METHOD FOR ASSESSING THE STABILITY OF A TEAR FILM

Description

Domaine technique

[0001] The invention relates to the field of detecting dry eye syndrome. Technique antérieure

[0002] Tests for assessing dry eye syndrome traditionally use a fluorescence test. The practitioner instills a drop of fluorescein into the patient's eye and the fluorescence phenomenon is observed using a slit lamp. The practitioner determines the time between blinking and the appearance of the first non-fluorescent area. This time is the measurement of the ocular film breakup time or BUT. This method leads to measuring a first breakup, whether due to a local defect such as dust or a real dry eye problem. The measured times are in the order of 6 to 10 seconds.

[0003] Another method of assessment, called "Non-Invasive Breakup Time" or NIBUT, is a non-invasive process for measuring breakup time. The tests of this method use, in particular, the reflection on the cornea of ​​a target provided with a pattern consisting of concentric circles and radial lines.

[0004] When a film break occurs, the reflected pattern shows distortions of lines and / or circles or missing areas.

[0005] The images of the reflected pattern are acquired in sequence by a camera and processed by a computerized processing system which analyzes the evolution of the pattern over time to detect deformations of the pattern and deduce the time of appearance of the film rupture zones.

[0006] WO 2018 / 156022 discloses a method for measuring the condition of the tear film using a device comprising: a multi-color stimulator 100 comprising a hemispherical surface 120 on which are mounted LEDs capable of projecting light rays onto the cornea in the form of a rectangular grid, a camera for recording an image of the illumination of the cornea, a computer for processing the recorded images and determining a local distortion of each image point and the interval between blinking of the eye and rupture of the tear film. Problème technique

[0007] Currently available devices using circular and radial line patterns cause problems of several kinds: given the shape of the targets, and the coverage of the cornea by disc sectors, the spatial resolution is intrinsically less good at the edges of the field than in the center.

[0008] Furthermore, in the center of the field, where the radial lines converge, it is difficult to determine the deformation of a pattern.

[0009] Increasing the spatial resolution would require tightening the pattern of the target. However, circle deformation detection algorithms that require finding the most probable circles in the image, evaluating whether a point in the image belongs to a probable circle and calculating the distance of a given point to the circle to which it is supposed to belong are complex and the processing and computation time to detect breakage zones from the image is significant.

[0010] It is therefore clear that: on the one hand, devices presenting the classic pattern of concentric circles and radial lines suffer from a lack of homogeneity of the spatial resolution, on the other hand, increasing the resolution at the edge of the field is penalizing on the calculation times.

[0011] Due to the low resolution of this process, the measurement may give inaccurate and overestimated values ​​of the occurrence of ruptures.

[0012] It should be noted in particular that measurements obtained by the fluorescence method and the non-invasive method are generally not well correlated: a factor of 2 between BUT and NIBUT measurements is commonly observed, with the fluorescence measurement showing a shorter time. This suggests a lack of sensitivity or resolution of the NIBUT devices.

[0013] Likewise, the times displayed in NIBUT on different available devices are not well correlated either.

[0014] In the diagnosis of dry eye syndrome, measurements obtained by fluorescein break-up time or measurements obtained by a NIBUT method do not correlate with the suffering expressed by the patient.

[0015] The previous methods do not reveal the overall state of the tear film, which is detrimental to the practitioner who needs more precision and repeatability of measurements to make his diagnosis. Exposé de l'invention

[0016] The invention improves the situation and proposes a method for measuring the stability of a tear film on a patient's cornea, based on the detection of modifications to this film and using a device comprising a pattern made up of horizontal or vertical light and dark lines, and observing the reflection of this pattern on the cornea.

[0017] To do this, in the context of this application we will look for micro-movements at the level of the tear film by analyzing the images of the lines of sight.

[0018] Micro-movement is a local variation in the overall thickness of the tear film. This variation induces slopes on the surface of the tear film, which will be materialized by a localized modification of the image of the lines of the target and in particular modifications of the width of the image of these lines or local deformations.

[0019] More specifically, the present application proposes a method for measuring the stability of a patient's tear film by means of a device comprising a backlit translucent plate provided with a target positioned in front of at least one eye of a patient, at least one digital photographic camera connected to a computing system provided with image processing and analysis means, a lens of the camera pointing towards the patient's eye in order to photograph a reflection of the pattern of the target on the patient's eye, characterized in that, the target being provided with a pattern consisting of a succession of alternating light and dark lines reflected on the patient's eye, the image processing and analysis means are configured to detect deformations of said light or dark lines of the pattern of the target reflected on the patient's eye and identify, by comparing the position of image points of the edge of the lines with respect to an estimated edge of the lines,micro-movements of the tear film revealed by these deformations, said method comprising, from a time origin t0 constituted by a blink of the eyelid, a succession of image captures, a succession of detections of micro-movement zones in the images and a succession of calculations and storage of data on positions and number of micro-movement zones of the tear film in each image by means of the image processing and analysis means and the calculation, from the succession of calculations of positions and number of micro-movement zones of the tear film, of a measurement of the average amplitude of the micro-movement zones as a function of time.,

[0020] The measurement carried out has the advantage of giving an image of the variations in thickness of the ocular film, film too thin or broken locally throughout a measurement sequence.

[0021] The features set out in the following paragraphs may optionally be implemented. They may be implemented independently of one another or in combination with one another: The method may comprise taking an image (200) every 0.1 to 0.5 seconds and preferably every 0.3 seconds.

[0022] Said succession of image captures advantageously ends with the first occurrence of one of the events: end of a time delay or detection of a following eyelid blink.

[0023] The method may include tracking the patient's eye or eyes using an iris tracking method to reposition the detected and stored tear film micro-movement areas relative to the eye being analyzed.

[0024] The estimated line edges can be calculated by polynomial regression.

[0025] The method may comprise the determination of a population of points representative of the micro-movement zones observed on the image of the lines by calculating, along the lines, the absolute value of the distance to the polynomial along the axis perpendicular to the general direction of the lines of the pixels P1 to Pn of the edges of the lines of the image which are separated from the polynomial by a distance dP in pixels greater than a threshold and storing for the representative points P1 to Pn the distances dP1 to dPn.

[0026] The method may comprise for each image, on said population of representative points, a calculation for each pixel P1 to Pn of the equivalent surface Sp = dP x pixel width x pixel height and a calculation of the sum ΣSp of the equivalent surfaces SP1 to SPn of the image points to their polynomial over the entire image and a calculation of a normalized sum NΣSp of said distances by means of the division of the sum ΣSp by the total length of the lines found on the image.

[0027] The method may include the calculation of an overall score consisting of the sum A of the normalized sums NΣSp over a series of images from the origin of times t0 to a given instant t.

[0028] The duration of the measurements can be chosen according to a patient population, a duration of around 6 seconds appearing sufficient to detect dry eye problems while avoiding reaching a risk of blinking.

[0029] The method may include storing the positions of the micro-movement zones and generating a map of the locations of the micro-movements image by image.

[0030] The method may comprise calculating and storing the time of appearance of the micro-movement points of the tear film and / or the kinetics of appearance of micro-movement points of the tear film.

[0031] The method may include storing in a database all or part of the calculated data in order to monitor the patient and compare said data over several examinations.

[0032] According to another aspect, there is provided a computer program comprising instructions for implementing all or part of a method as defined herein when this program is executed by a processor.

[0033] According to another aspect of the invention, there is provided a non-transitory, computer-readable recording medium on which such a program is recorded.

[0034] The invention and the variants thereof can make it possible, in general, to propose a method for measuring the evolution of irregularities in a tear film over a defined examination period offering increased precision compared to existing methods for detecting breaks in such a tear film.

[0035] Such a solution, which is based on a measurement of the deformations of the film and their evolution over time, makes it possible to resolve the problems posed by known solutions by offering high repeatability of measurements. Brève description des dessins

[0036] Other characteristics, details and advantages of the invention will appear on reading the detailed description below, and on analyzing the attached drawings, in which: [ Fig. 1 ] is a flat front view of a target provided with a pattern of the invention; [ Fig. 2 ] is a schematic view of a device of the present application; [ Fig. 3 ] shows a first example of measurement support usable within the framework of the present application; [ Fig. 4 ] represents an image of a patient's eye following a first stage of treatment; [ Fig. 5A ], [ Fig. 5B ], [ Fig. 5C ], [ Fig. 5D], [Fig. 5E ] represent various stages of processing the image of the eye of the figure 4 ; [ Fig. 6A], [Fig. 6B ] represent details of the figure 5E ; [ Fig. 7 ] represents an image of the eye of the figure 4 after image analysis; [ Fig. 8 ] represents an image of a patient's eye looking towards a camera; [ Fig. 9A], [Fig. 9B ], [ Fig. 9C ] represent steps in determining the position of the iris of the eye of the figure 8 according to one aspect of the request; [ Fig. 10 ] represents an image of a patient's eye looking away from a camera; [ Fig. 11A], [Fig. 11B ], [ Fig. 11C ] represent steps in determining the position of the iris of the eye of the figure 10 according to one aspect of the request; [ Fig. 12 ] schematizes a method for detecting offset points according to an aspect of the request; [ Fig. 13 ] schematizes a process for repositioning deformations of image lines of a test pattern; [ Fig. 14 ] Diagram of the first steps of the application process; [ Fig. 15 ] schematizes subsequent steps in the application process; [ Fig. 16 ] represents a first graph with smoothed value curves for several patients; [ Fig. 17 ] represents a second graph with cumulative value curves for several patients. Description des modes de réalisation

[0037] The drawings and description below describe one or more examples of embodiment which may therefore not only serve to better understand the objects of this application, but also contribute to its definition, where appropriate.

[0038] The method according to the present application uses a measuring device comprising a target 10 shown in figure 1 and made from a transparent polymer film comprising a pattern 11 made with alternating straight and parallel lines 12, 13. The pattern comprises opaque lines 12, for example black lines separated by transparent lines 13 allowing light from a light source to pass through a translucent support behind the pattern to produce clear lines so that the clear lines of the pattern are reflected on the eye or eyes of a patient.

[0039] According to the example, the pattern 11 of the target 10 comprises twelve opaque lines 12 not including the upper and lower borders of the target. These opaque lines are separated by transparent lines 13 and centered around a central translucent line. The target can be mounted on a plastic frame to make it easier to handle.

[0040] By convention, we will call a horizontal axis an axis parallel to an axis passing through the patient's eyes and a vertical axis the axis perpendicular to this axis and in the example illustrated, the lines of the pattern are horizontal.

[0041] As represented in figure 2 , the target 10 on its plastic support 10b is positioned on a translucent support 10a itself pierced with holes for the camera lenses to pass through and the device uses a diffuse light source 23 behind the support of the pattern to illuminate the pattern. The reflection of the pattern is observed by one or two digital cameras and, to observe the two eyes, two digital cameras 21, 22 are provided.

[0042] The diffuse light source 23 can be produced by means of an integrating box or sphere or in a manner similar to an LCD screen backlight for example.

[0043] The target is provided with two recessed areas 14 centered in opaque frames 15 on a horizontal center line. The recessed areas are separated by a distance corresponding to an average eyepiece distance as shown in figure 1 . Back to the figure 2 , the cameras are positioned behind the recessed areas and are in front of the eyes 101 of the patient 100. The camera lenses film or photograph the patient's eyes through the recessed areas. The recessed areas can be replaced by transparent substrate areas of the target if the optical properties of the target substrate and its level of cleanliness are compatible with image formation (transparent and non-diffusing).

[0044] The cameras are for example CMOS type cameras with 1 / 4" sensor. According to the non-limiting example shown, the transparent areas are circular holes with a diameter adapted to the camera lenses. For cameras with optics with a focal length of around 4mm, holes of around 14mm are made and the opaque frames are opaque squares of around 16mm x 16mm. These frames are intended to precisely terminate the lines upstream of the holes receiving the camera lenses.

[0045] In the exemplary embodiment, the camera(s) provide images with a definition of 1920x1080 which is sufficient for eye film analysis without increasing the computational load of the system.

[0046] The video or camera signals are sent to a computerized processing device 30 or calculation system, internal or external to the measuring device and, to avoid duplicating this processing device, the video signals from the two cameras pass through a multiplexer, on an electronic card 24 of the device, the multiplexer making it possible to send one or other of the video channels as desired to the processing device 30.

[0047] The cameras film or photograph the image of the pattern lines reflected in the patient's cornea. Since the cornea at first approach can be considered as a spherical diopter, it has a significant curvature of field and the image has a significant distortion. In order to at least partially compensate for the distortion and to preserve on the captured images lines whose width varies little from the median axis of the pattern towards the edge of the image, the pattern of the target includes lines with an increasing period from the central axis of the pattern towards the edges of the support parallel to the lines. The width of the lines is of the order of 2 to 4 mm depending on the desired resolution.For example, a central white line may have a height of approximately 2.8 mm, the adjacent black lines a height of 2.2 mm while the last white lines have a height of 3.8 mm and the black lines preceding these white lines have a height of 2.8 mm, the progression being optimized to compensate for the curvature of a standard eye. In the method described, it is the light lines that are used to search for variations in width or deformation of the lines but the dark lines can also be used. The number and width of the light lines may be different from the examples given but is chosen to obtain sufficient definition for significant detection of film deformation zones depending on the resolution of the camera(s).

[0048] In the principle of the present application, the white lines are reflected by the corneal diopter and emerge when they are backscattered in the bulbar conjunctiva of the eye and the iris where the alternation of dark lines and light lines rather creates a more or less luminous continuous background depending on the ratio between the transparent surfaces of the target and its total surface.

[0049] In figure 2 , the target seen from above is curved around a vertical axis to form a portion of a cylinder and follow the curvature of the patient's head 100 so that the image of the target covers a large portion of the cornea.

[0050] The reasons are: a. Optical conjugation: The camera sees the reflection of the target by the cornea considered at first glance as a spherical mirror of radius approximately 8 mm, or 4 mm focal length. Given this short focal length, the target (the object in the optical conjugation) must be large so that the size of the image by the cornea is sufficiently large, comparable in size to the external diameter of the iris. This is what determines the size of the mask. b. Photometry: For the target to be visible, light rays from the extreme edges of the target must enter the pupil of the lens. The cornea being a mirror of strong curvature, it is necessary that at the edge of the field the light rays arriving on the cornea are grazing.

[0051] This is what justifies the curvature of the sight.

[0052] According to the figure 3 the measuring device is mounted on an ophthalmological frame with chin and forehead support comprising uprights 44, 46, a casing 43 which integrates the cameras and which receives the target 10 on a curved frontal face in front of which the patient will be placed, chin resting on a support 45. On such a frame, the patient's eyes are at a distance from the target of approximately 50 mm for a camera focal length of 4 mm. The measuring device can also be integrated into a helmet worn by the patient.

[0053] The focal length and distance are chosen to allow the entire eye to be seen, taking into account variations in the positioning of the eye in relation to the camera (interocular distance differs from one individual to another). Then a window of interest is chosen, centered on the patient's pupil, by the action of the operator who points the center of the pupil on the image.

[0054] To ensure that the lines of the sight are sharp on the image, the focus is adjusted with a wheel 42.

[0055] The purpose of the measuring method of the present application is to detect and measure the extents and zones of instability of the tear film which result in deformations of the reflected lines. It comprises a repetition of shots of one or both eyes of the patient according to a repetition frequency of the order of 0.1 to 0.5 seconds and in practice 0.3 seconds after a blink of the eyes or the eye.

[0056] The method is described mainly in the context of a horizontal line pattern, that is to say along an axis passing through the patient's two pupils, but is adaptable in particular by means of a 90° rotation of the pixel series processing means and the anisotropic bandpass filter described below. Furthermore, the method described in the context of the detection of light lines can be applied to the detection of dark lines.

[0057] Using horizontal or vertical lines rather than slanted ones is attractive because image processing along rows or columns of pixels is easiest to perform.

[0058] The measurement steps in the computing system 30 include image processing steps which, starting from a capture of the original image of the patient's eye, include; a conversion of the image into grayscale according to step 205 of the figure 12 and an example of the result is represented in figure 4 where we distinguish the image 51 of the pattern on the iris 50 with the image of the frame 52 surrounding the camera lens. In this image and the original color image, the shape of the reflected light lines has local defects (in particular irregular line edges which already reveal deformations of the tear film); an application of an anisotropic bandpass filter, step 210 of the figure 12 , in a direction perpendicular to the direction of the reflected lines. This direction is a vertical direction along the pixel columns of the image in the case of a horizontal line pattern or a vertical line pattern with the image rotated 90° to obtain clear lines oriented in a horizontal direction. The filter is configured to keep the transitions between the gray levels sharp and to suppress or attenuate the modulations of low spatial frequency and higher spatial frequency in the perpendicular direction. The filter output image is represented at figure 5A . This filter is particularly useful for overcoming vignetting phenomena and lighting homogeneity defects. This transformation accentuates the lines of the eyelid 53, the light lines 54 of the pattern of the test chart and preserves the image of the frame 55.

[0059] Then, still in the case of clear lines oriented in a horizontal direction, the method includes an analysis of the image 220 by columns of pixels according to the figure 12 which will search for vertical light segments. The resulting image then has lines 56, 57, 58, 59 as shown in figure 5B . Once this analysis has been carried out, the light segments are qualified by their size, steps 230, 235. This makes it possible to eliminate segments that are too large or too short and that clearly do not correspond to line segments of the pattern, for example segments forming part of the outline of the eyelids. At the end of this step, the figure 5C represents the image where the column segments constituting the isolated line 61, the lines of the pattern image 62, the background 64 and the frame 63 remain. It should be noted that the eyelashes cause significant fragmentation of the lines 65 at the top of the image. In the case of a vertical line test chart, the analysis and qualification of the segments is done by lines of pixels.

[0060] When the segmentation is complete, the processing method comprises a marking / cataloging algorithm 240, 250, 260 of the light segments to obtain the objects representative of light branches of lines of the pattern and to reject the light objects not having the desired shape which are then considered as artifacts. This algorithm firstly comprises a connection of the contiguous column segments to reconstitute horizontal branches. The result of this marking / cataloging is represented in figure 5D for which each line found is assigned a color here represented in grayscale. This cataloging makes it possible to create complete lines 70 or isolated branches 71, 72.

[0061] A next step is a connection 280 of the branches of light lines of the same level (for example of similar width and altitude) in the image and then a polynomial regression 285 using a polynomial of order higher than two in order to calculate an RMS curve of the shape of the line edges. This step is represented in figure 5E . In this figure we notice in particular connections by the lower RMS curves 74 and upper 74' of the branches 73a, 73b of the lines broken up at the level of the image of the frame surrounding the lens of the camera and of the branches 73c, 73d of the lower line. The enlargement of the figure 6A allows to better distinguish the RMS curves 74, 74' between the branches 73c, 73c at the bottom of the image. Then a detection of the offset points 500 is carried out at the locations where the edges of the lines have measurement points which deviate from the shape given by the polynomial.

[0062] Unlike a method of simply searching for proven film ruptures, in the context of the present invention, fluctuations in film thickness are searched for which result in micro-movements in the film.

[0063] The criterion taken into account for detection is here only a threshold criterion (for example a threshold of one pixel) to filter the noise in the image, all the deviations between the points and the polynomial greater than this threshold being taken into account.

[0064] This is for example represented in figure 6B in zone 76 where the edge of line 77b of line 77a does not reach curve 74'.

[0065] As seen above, the method can be based on a treatment of dark lines. Grouping the transition pairs (rising and falling in the case of light lines) allows a consistency check to be made on the width of the segment obtained and to reject segments that are too wide or too narrow to be part of the image of the test pattern. Once the branches have been determined, the polynomial regression is carried out on each side of the branch: a polynomial for the rising transitions, a polynomial for the falling transitions. The method of the invention therefore makes it possible to search for dark segments and dark lines and arrive at the same polynomial regressions and the same final result.

[0066] The result of the measurements is a map of micro-movements in the tear film represented in the figure 7 view of a patient's eye 6 seconds after blinking with magnification for which the mapping of the areas of reduced or zero film thickness 780 appears positioned on lines 51 in the initial color image of the eye reproduced here in grayscale.

[0067] Micro-movements correspond to evolving local deformations or sagging of the film, for example in the form of bowls which can go as far as breaking the film which result in deformations of the lines by local modification of the curvature of the diopter. The evolution of these deformations when observing the eye with magnification appears in the form of micro-movements or undulations.

[0068] As seen above, images are captured approximately every 0.3 seconds. A time origin is defined as an eyelid blink and repeating the measurement for each image over a period of time allows the construction of a map of ocular film defects as a function of time.

[0069] One issue to consider is that the patient's gaze may change direction during the image acquisition period.

[0070] As the camera observes the reflections of the pattern on the cornea which behaves in first approximation like a spherical diopter, the position of the pattern image remains almost invariant in the image delivered by the camera while the position of the iris changes if the patient turns the eyes. Therefore a given point of the image of the target is not linked to a fixed point of the cornea, but on the contrary to a point dependent on the direction of gaze. This implies that the measurement must be referenced in relation to the position of the observed eye and not in relation to the image of the pattern.

[0071] To do this, it is necessary to follow the position of the eye in each image. It was preferred to follow the outer contour of the iris of the eye because there is a significant contrast with the bulbar conjunctiva, which is light in color and on which there is no reflection of the target, while the pupil is more difficult to follow because of the reflection of the target, which complicates the analysis of the image.

[0072] The following method can be implemented within the scope of the present application or independently for other eye measurements. This method is also independent of the orientation of the sight lines.

[0073] THE figures 8 , 9A , 9B And 9C correspond to the image processing on a patient's eye looking at the camera and the figures 10 , 11A , 11B And 11C correspond to image processing on a patient's eye whose gaze moves away from the camera.

[0074] According to the figure 8 , the eye 80 looks straight ahead, the image of the pattern 83 is centered relative to the iris 82 itself centered relative to the eyelid 81.

[0075] The image processing process to find the position of the iris is shown diagrammatically in figure 13 . It involves a first transformation of the image by applying an anisotropic bandpass filter 400 applied horizontally to detect light transitions along a horizontal axis. In this operation we seek to distinguish the descending transitions (light to dark) and the ascending transitions (dark to light) and for understanding, the light to dark transitions (descending transition) are translated by the dark crescent 84 of the grayscale representation of the figure 9A and the dark to light transitions (rising transition) are translated by the light crescent 85 of the figure 9A . Parts of the image without significant transitions become medium gray, such as the crescent 86 for example, and the outline of the iris is materialized by a portion of a ring 87 which is found on the left side next to a light to dark transition and on the other side of the eye next to a dark to light transition.

[0076] Back to the figure 13 , a second operation consists of a segmentation of the image 410 to search for the pairs of rising and falling transitions, 84, 85 in figure 9A for example, which form boundaries of the bright areas in the image, notably around the bulbar conjunctiva. These pairs of transitions materialized by the boundaries 88, 89 and 90, 91 on the figure 9B frame 92 areas potentially defining the bulbar conjunctiva.

[0077] From this transformation the process comprises according to the figure 13 a filtering of the image 420 which removes the central area containing the pattern and the upper and lower areas of the image. On the remaining parts, a calculation of an RMS circle of the circumference of the iris is carried out from the right ends of the segments to the left of the image and from the left ends of the segments to the right of the image 430. For this calculation, the points too far from the RMS circle which correspond to imperfections notably caused by the eyelashes or the eyelid are rejected at step 440 and, for the remaining points, a new RMS circle is calculated to follow the contour of the iris, this circle 93 is represented in figure 9C on the original image of the eye.

[0078] There figure 10 represents an eye 80' looking sideways and whose iris 82' is offset relative to the pattern 83'. For this position of the eye, the transitions 84', 85' around dark areas 86', 87' corresponding to homogeneous colors are offset laterally in the figure 11A while in the figure 11B we note that the arcs of circles 89' and 90' corresponding to the edge of the iris remain detectable. The application of the tracking process also leads to the recreation of the RMS circle 93' which will be repositioned on the original image as represented in figure 11C The detected deformation zones are then relocated according to step 460 according to the position of the circle defining the outline of the iris. This makes it possible to anchor the deformations of the tear film to the outline of the eye rather than to the image.

[0079] This sequence is performed for each image preferably after the analysis of the line pattern described above.

[0080] As mentioned above, this process is applied here to the repositioning of film deformations, but it can also be used for other types of detection and methods using eye position tracking.

[0081] According to one aspect of the application, the device may include a manual trigger that arms the device, the triggering of the sequence of images then being carried out on the occurrence of an event such as a series of two blinks of the patient's eyelids. To do this, the system includes a blink recognition method that makes it possible to start the measurement sequence automatically. Similarly, the system can stop the measurement sequence automatically upon detection of a subsequent blink of the eyelids or stop the sequence automatically after a time delay, for example 15 seconds.

[0082] The shooting sequence can consist of 30 to 50 images for example and in the case of a 15 second shooting sequence with shots taken every 0.3 seconds the sequence consists of 45 images. The analysis of the images can be done after the shooting sequence and due to the chosen solution of working on a pattern made of lines, the processing time remains reduced, for example 15 seconds with a standard computer.

[0083] According to the figure 14 , the method comprises, from a time origin t0 constituted by a blink of the eyelid, a succession of images 200 taken over a sequence which will last around fifteen seconds and taking images every 0.1 to 0.5 seconds and preferably every 0.3 seconds to have a good compromise between the detection of variations in the ocular film and the quantity of data to be processed.

[0084] The sequence of image captures ends with the first occurrence of one of the events: end of a time delay, for example 15 seconds which gives 45 images with one image every 0.3 seconds or detection of an eyelid blink following 570.

[0085] Between the image captures or delayed, the method comprises a succession of detections of micro-movement zones in the images 510 and calculations and storage of data on positions, surface area and number of micro-movement zones 520 of the tear film in each image by means of image processing and analysis means according to, for example, the method of figure 12 and in particular polynomial regression which will make it possible to determine line edges estimated by means of a polynomial regression polynomial.

[0086] The method then comprises calculating 530 and storing 540 a measurement of the amplitude of the micro-movement zones as a function of time. This measurement provides an overview of the reductions in film thickness across the entire extent of the lines viewed.

[0087] To realign the positions of the lines between the images, the process will involve tracking the eye of the patient who is the subject of the figure 13 .

[0088] In a next step according to the figure 15 , the process will make it possible to determine a population of points representative of the micro-movement zones observed on the image of the lines.

[0089] This is done by a calculation 610, along the lines, of the absolute value of the distance to the polynomial along the axis perpendicular to the general direction of the lines of the pixels P1 to Pn of the edges of lines of the image which are separated from the polynomial by a distance dP in pixels greater than a threshold 620. These representative points P1 to Pn and the distances dP1 to dPn are then stored. In the case of horizontal lines, the distance is vertical and in the case of vertical lines, the distance is horizontal.

[0090] To determine an extent of areas where the film is deformed, the method comprises for each image, a calculation 640 for each pixel P1 to Pn of the equivalent surface of the deformation at the level of the line edge Sp = dP x pixel width x pixel height.

[0091] Then a calculation 650 of the sum ΣSp of the equivalent surfaces SP1 to SPn between the image points and their polynomial over the entire image and of a normalized sum NΣSp of the said distances by means of the division of the sum ΣSp by the total length LT of the lines found on the image.

[0092] This number represents an overall extent of deformations in the line edge areas.

[0093] To obtain a reproducible value between examinations, the method comprises the calculation 660 of an overall score consisting of the sum A of the normalized sums NΣSp on a series of images from the origin of times t0 to a given instant t.

[0094] According to the examples of the curves of the figures 16 And 17 we note that a sum over a duration of around 6 seconds, i.e. 18 images when the images are taken every 0.3 seconds, makes it possible to distinguish the differences between patients.

[0095] The method may further comprise the generation 550 of a map of the locations of the micro-movements image by image, which makes it possible to distinguish the more marked zones of instability of the ocular film and to observe the variation of the deformations of the film over time.

[0096] It is also possible to calculate and store the time of appearance of micro-movement points of the tear film and / or the kinetics of appearance of micro-movement points of the tear film 560 which can be useful for comparing results during a treatment.

[0097] In addition, all or part of the calculated data can be stored in order to monitor the patient and compare said data over several examinations.

[0098] When the measurement is completed, the practitioner has at his disposal, on the one hand, a spatial and temporal mapping of the micro-movements of the tear film within the corneal surface and, on the other hand, a temporal curve tracing the evolution of these micro-movements of the tear film as a function of time. The amplitude at each point of the curve will reveal the level of stability at a given moment. The slope of this temporal curve will reveal the kinetics of the appearance of the ruptures of the tear film. This allows for a more precise interpretation of the examination carried out.

[0099] There figure 16 represents a graph 700 with on the ordinate 701 the normalized sum NΣSp filtered with a moving average on three images and on the abscissa 702 the images for five patients C1 to C5.

[0100] Patients C1 and C2 show little micro-movement, i.e. low instability due to their ocular film as a function of time, patient C3 shows medium instability but did not hold the eye open during the complete sequence, while patients C4 and C5 show high instability representative of dry eye.

[0101] There figure 17 represents a graph 710 with the cumulative value A between successive images on the ordinate 711 and the images on the abscissa 702. This graph amplifies the visibility of the growth of micro-movement zones for patients C4 and C5 with dry eye, while patients C1 and C2 show very little progression, with patient C3 showing average progression up to blinking at image 28.

[0102] On both graphs, temporal zones 703, 713 of approximately two seconds centered around the 18th image, or in this case 6 seconds after blinking, have been represented. It can be seen from the curves that this examination duration of around 6 seconds can already allow the results to be normalized and the patients without dry eye problems to be distinguished from those who do have them without the risk of reaching an examination duration that could cause untimely blinking.

[0103] The invention is not limited to the examples described above, only by way of example, but it encompasses all the variants such as a distribution or a progression of the height of the different lines that may be envisaged by those skilled in the art within the framework of the protection sought. In particular, as stated above, the lines of the pattern which are horizontal parallel lines according to the example shown may be replaced by vertical parallel lines, a rotation of the image making it possible, for example, to apply the image processing means for detecting the deformations of the lines to this configuration without changing the direction thereof.

Claims

1. A method for measuring the stability of a tear film of a patient by means of a device comprising: - a backlit translucent plate (10a) equipped with a test chart (10) that is positioned in front of at least one eye (101) of a patient (100), the test chart being provided with a pattern consisting of a succession of alternating light and dark lines (12, 13) that reflect from the eye of the patient, - at least one digital photographic camera (22, 23) connected to a computing system (30) provided with means for processing and analyzing images, an objective of the camera pointing toward the patient's eye in order to photograph a reflection of the pattern of the test chart from the patient's eye, said method being such that, - the means for processing and analyzing images are configured to - detect deformations (76) of said light or dark lines of the pattern of the test chart reflected from the eye of the patient and - identify, via a comparison of the position of image points on the edge of the lines with respect to an estimated line edge (74, 74') of the lines, tear-film micro-movements revealed by these deformations, the method comprises, from a start time (t0) at which an eyelid blinks: - successively capturing (200) images, - successively detecting (510) regions of micro-movement in the images and successively computing and storing data on the positions, area and number of regions of micro-movements (520) of the tear film in each image by means of the means for processing and analyzing images, and - computing (530) and storing (540), on the basis of the successively computed positions, area and number of regions of micro-movement of the tear film, a measurement of the amplitude of the regions of micro-movement as a function of time.

2. The method for measuring the stability of a tear film as claimed in claim 1, comprising capturing (200) an image every 0.1 to 0.5 seconds and preferably every 0.3 seconds.

3. The method for measuring the stability of a tear film as claimed in claim 1 or 2, wherein the step of successively capturing images ends (570) with the first occurrence of one of the following events: end of a time delay or detection of a next eyelid blink.

4. The method for measuring the stability of a tear film as claimed in claim 1, 2 or 3, comprising tracking (400, 410, 420, 430, 440, 460) the patient's eye or eyes by means of an iris-tracking method, so as to reposition the regions of deformation of the tear film that are detected and stored with respect to the analyzed eye.

5. The method for measuring the stability of a tear film as claimed in any one of the preceding claims, wherein the estimated line edges are computed by polynomial regression (285).

6. The method for measuring the stability of a tear film as claimed in claim 5, comprising determining a population of points representative of the regions of micro-movement observed in the image of the lines by computing (610), along the lines, the absolute value of the distance to the polynomial along the axis perpendicular to the general direction of the lines of pixels P1 to Pn of line edges of the image that are separated from the polynomial by a distance dP in pixels larger than a threshold (620), and storing (620), for the representative points P1 to Pn, the distances dP1 to dPn.

7. The method for measuring the stability of a tear film as claimed in claim 6, comprising, for each image, for said population of representative points, computing (640), for each pixel P1 to Pn, the equivalent area Sp = dP x pixel width x pixel height, and computing (650) the sum ΣSp of the equivalent areas SP1 to SPn from the image points to their polynomial over the entire image and a normalized sum NΣSp of said distances by means of division of the sum ΣSp by the total length LT of the lines found in the image.

8. The method for measuring the stability of a tear film as claimed in claim 7, comprising computing (660) an overall score consisting of the sum A of the normalized sums NΣSp over a series of images from the start time t0 to a given time t.

9. The method for measuring the stability of a tear film as claimed in any one of the preceding claims, comprising generating (550) a map of the locations of the micro-movements image by image.

10. The method for measuring the stability of a tear film as claimed in any one of the preceding claims, comprising computing and storing (560) the time of appearance of the points of micro-movement of a tear film and / or the rate of appearance of points of micro-movement of the tear film.

11. The method for measuring the stability of a tear film as claimed in any one of the preceding claims, comprising storing, in a database, all or some of the computed data with a view to monitoring the patient and comparing said data over a plurality of examinations.

12. A computer program comprising instructions for implementing the method as claimed in any one of the preceding claims when this program is executed by a processor.

13. A computer-readable non-volatile storage medium on which is stored a program for implementing the method as claimed in any one of claims 1 to 11 when this program is executed by a processor.