a program stored thereon.

[0036] The present invention and its variations may generally allow for providing a method for measuring changes in irregularities of a tear film over a defined examination period that is more accurate than existing methods for detecting breakup of such a tear film.

[0037] Such a solution based on the measurement of the deformation of the membrane and its evolution over time allows solving the problems posed by the known solutions while achieving a high measurement repeatability. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Other features, details, and advantages of the present invention will become apparent upon reading the following detailed description and analyzing the accompanying drawings, in which:

[0039] [ Figure 1 ] is a plan front view of a test pattern having the pattern of the present invention;

[0040] [ Figure 2 ] is a schematic diagram of the device of this patent application;

[0041] [ Figure 3 ] shows a first example of a measuring device that can be used in the context of the present patent application;

[0042] [ Figure 4 ] showing an image of the patient's eye after a first treatment step;

[0043] [ Figure 5A ]、[ Figure 5B ]、[ Figure 5C ]、[ Figure 5D ]and[ Figure 5E ]Display processing Figure 4 The various stages of the eye's image;

[0044] [ Figure 6A ]and[ Figure 6B ]exhibit Figure 5E Details;

[0045] [ Figure 7 ] is shown after image analysis Figure 4 images of eyes;

[0046] [ Figure 8 ] shows an image of the patient's eyes as they look toward the camera;

[0047] [ Figure 9A ]、[ Figure 9B ]and[ Figure 9C ] shows that according to one aspect of the present patent application Figure 8 Steps for the position of the iris of the eye;

[0048] [ Figure 10 ] shows an image of the patient's eye as they look away from the camera;

[0049] [ Figure 11A ]、[ Figure 11B ]and[ Figure 11C ] shows that according to one aspect of the present patent application Figure 10 Steps for the position of the iris of the eye;

[0050] [ Figure 12] schematically illustrates a method for detecting an offset point according to one aspect of the present patent application;

[0051] [ Figure 13 ] schematically illustrates a method for repositioning deformation of image lines of a test chart;

[0052] [ Figure 14 ] schematically shows the first step of the method of the present patent application;

[0053] [ Figure 15 ] schematically shows the subsequent steps of the method of the present patent application;

[0054] [ Figure 16 ] A first graph showing smoothed curves for a plurality of patients;

[0055] [ Figure 17 ] A second graph showing a curve of cumulative values for multiple patients. DETAILED DESCRIPTION

[0056] The drawings and the following description describe one or more examples of embodiments, which may therefore serve not only to better understand the subject matter of the present patent application but also, if appropriate, contribute to its definition.

[0057] The method according to the present patent application comprises using Figure 1 The test shown in Figure 10 The invention relates to a measuring device for measuring blood pressure, wherein the test chart is produced using a transparent polymer film comprising a pattern 11 consisting of alternating straight and parallel lines 12, 13. The pattern comprises opaque lines 12, for example black lines, separated by transparent lines 13, which allow the passage of light from a light source, which passes through a translucent carrier behind the pattern in order to generate light rays, which are reflected from one or more eyes of the patient.

[0058] In this instance, the test Figure 10 The pattern 11 comprises twelve opaque lines 12, excluding the top and bottom borders of the test chart. These opaque lines are separated by transparent lines 13 and centered around a translucent center line. The test chart can be mounted on a plastic frame for easier handling.

[0059] By convention, the axis parallel to the axis through the patient's eye will be referred to as the horizontal axis, and the axis perpendicular to this axis will be referred to as the vertical axis; in the example shown, the lines of the pattern are horizontal.

[0060] like Figure 2 As shown in the test on the plastic carrier 10b Figure 10Positioned on a translucent carrier 10a, which itself has a hole drilled through it for the camera objective to pass through, the device illuminates the pattern using a diffuse light source 23 behind the pattern's carrier. The reflection of the pattern is observed by one or two digital cameras, and for observation of both eyes, two digital cameras 21, 22 are provided.

[0061] For example, the diffuse light source 23 may be produced by means of an integrating box or sphere or similar to an LCD backlight.

[0062] The test chart has two empty areas 14 centered on the horizontal midline with an opaque border 15. The empty areas are separated by a distance corresponding to the average eye distance, such as Figure 1 Return to Figure 2 The camera is positioned behind the empty area and in front of the eye 101 of the patient 100. The objective lens of the camera takes a picture or video of the patient's eye through the empty area. If the optical properties of the substrate of the test chart and its cleanliness are compatible with image formation (transparent and non-scattering), the empty area can be replaced by an area of the transparent substrate of the test chart.

[0063] The camera is, for example, a CMOS camera with a 1 / 4" sensor. According to the non-limiting example shown, the transparent area is a circular hole adapted to the diameter of the camera objective. For a camera with optics having a focal length of approximately 4 mm, a hole of approximately 14 mm is provided, and the opaque borders are opaque squares of approximately 16 mm x 16 mm. The goal of these borders is to terminate clearly upstream of the hole that receives the camera objective.

[0064] In this example of an embodiment, the camera or cameras deliver images with a resolution of 1920x1080, which is sufficient to obtain an analysis of the tear film without unduly increasing the computational load on the system.

[0065] The video signals or the signals of the cameras are sent to a computerized processing device 30 or computing system inside or outside the measuring device, and in order to avoid duplication of this processing device, the video signals of the two cameras are passed on the electronic board 24 of the device via a multiplexer that allows either of the video channels to be sent to the processing device 30 as needed.

[0066] A camera photographs or films an image of the patterned lines reflected from the patient's cornea. Because the cornea can be roughly considered a spherical refractive interface, it has a high curvature of field, and the image exhibits significant distortion. To at least partially compensate for this distortion and ensure that the line width in the captured image barely changes from the central axis of the pattern to the edge of the image, the test chart pattern includes lines whose period increases from the central axis of the pattern to the edge parallel to the line carrier. Depending on the desired resolution, the line width is approximately 2 to 4 mm. For example, the central white line may have a height of approximately 2.8 mm, the adjacent black lines 2.2 mm, the final white line 3.8 mm, and the preceding black lines 2.8 mm. This process is optimized to compensate for the curvature of a standard eye. In the described method, light rays are used to detect line width variations or deformations, but dark lines could also be used. The number and width of the rays may vary from the given example but are selected to achieve sufficient clarity to meaningfully detect areas of membrane deformation given the resolution of the camera or cameras.

[0067] In the context of this patent application, the white lines are reflected and appear by the corneal refractive interface, while they are backscattered from the conjunctiva and iris of the eye, in contrast to the dark lines and the backscattering of light that alternate to form a continuous background whose brightness depends on the ratio between the transparent area of the test chart and its total area.

[0068] exist Figure 2 In FIG, the test chart seen from above is bent about a vertical axis to form a portion of a cylinder and conforms to the curvature of the patient's head 100 so that the image of the test chart covers most of the cornea.

[0069] The reasons are:

[0070] a. Optical conjugation:

[0071] The camera sees the reflection of the test pattern from the cornea, which is roughly treated as a spherical mirror with a radius of about 8mm or a focal length of 4mm. Given this short focal length, the test pattern (the object in the optical conjugate) must be large enough so that the size of the image reflected by the cornea is large enough, that is, comparable to the outer diameter of the iris. This determines the size of the mask.

[0072] b. Photometry:

[0073] In order for the test chart to be visible, light rays from the extreme edges of the test chart must enter the pupil of the objective. Since the cornea is a mirror with a high curvature, it is necessary for the light rays reaching the cornea to be grazing at the edge of the field of view.

[0074] This justifies the curvature of the test chart.

[0075] exist Figure 3In the embodiment, the measuring device is mounted on an ophthalmic structure with a chin and forehead support, said structure comprising uprights 44, 46 and a housing 43 in which the camera is integrated and which receives the test piece on a forward curved surface. Figure 10 The patient is placed in front of the forward curved face with his lower jaw resting on the support 45. With this configuration, the distance between the patient's eyes and the test chart is approximately 50 mm for a camera with a focal length of 4 mm. The measurement device can also be integrated into a headset worn by the patient.

[0076] Allowing for variations in the position of the eye relative to the camera (different eye distances from one person to the next), the focus and distance are chosen to allow the entire eye to be seen. Next, a window of interest centered on the patient's pupil is selected via an action by the operator, who marks the center of the pupil in the image.

[0077] Thus, the lines of the test chart are sharp in the image and the focus is adjusted using the thumb wheel 42 .

[0078] The goal of the measurement method of the present patent application is to detect and measure the degree and area of instabilities in the tear film, which cause deformations in the reflection lines. The measurement method comprises repeatedly capturing images of one or both eyes of the patient at a repetition rate of approximately 0.1 to 0.5 seconds, and in practice 0.3 seconds, after one or more eyes have blinked.

[0079] The method is primarily described in the context of a test pattern consisting of horizontal lines, i.e., lines extending along an axis passing through the patient's two pupils, but can be adapted, in particular by means of means for processing a series of pixels and a 90° rotation of the anisotropic bandpass filter described below. Furthermore, the method described in the context of light detection is applicable to the detection of dark lines.

[0080] Using horizontal or vertical lines rather than slanted lines is advantageous because it is easier to perform image processing along rows or columns of pixels.

[0081] In the computing system 30, the measurement step includes an image processing step starting with capturing a raw image of the patient's eye, which includes:

[0082] - Convert the image to Figure 12 Step 205) into grayscale, Figure 4 An example of the result is shown in FIG, in which an image 51 of the pattern on the iris 50 is visible, wherein the image of the frame 52 surrounds the objective lens of the camera. In this image and in the original color image, the shape of the reflected light includes local defects (especially irregular line edges that alone suggest deformation of the tear film);

[0083] -application( Figure 12In step 210, a bandpass filter is applied that is anisotropic in a direction perpendicular to the direction of the reflected light. In the case of a test pattern consisting of horizontal lines or a test pattern consisting of vertical lines, this direction is the vertical direction, the direction of the pixel columns of the image, in which case the image is rotated 90° to obtain light directed in the horizontal direction. The filter is configured to pass abrupt transitions between grayscale levels and remove or attenuate modulation of low and higher spatial frequencies in the vertical direction. Figure 5A The output image from the filter is shown in . This filter has the advantage of removing dark corners and defects in the uniformity of the illumination. This transformation emphasizes the edge of the eyelid 53, the light 54 of the test chart pattern, and maintains the image of the border 55.

[0084] Next, also in the case of light directed in the horizontal direction, the method comprises analyzing the image in the 220 pixel columns, such as Figure 12 , to search for bright vertical segments. The resulting image then includes Figure 5B , such as the lines 56, 57, 58, 59 shown in FIG. Once this analysis has been performed, the bright segments are checked according to their size in steps 230 and 235. This allows for removing segments that are too large or too short and that clearly do not correspond to segments of a pattern line, as is the case, for example, if the segments form part of the eyelid contour. Figure 5C The image at the end of this step is shown, in which the column segments of the independent lines 61, the image lines of the pattern 62, the background 64, and the border 63 are preserved. It should be noted that the eyelashes have caused a large breakup of the line 65 at the top of the image. In the case of a test pattern consisting of vertical lines, the segments are analyzed and checked in pixel rows.

[0085] When the segmentation is finished, the processing method includes an algorithm for marking / classifying 240, 250, 260 the bright segments to obtain objects that represent bright lengths of pattern lines and discarding bright objects that do not have the desired shape and are therefore considered artifacts. This algorithm first joins adjacent column segments to reconstruct the horizontal length. The results of this marking / classification are shown in Figure 5D In which each line found has been assigned a color represented here in grayscale. This classification allows to create complete lines 70 or individual lengths 71, 72.

[0086] In a subsequent step, the lengths of bright lines of the same level (e.g., having similar width and altitude) in the image are joined, and then polynomial regression 285 using a polynomial of order greater than two is applied in order to calculate the RMS curve of the shape of the line edge. This step is shown in Figure 5E In this figure, it will be particularly noted that the lengths 73a, 73b of the lines segmented by the image of the frame surrounding the camera objective and the lengths 73c, 73d of the bottom line have been joined by the bottom 74 and top 74' RMS curves. Zoom in Figure 6AThis allows the RMS curves 74, 74' between the lengths 73c, 73c at the bottom of the image to be seen more clearly.Next, offset points 500 are detected in positions where the line edges include measurement points that deviate from the shape given by the polynomial.

[0087] In contrast to methods where only confirmed film ruptures are sought, in the context of the present invention, fluctuations in film thickness are sought, these causing micro-movements in the film.

[0088] As far as detection is concerned, the criterion considered here is simply a threshold criterion (eg, a threshold of one pixel) for filtering out noise in the image, taking into account all deviations between points and the polynomial that are larger than this threshold.

[0089] This is for example Figure 6B 7. In FIG. 7, it is shown in region 76 where line edge 77b of line 77a does not conform to curve 74'.

[0090] As seen above, the method can be based on the processing of dark lines. Grouping pairs of transitions (rising and falling transitions in the case of light) allows checking the consistency of the width of the obtained segments and discarding segments that are too wide or too narrow to form part of the image of the test chart. Once the length has been determined, a polynomial regression is performed on each side of the length: one polynomial for rising transitions and one for falling transitions. Thus, the method of the present invention can target dark segments and dark lines equally well, obtaining the same polynomial regression and the same final result.

[0091] The result of the measurement is a map of micro-movements in the tear film, a map of this type is shown in Figure 7 , which is a magnified view of the patient's eye 6 seconds after blinking, and in which the region 780 of smaller or zero film thickness is shown positioned on line 51 in the initial color image of the eye, which is shown here in grayscale.

[0092] Micro-movements correspond to gradual localized sagging or deformations of the membrane, for example in the form of depressions that can become as large as to destroy the membrane and cause deformations of the line because they locally modify the curvature of the refractive interface. When the eye is observed under magnification, these changes in deformation appear in the form of micro-movements or undulations.

[0093] As seen above, images are captured approximately every 0.3 seconds. The onset time is defined by the eyelid blink, and repeated measurements for each image over a period of time allow for the construction of a map of defects in the tear film over time.

[0094] One issue to consider is that the patient's gaze may change direction during the image acquisition cycle.

[0095] Because the camera observes the reflection of the pattern from the cornea, which behaves roughly like a spherical refractive interface, the position of the pattern's image remains largely unchanged in the image delivered by the camera, while the position of the iris changes if the patient turns their eye. Therefore, a given point in the image of the test chart is not connected to a fixed point on the cornea, but rather to a point that depends on the direction of gaze. This means that measurements must be referenced to the observed position of the eye, rather than to the image of the pattern.

[0096] To do this, the position of the eye in each image must be tracked. It is preferable to fit the outline of the iris of the eye because it forms a high contrast with the bulbar conjunctiva, which is lighter in color and does not reflect the test pattern from it. The pupil is easier to fit due to the reflection of the test pattern, which complicates the analysis of the image.

[0097] The following method can be implemented in the context of this patent application or independently to perform other measurements on the eye.In addition, this method is independent of the orientation of the lines of the test chart.

[0098] Figure 8 、 9A , 9B and 9C correspond to processing operations performed on an image of the patient's eye looking at the camera, and Figure 10 、 11A , 11B, and 11C correspond to processing operations performed on the image of the patient's eye whose gaze has shifted away from the camera.

[0099] exist Figure 8 In FIG. 8 , eye 80 is looking straight ahead, and the image of pattern 83 is centered relative to iris 82 , which is itself centered relative to eyelid 81 .

[0100] Figure 13 The image processing method for finding the position of the iris is schematically shown in . It comprises a first transformation of the image by applying an anisotropic bandpass filter 400, which is applied horizontally in order to detect brightness transitions along the horizontal axis. In this operation, it is desirable to distinguish between falling transitions (light to dark) and rising transitions (dark to light), and for the sake of clarity, this has been represented by Figure 9A The dark crescent 84 in the grayscale image in FIG. 8 represents a light to dark transition (a falling transition) and has been Figure 9A The light crescent 85 in the image represents the dark to light transition (rising transition). Positions of the image where a significant transition does not exist, such as the crescent 86, become average grayscale, and the outline of the iris is represented by the ring portion 87, which is located just next to the light to dark transition on the left hand side and just next to the dark to light transition on the other side of the eye.

[0101] Return to Figure 13The second operation consists in segmenting the image 410 in order to find pairs of rising and falling transitions (e.g., Figure 9A 84, 85). Figure 9B These transition pairs are indicated by boundaries 88, 89 and 90, 91, with a border region 92 potentially bounding the bulbar conjunctiva.

[0102] After this transformation, the method comprises Figure 13 The image is filtered 420 as shown in FIG, which removes the central area including the pattern and the top and bottom areas of the image. Based on the remaining parts, an RMS circle is calculated 430 for the periphery of the iris from the right end of the left-hand segment of the image and from the left end of the right-hand segment of the image. For this calculation, points that are too far from the RMS circle, corresponding to defects caused in particular by eyelashes or eyelids, are discarded in step 440, and for the remaining points, a new RMS circle is calculated to fit the contour of the iris, this circle 93 being Figure 9C is shown on the original image of the eye.

[0103] Figure 10 An eye 80' is shown looking to the side, with its iris 82' offset relative to pattern 83'. For this position of the eye, Figure 11A In the example, the transitions 84', 85' around the dark areas 86', 87' corresponding to uniform colors are laterally offset, but in Figure 11B The arcs 89' and 90' corresponding to the edges of the iris remain detectable. The application of the tracking method again makes it possible to regenerate the RMS circle 93', which will be Figure 11C . In step 460, the detected deformed area is then relocated depending on the position of the circle defining the outline of the iris. This makes it possible to anchor the deformation of the tear film to the outline of the eye rather than to the image.

[0104] This sequence is preferably performed for each image after the analysis of the line patterns described above.

[0105] As stated above, this method is applied here to deformation of the repositioning film, but it can also be used for other types of detection and methods that require tracking eye position.

[0106] According to one aspect of the present patent application, the device may include a manual trigger for providing the device, which then triggers the image capture sequence upon the occurrence of an event, such as a series of two blinks of the patient's eyelid. To this end, the system includes a method for recognizing an eyelid blink, which allows the measurement sequence to be automatically started. Similarly, the system may automatically stop the measurement sequence upon detecting a subsequent eyelid blink, or automatically stop the sequence after a time delay of, for example, 15 seconds.

[0107] For example, an image capture sequence may include 30 to 50 images, and in the case of an image capture sequence of 15 seconds in length with an image captured every 0.3 seconds, the sequence includes 45 images. The images can be analyzed after the image capture sequence, and due to the chosen solution, i.e. the option to work with patterns consisting of lines, the processing time remains low, for example 15 seconds using a standard computer.

[0108] exist Figure 14 In the method, the method comprises capturing 200 images continuously over a sequence lasting about fifteen seconds from the start time t0 of an eyelid blink, and during the start time, capturing images every 0.1 to 0.5 seconds and preferably every 0.3 seconds to obtain a good compromise between the detection of tear film changes and the amount of data to be processed.

[0109] The continuous image capture ends 570 with the first occurrence of one of the following events: the time delay ends, for example 15 seconds, resulting in 45 images if one image is captured every 0.3 seconds; or the next eyelid blink is detected.

[0110] Between image captures or in replay mode, ie not in real time, the method comprises continuously detecting 510 micro-movement areas in the image and by means of means for processing and analyzing the image, e.g. according to Figure 12 The method of and in particular using polynomial regression to continuously calculate and store 520 data on the position, area and number of micro-movement areas of the tear film in each image will make it possible to determine the estimated line edge by means of polynomial regression.

[0111] The method then includes calculating 530 and storing 540 a measure of the amplitude of the micro-movement area as a function of time. This measure provides a deep understanding of the thickness reduction of the film across the entire range of the imaging line.

[0112] To realign the lines between the images, the method tracks the patient's eyes. Figure 13 theme.

[0113] In the following steps, if Figure 15 As shown in , the method will allow determining a population of points representing areas of micro-movement observed in an image of a line.

[0114] This is done by calculating 610 the absolute value of the distance to the polynomial along an axis perpendicular to the general direction of the line of pixels P1 to Pn separated from the polynomial by a distance dP in pixels greater than a threshold 620 from the edge of the image line. These representative points P1 to Pn and the distances dP1 to dPn are then stored. In the case of a horizontal line, the distances are vertical, and in the case of a vertical line, the distances are horizontal.

[0115] In order to determine the extent of the area of membrane deformation, the method comprises, for each image, calculating 640 the equivalent area of deformation in the line edge for each pixel P1 to Pn, ie Sp = dP x pixel width x pixel height.

[0116] Then, the following are calculated 650: the sum ΣSp of the equivalent areas SP1 to SPn between image points and their polynomials over the entire image, and the normalized sum NΣSp of the distances obtained by dividing the sum ΣSp by the total length LT of the lines found in the image.

[0117] This number represents the total extent of the deformation in the area around the edge of the line.

[0118] In order to obtain a value that is reproducible from one examination to the next, the method comprises calculating 660 an overall score consisting of the sum A of the normalized sums NΣSp over a series of images from a starting time t0 to a given time t.

[0119] exist Figure 16 and 17 In the example of the curve in , it can be seen that summing over a period of approximately 6 seconds, ie 18 images when an image is taken every 0.3 seconds, allows to discern differences between patients.

[0120] The method may further include generating 550 a map of the locations of the micromovements on an image-by-image basis, which allows identification of more marked areas of instability in the tear film and observation of changes in deformation of the film over time.

[0121] It is also possible to calculate and store the appearance time of the micro-movement points of the tear film and / or the appearance rate 560 of the micro-movement points of the tear film, which may help to compare the results during treatment.

[0122] Furthermore, all or some of the calculated data may be stored in order to monitor the patient and compare the data over multiple exams.

[0123] Once the measurement is complete, the physician will have, on the one hand, a spatial and temporal map of the tear film micromovements on the corneal surface, and, on the other hand, a time-dependent curve that tracks the changes in these micromovements of the tear film over time. The amplitude at each point of the curve will reveal the level of stability at a given time. The slope of this time-dependent curve will reveal the incidence of breakups in the tear film. This allows for optimized interpretation of the performed examination.

[0124] Figure 16 A graph 700 is shown where the y-axis 701 represents the normalized sum NΣSp filtered with a moving average of three images, and the x-axis 702 represents images of five patients C1 to C5 .

[0125] Patients C1 and C2 exhibited minimal micromovements, ie, low instability of their tear film over time, patient C3 exhibited moderate instability but did not keep their eyes open throughout the sequence, and patients C4 and C5 exhibited high instability indicative of dry eye.

[0126] Figure 17 Graph 710 is shown, where the y-axis 711 represents the cumulative value A of consecutive images and the x-axis 702 represents the image. This graph magnifies the visibility of the development of micro-movement areas in patients C4 and C5 with dry eye, while patients C1 and C2 showed minimal progression and patient C3 showed moderate progression until blinking at image 28.

[0127] In both graphs, a time region 703, 713 of approximately two seconds centered around image 18, i.e., in the present case, 6 seconds after a blink, is shown. As can be seen in the graphs, this examination time of approximately 6 seconds alone is sufficient to standardize the results and distinguish between patients without dry eye problems and those who do suffer from it, without the risk of an examination time so long that an untimely blink may occur.

[0128] The present invention is not limited to the examples described above, which are merely examples, but encompasses any variations within the scope of the claimed invention, such as any other distribution or variation of the line heights that can be envisioned by a person skilled in the art. Specifically, as stated above, the pattern lines, which are parallel horizontal lines in the example shown, can be replaced by parallel vertical lines. Rotation of the image, for example, allows image processing means for detecting deformation of the lines to be applied to this configuration without changing their orientation.

Claims

1. A method for measuring the stability of a patient's tear film by means of a device, the device comprising: a backlit translucent plate provided with a test pattern positioned in front of at least one eye of the patient; and at least one digital camera connected to a computing system having means for processing and analyzing images, the objective lens of the camera being directed toward the patient's eye in order to capture the reflection of the pattern of the test chart from the patient's eye, characterized in that the test chart has a pattern consisting of a series of alternating light and dark lines reflected from the patient's eye, and the means for processing and analyzing images being configured to detect deformations of the light or dark lines of the pattern of the test chart reflected from the patient's eye and to identify tear film micromovements revealed by these deformations by comparing the positions of image points on the edges of the lines with an estimated line edge of the lines, the method comprising, starting from the start time t0 of an eyelid blink: - continuously capture images, - continuously detecting areas of micro-movement in the images, and continuously calculating and storing data on the position, area and number of areas of micro-movement of the tear film in each image by means of means for processing and analyzing images, and - Based on the continuously calculated position, area and number of areas of micro-movement of the tear film, calculating and storing a measure of the width of said areas of micro-movement as a function of time. The method of claim 1 , comprising capturing an image every 0.1 to 0.5 seconds.

3. The method according to claim 1, characterized in that The step of continuously capturing images ends with the first occurrence of one of the following events: the time delay ends or the next eyelid blink is detected.

4. The method according to claim 1, comprising tracking one or more eyes of the patient by means of an iris tracking method in order to relocate the deformed areas of the tear film detected and stored relative to the analyzed eye.

5. The method according to claim 1, wherein The estimated line edges are calculated by polynomial regression.

6. The method according to claim 5, comprising: - determining, along said line, a group of points representing an area of micro-movement observed in the image of said line by calculating the absolute value of the distances between pixels P1 to Pn of the edge of said line of said image along an axis perpendicular to the general direction of said line, said pixels being separated from said polynomial by a distance dP, said distance dP being greater than a threshold value for the number of pixels, and - Storing the distances dP1 to dPn for the representative points P1 to Pn.

7. The method according to claim 6 , comprising calculating, for each image, for the group of points, an equivalent area Sp = dP×pixel width×pixel height for each pixel P1 to Pn, and calculating a sum ΣSp of equivalent areas SP1 to SPn from the image point to its polynomial over the entire image and a normalized sum NΣSp of the distances obtained by dividing the sum ΣSp by the total length LT of the lines found in the image. 8 . The method according to claim 7 , comprising calculating an overall score consisting of the sum A of the normalized sums NΣSp within a series of images from the starting time t0 to a given time t.

9. The method of claim 1, comprising generating a map of the locations of micro-movements on an image-by-image basis. 10 . The method according to claim 1 , comprising calculating and storing an appearance time of a micro-movement point of a tear film and / or an appearance rate of the micro-movement point of the tear film.

11. The method of claim 1 , comprising storing all or some of the calculated data in a database for monitoring the patient and comparing the data over multiple examinations.

12. A computer-readable storage medium storing a computer program comprising instructions for implementing the method according to claim 1 when the program is executed by a processor.

13. A computer-readable non-volatile storage medium having stored thereon a program for implementing the method according to claim 1 when the program is executed by a processor.

Images