GB2586446A

GB2586446A - Improvements relating to ocular imaging

Info

Publication number: GB2586446A
Application number: GB1911312.5A
Authority: GB
Inventors: Alexander Rabinowicz Simon; Vaghela Uddhav
Original assignee: Vui Diagnostics Ltd
Current assignee: Vui Diagnostics Ltd
Priority date: 2019-08-07
Filing date: 2019-08-07
Publication date: 2021-02-24
Also published as: GB201911312D0

Abstract

The device comprises a housing 2 configured to be attachable to the subject’s head and an imaging unit 4 movably coupled to the housing and configured to capture a plurality of images of an eye of the subject. The imaging unit is configured to be moved relative to the housing about a predetermined reference position, such that at least some of the plurality of images capture different portions of the subject’s eye. The device may comprise an adjustment unit (14, fig 2B) having a rotational component 16 to rotate the imaging unit about the reference position and a translational component 18 to translate the imaging unit radially relative to the reference position. There is also described a method of imaging an eye of a subject using a head-mounted imaging device.

Description

IMPROVEMENTS RELATING TO OCULAR IMAGING

TECHNICAL FIELD

The present invention relates to improvements in techniques relating to imaging of a subject's eye, and more specifically to a device and method for collecting and processing ocular images of the subject.

BACKGROUND

The visualisation, monitoring and/or diagnosis of ocular conditions and diseases (e.g. macular degeneration, glaucoma etc.), as well as other systemic diseases that can affect ocular functions (e.g. diabetes mellitus, arterial hypertension etc.), requires the visualisation and examination of a subject's (or patient's) fundus oculi'. This term is generally used in the field of ocular imaging to refer to the concave interior of the subject's eye, and encompasses the retina, choroid, sclera, optic disk and vasculature / blood vessels. For ease of reference, subsequent references to imaging of the 'subject's eye' should be taken to refer to imaging at least a portion of the subject's / patient's fundus oculi unless explicitly stated otherwise.

Various types of ophthalmoscopic instruments exist and may be used by medical practitioners / doctors / ophthalmologists / optometrists to examine and visualise a subject's fundus oculi. Some examples of existing instruments are shown in Figure 1.

The so-called 'gold standard' for high-resolution ophthalmic imaging involves the use of non-portable benchtop systems that can provide cross-sectional and/or wide field-of-view (FoV) images of the retina and anterior segment (also known as the anterior cavity, this includes structures in the eye in front of the vitreous humour such as the cornea, iris, ciliary body and lens). An example of such a system may be seen in Figure 'IA; as will be appreciated, the subject would place their chin on the support in order to position their eye(s) in the appropriate location for the imaging to be carried out.

These benchtop systems include substantial lighting, optical and computerised systems, with associated patient-specific mechanical calibration and adjustment capabilities. Such systems are hence advantageous in that they facilitate reliable, repeatable image capture with minimal operator variability. However, due to their size and design / operation complexity, the cost per unit is high. As a result, these benchtop systems have limited availability in under-resourced areas; their size and cost also makes them impractical for use in the primary care setting.

At the other end of the spectrum are hand-held ophthalmoscopes, which are portable and comprise simpler lighting and optical systems. Such systems work by directing illumination into the subject's eye, the illumination being directed co-axial to the eye of the doctor, operator or medical practitioner carrying out the examination, thereby enabling direct visualisation of the retina. An example of a hand-held direct ophthalmoscope is shown in Figure 1B. As an alternative to this hand-held direct ophthalmoscope, the camera sensor of a smartphone (or other hand-held mobile device) can be used to carry out retinal imaging in a similar manner to that described above. Use of a smartphone or other mobile device with an in-built display screen allows for the imaged portions of the eye to be easily visualised via the display screen during the imaging process. In addition, adaptors (such as those described in US 2016/0296111 and WO 2015/100291) can be attached to the smartphone camera to provide improved imaging capability. In some cases, a doctor or other medical practitioner may utilise an indirect ophthalmoscope, an example of which is shown in Figure 1C -this comprises a light source attached to a headband (worn on the head of the doctor) which is used in combination with a hand-held lens to view images of the subject's eye.

However, whilst such portable devices may overcome the usability and affordability issues associated with the benchtop systems, the hand-held systems do have their own associated shortcomings. In particular, the imaging capabilities of these hand-held devices are limited by their narrow field-of-view. Furthermore, the operation of such devices and hence, to an extent, the results obtained are user-dependent. For example, the positioning and/or motion of the hand-held device relative to the subject's eye; as well as the variability in proximity to the subject's eye during examination can cause a fluctuation in the field-of-view captured or that is visible to the operator of the device. In addition, as a result of the imaging techniques used with existing portable ophthalmoscopes, the device operator will only be able to transiently observe specific segments of the eye at any given time. Obtaining a global impression of the subject's eye and / or making a medical diagnosis on the basis of these images is very difficult, even for trained specialists in the field.

It is against this background, and with the aim of overcoming at least some of the disadvantages set out above, that the present invention has been devised.

SUMMARY OF THE INVENTION

According to an aspect of the invention, there is provided a head-mounted device for ocular imaging of a subject. The device comprises a housing configured to be attachable to the subject's head. The device also comprises an imaging unit movably coupled to the housing and configured to capture a plurality of images of an eye of the subject. The imaging unit is configured to be moved relative to the housing about a predetermined reference position, such that at least some of the plurality of images capture different portions of the subject's eye.

Advantageously, the above-described device is able to alter the field-of-view (FOV) of the imaging unit by moving the imaging unit relative to the housing (and hence relative to the subject's eye that is being imaged). Different (at least partially overlapping) portions of the subject's eye can therefore be captured during a single imaging scan cycle. As a result, when the plurality of individual images is subsequently combined during image post-processing, the final composite image has an effective FOV that is greater than that of the imaging unit's sensor on its own, since this effective FOV corresponds to the combined FOV of the plurality of captured images. In this manner, an overall picture of a desired portion of the subject's eye (e.g. the retina or cornea) can be built up from multiple images, without requiring the subject to move their head during the imaging process. This also improves the repeatability of the movements of the imaging unit, improving the comparability of images taken at different points in time. This may improve the ease with which subsequent diagnoses can be performed on the basis of the final images. Furthermore, as will be appreciated, the provision of the entire imaging system in a head-mounted device improves the portability of the device, and the ease with which it may be used.

The device may further comprise an adjustment unit configured to movably couple the imaging unit to the housing and to move the imaging unit relative to the housing. In such an instance, the adjustment unit may be programmed to carry out a specific predetermined sequence of movements as part of the imaging cycle. This further improves the repeatability of an imaging scan, thereby enabling subsequent scans of a subject's eye to be more easily compared with previous scans, thereby enabling changes in the imaged portions of the eye to be more easily detectable.

In some embodiments, the adjustment unit may be configured to cause the imaging unit to carry out spiral orbital motion about the predetermined reference position. The use of spiral orbital motion is advantageous as it enables at least partial overlap to be obtained between captured images of the subject's eye, thereby improving the ease with which the subsequent image processing can be carried out.

In such embodiments, the adjustment unit may comprise a rotational component configured to rotate the imaging unit about the predetermined reference position; and a translational component configured to translate the imaging unit radially relative to the predetermined reference position. In this instance, advantageously, the adjustment unit may move the imaging unit in multiple degrees of freedom to effect the resultant spiral orbital motion. This allows each mode of movement to be controlled separately from one another, thereby simplifying the physical implementation mechanisms that combine to provide the more complicated spiral orbital motion.

The imaging unit may be configured to be coupled to the rotational component via the translational component. Advantageously, as the mechanism for attachment of the imaging unit to the rotational component also provides the mechanism for translational movement of the imaging unit, this simplifies the number of moving parts utilised during the device construction. This in turn leads to increased compactness of the overall device, further improving its portability.

In some embodiments, the rotational component may comprise a rotary disk having a curved outer surface to which the imaging unit is configured to be coupled. By coupling the imaging unit to the surface of the rotary disk component, smooth movement of the imaging unit along a consistent, repeatable path is maintained. In addition, it is noted that where the curved outer surface of the rotary disk component effectively mirrors the curvature of the subject's eye, coupling the imaging unit to the outer curved surface of the eye means that the curvature of the disk's surface will guide the movement of the imaging unit, and maintain an approximately constant distance between the imaging unit and the portion of the subject's eye that is being imaged. This makes it easier to ensure that steady focused images are captured, even when the imaging unit is moved (rotated and/or translated) during an imaging scan.

The curved outer surface of the rotary disk may comprise an arcuate slot, and the translational component may be configured to interface with and move relative to the slot. 35 Advantageously, this ensures that the camera moves outwards away from the subject's eye as it rotates around the eye -spiral motion is thereby achieved and the imaging unit is able to capture a large area of the desired portion of the subject's eye (e.g. the retina) at different angles and perspectives to alter the portion of the eye contained within the FoV of the imaging unit during the course of the scan. The above-described configuration also improves the repeatability of the movements that are carried out by the adjustment unit.

In some embodiments, the adjustment unit further comprises a guide component configured to interface with and guide the rotational motion of the rotational component. This enables smooth repeatable rotational motion to be achieved during the imaging scan cycle; this is particularly in the case where the guide component is also configured to attach the rotational component within the device housing.

In some embodiments, the housing may comprise a headset configured to be mounted on the subject's face and over at least one of the subject's eyes. In this manner, the imaging unit and adjustment unit may be easily mounted to the subject's head and located in the appropriate position relative to the subject's eye for the predetermined reference position to be determined easily and reliably.

In some cases, the housing may comprise a cover plate or back cover configured to close an open side of the housing anterior to the subject's face. Advantageously, this protects the movable components of the device, and also minimises the amount of ambient light to which the subject's eye is exposed during imaging. In addition, other components may also be located on the back cover of the housing to, for example, assist with focusing a subjects gaze during the imaging scan.

For example, at least one illumination source may be provided and located on the cover plate and within the housing. When illuminated, this source therefore serves as a target for the subject to focus and fixate on during imaging -this technique is generally referred to as 'static pupillary fixation'. In other words, the illumination source provides a steady reference point, for the contralateral, non-imaging eye, upon which the subject may fixate and thereby maintain a forward gaze within the imaged eye.

In some cases, an array of illumination sources may be provided and located on the cover plate and within the housing, the array of illumination sources being arranged to be illuminated according to a predetermined sequence. Each of the sources therefore provides an individual target for the subject to focus and fixate on as described above, such that the gaze direction of the subject can be altered by changing the source (or sources) that is illuminated -this technique is generally referred to as 'dynamic pupillary fixation'. By illuminating the sources in a particular sequence, when the subject's eye focuses on each consecutive source, the portion of the subject's eye that is imaged by the imaging unit will vary. This advantageously enables the proportion of the subject's eye (or specific sub-components thereof) captured in the overall FOV of the final composite image to be increased.

In some embodiments, the housing or headset may comprise a cover configured to be mounted over one of the subject's eyes. For example, this cover may take a form akin to that of an eye patch which may be held on the subject's face (or fastened to the subject's face using straps for example) covering the eye that is to be imaged. Advantageously, the same adjustment mechanisms described above can be implemented in such a cover which makes the device more compact and increases its portability, since it only needs to cover one eye at a time.

In some embodiments, the adjustment unit may comprise a plurality of adjustment mechanisms. Each of the plurality of adjustment mechanisms may be configured to cause the imaging unit to carry out translational motion relative to the predetermined reference position; and/or pivotal motion about a central viewing axis passing through the predetermined reference position. In such embodiments, the adjustment mechanism would not cause the imaging unit to carry out spiral orbital motion. Instead, the imaging unit may be translated along x and y directions relative to the user's eye (i.e. horizontally and vertically); the imaging unit may also be pivoted or titled back and forth (i.e. towards and away from the subject's face) in combination with such translational movements. The above-described configuration advantageously increases the number of degrees of freedom in which the imaging unit may move, and hence enables greater flexibility and control of the articulation and positioning of the imaging unit relative to the subject's eye.

In such cases, the plurality of adjustment mechanisms may comprise any one or more of the following: a dual lead screw configuration; a pulley system; or a rack and pinion mechanism.

In some embodiments, the device may further comprise a complementary arrangement configured to enable translational movement of the imaging unit within the housing. For example, a linear guiderail provided on the housing with a complementary bearing may be provided on the adjustment unit, thereby enabling the adjustment unit to slide along the guide rail and effect horizontal translation within the housing. Where the guiderail extends along the majority of the length of the housing, this advantageously enables bilateral imaging of the subject's eyes to be carried out.

In some embodiments, the imaging unit may comprise a light source for illumination of the subject's eye; and an image sensor such as a camera configured to capture light reflected from the subject's eye. This maintains the compactness of the device configuration, and also ensures that the light source and image sensor maintain a fixed alignment within the imaging unit, whilst the imaging unit is being rotated and/or translated within the housing by the adjustment unit.

In some embodiments, the light source is Infrared light. This advantageously ensures that wider pupil diameters can be maintained, and since the pupil diameter constitutes the primary limitation to the available imaging window of the interior of the eye, an increase in pupil diameter increases the amount of the desired portion of the subject's eye (e.g. retina,) that can be imaged. However, other wavelengths of light, for example different wavelengths of visible light, can be used instead of IS light. The wavelength of light selected depends on the desired imaging depth and detail, and the type and amount of resulting information that is to be extracted from the images. This is because different wavelengths of light can achieve different tissue penetration, and thus when detected by the image sensor, will show results that have different clinical significance. For example, blue light wavelengths could be used to image the superficial retina; whilst green light wavelengths could penetrate deeper and image intermediate retinal layers.

According to another aspect of the invention, there is provided a method of imaging an eye of a subject using a head-mounted imaging device comprising a housing and an imaging unit movably coupled to the housing. The method comprises attaching the housing of the device to the subject's head; and moving the imaging unit relative to the housing and about a predetermined reference position. The method further comprises, while moving the imaging unit, capturing a plurality of images of the eye of the subject with the imaging unit, such that at least some of the plurality of captured images capture different portions of the subject's eye As previously set out in relation to the device, the above-described method is able to increase the effective field-of-view (FOV) of the imaging unit -by moving the imaging unit relative to the housing (and hence relative to the subject's eye that is being imaged), different (at least partially overlapping) portions of the subject's eye can therefore be captured during a single imaging scan cycle, such that, the final composite image has an effective FOV that corresponds to the combined FOV of the plurality of captured images. In this manner, an overall picture of a desired portion of the subject's eye (e.g. the retina or cornea) can be built up from multiple images, without requiring the subject to move their head during the imaging process, improving the repeatability of the movements of the imaging unit, and the resulting image generated.

In some embodiments, moving the imaging unit relative to the housing comprises moving the imaging unit along a spiral orbital path relative to the predetermined reference position. For example, the predetermined reference position may correspond to the corneal apex, and effecting spirally outward orbital motion means that during the course of an imaging scan cycle, as much of the desired portion of the subject's eye (e.g. the retina) can be imaged as possible.

In some embodiments, moving the imaging unit along the spiral orbital path comprises iteratively repeating the following steps: rotating the imaging unit about the predetermined reference position; and translating the imaging unit radially outwardly from the predetermined reference position by a set increment. This provides a mechanism for defining the imaging scan parameters, ensuring that a given imaging scan (for a particular subject) is repeatable subsequently, since the two sets of motions can be controlled and pre-programmed. In addition, the set increment may be varied to alter the overlap between images if so desired; this can complement or offset changes in the type of light used and/or the degree of pupillary dilation that occurs.

In some embodiments, the method may further comprise: prior to the moving and capturing steps, calibrating the imaging unit by adjusting the position of the imaging unit to correspond with the predetermined reference position. The calibration process and the associated adjustment of the imaging unit relative to the subject's eye, advantageously ensures accurate alignment of the camera / image capture sensor with the eye, and specifically with the predetermined reference position (e.g. the corneal apex), before any imaging is carried out. Such calibration adjustments can be achieved manually, or using (preferably) automated mechanised systems.

Calibrating the imaging unit may comprise adjusting the position of the imaging unit relative to the housing in any one or more of the following ways: (a) lateral translation of the imaging unit; (b) vertical translation of the imaging unit; (c) anterior-posterior translation of the imaging unit; or (d) rotation of the imaging unit.

Advantageously, the above-described method ensures optical alignment of the imaging unit with the subject's pupil and corneal apex (via calibration and adjustment in the x-and y-directions -horizontally and vertically). It also ensures that the camera is located at an optimal distance from the corneal apex for steady imaging (via calibration and adjustment in the anterior-posterior, z, direction).

In some embodiments, the method may further comprise: segmenting a specific portion of the subject's eye from each of the plurality of captured images; and combining the segmented images to generate a composite image of the specific portion of the subject's eye. For example, portions of the subject's retina or cornea may be segmented out from each of the captured images and stitched together so as to obtain a composite image that has a larger effective FOV than would be obtainable via capture of a single image on its own. This improves the amount of information and detail obtainable from the imaging scan.

In such cases, combining the segmented images may comprise: identifying corresponding features present in consecutive ones of the plurality of segmented images; and mosaicking the segmented images together by matching the corresponding features identified. This may be carried out automatically using an image mosaicking algorithm. In some instances, some or all of process (i.e. the calibration, imaging scan cycle and the subsequent post-processing and generation of a composite image) may be automated. This improves the repeatability and consistency of the imaging scan and of the resultant image obtained, making subsequent review and comparison of the output images for diagnostic purposes easier.

It will be appreciated that the features and advantages described in relation to the device are also applicable to the method, and vice versa.

Within the scope of this application it is expressly intended that the various aspects, embodiments, examples and alternatives set out in the preceding paragraphs, in the claims and/or in the following description and drawings, and in particular the individual features thereof, may be taken independently or in any combination. That is, all embodiments and/or features of any embodiment can be combined in any way and/or combination, unless such features are incompatible. The applicant reserves the right to change any originally filed claim or file any new claim accordingly, including the right to amend any originally filed claim to depend from and/or incorporate any feature of any other claim although not originally claimed in that manner.

BRIEF DESCRIPTION OF THE DRAWINGS

The above and other aspects of the invention will now be described, by way of example only, with reference to the accompanying drawings, in which: Figures 1A, 16 and 1C show various examples of known ophthalmoscopic instruments that are currently used by medical practitioners / optometrists; Figures 2A and 2B show perspective diagrams of a device for ocular imaging according to an embodiment of the present invention; Figure 3 is a flow diagram illustrating steps in a method of ocular imaging according to an embodiment of the present invention; Figures 4A, 4B and 4C show perspective diagrams (Figs. 4A and 4B) and a side plan view with expanded inset (Fig. 4C) of a primary adjustment component comprised in the device of Figures 2A and 28 according to an embodiment of the present invention; Figures 5A and 56 show a perspective exploded diagram and a perspective exploded schematic diagram respectively of a secondary adjustment component and an imaging unit comprised in the device of Figures 2A and 2B according to an embodiment of the present invention; Figure 6 shows a schematic diagram of the internal components of the imaging unit of Figures 5A and 5B according to an embodiment of the present invention; Figures 7A, 76 and 7C provide illustrations of various calibration mechanisms for the device of Figures 2A and 2B according to an embodiment of the present invention; Figure 8A shows a perspective diagram of a device for ocular imaging according to another embodiment of the present invention, and Figure 8B shows a cover plate for the housing of the device in Figure 8A; Figures 9A, 9B and 9C show back lower-left perspective, back upper-right perspective and back plan views respectively of a device for ocular imaging according to another embodiment of the present invention; and Figures 10A 10B, and 10C show front, back and side diagrams of a device for ocular imaging according to another embodiment of the present invention.

In the drawings, like features are denoted by like reference signs.

DETAILED DESCRIPTION

A specific embodiment of the invention will now be described in which numerous features will be discussed in detail in order to provide a thorough understanding of the inventive concept as defined in the claims. However, it will be apparent to the skilled person that the invention may be put into effect without the specific details and that in some instances, well known methods, techniques and structures have not been described in detail in order not to obscure the invention unnecessarily.

With reference to Figures 2A and 2B, there is provided a device 1 for imaging portions of the fundus oculi of a subject, for example, for carrying out retinal or corneal imaging of the subject, in accordance with an embodiment of the present invention.

In its most general form, the device 1 comprises a housing, support or stage 2 that is configured to be attached or mounted to the head of a subject undergoing an eye examination or scan. The device 1 also comprises an imaging unit 4 which is movably coupled to the housing 2. The imaging unit 4 comprises a light source 6 (as seen in Figures 5A and 53) configured to illuminate the subject's eye, and an image sensor or image capture device (such as a camera) 8 (as seen in Figures 5A and 5B) configured to capture illumination reflected from the subject's eye. The imaging unit 4 is hence configured to capture a plurality of images of the subject's eye whilst the imaging unit 4 is moving. In other words, when the device housing 2 is mounted to the subject's head, the imaging unit 4 is movable relative to the housing 2 whilst imaging or scanning one of the subject's eyes.

The portions of the eye that are imaged correspond to those portions of the eye that fall within a field-of-view of the imaging unit (i.e. the imaging unit's sensor or image capture device) at the time the image is captured. The field-of-view of the imaging unit will generally be unable to encompass the entire eye (or even the entire retina) in a single image due to the physical limitations imposed by the pupillary diameter which defines the imaging window for the patient's eye (as is described in more detail subsequently). However, by capturing a plurality of images of the subject's eye whilst the imaging unit 4 is moving, the position and/or orientation of the imaging unit 4 relative to the subject's eye will vary. The field-of- view of the imaging unit 4 may therefore encompass different (yet still potentially overlapping) portions, sections or parts of the subject's eye over the course of the imaging process. The plurality of images can subsequently be combined to build up or generate a final composite image having an effective field-of-view corresponding to the combined fields-of-view of the plurality of captured images used to create the composite image. In other words, the final image may comprise all of the different portions of the subject's eye that were contained in each of the plurality of images captured initially. For example, in the case of retinal imaging, the imaging unit 4 may capture multiple images of the subject's eye at different positions / orientations, each image containing at least some of the retina. A composite image of the retina can then be generated by combining the individually-captured images. This is described in greater detail below with reference to Figure 3.

To begin with, prior to imaging the subject's eye, an initial reference position is predefined or predetermined with respect to the subject's eye, and the imaging unit 4 is aligned with this predetermined reference position -this initial position of the imaging unit 4 will hereafter be referred to as the 'neutral starting position'. When carrying out a scan of the subject's eye, the imaging unit 4 is then moved about or around that predetermined reference position, executing a series of pre-programmed movements whilst carrying out a scan of the subject's eye and capturing images of the subject's eye at predetermined intervals. The movements of the imaging unit 4 and the image capture timings are programmed / defined to ensure that at least some of the images captured during the scan will contain different (although potentially still overlapping) portions of the eye being imaged. In other words, over the course of the scan, a field-of-view of the imaging unit 4 (specifically of the image sensor or image capture device 8 comprised therein) will cover or encompass varying portions of the subject's eye.

The ability of the above-described device 1 to vary the portion of the subject's eye that is captured in a given image during the course of the imaging process is particularly advantageous in view of the shape of a human eyeball. As is known in the field of ophthalmology, a subject's eyeball is generally convex (i.e. curved and extending outwards away from the subject's eye socket and head), and this curvature of the eye needs to be taken into consideration when attempting to image portions of the eye such as the retina or cornea, in order to ensure that these portions are captured in as much detail as possible. In the above example, the predetermined reference position is defined to correspond to the corneal apex of the eye -i.e. the point of maximum curvature of the eye, or in other words, the most anterior or outwardly-extending portion of the subject's eye. Using the corneal apex to provide an initial reference position for the imaging unit 4 will therefore ensure that the imaging unit is initially aligned with the pupil at the centre of the subject's eye. The pupil acts as a window to view the fundus oculi, and therefore appropriate illumination and positioning of the imaging unit relative to the pupil results in the diameter of the pupil being substantially equivalent to the visible fundus oculi.

An overview of an ocular imaging process 300 carried out by the above-described device will now be provided, with reference to Figure 3. It will be appreciated that this process may also be applied in order to image specific portions of the subject's eye (e.g. the retina or the cornea) via appropriate adjustment of factors such as the focal length of the image capture device 8.

To begin with, in Step 305 the device 1 is mounted, attached or otherwise held on the subject's head / face (for example, using straps or other fastening / attachment mechanisms), such that the portion of the device comprising the imaging unit 4 is positioned generally proximal to and facing the eye of the subject that is to be scanned. Once the device 1 has been mounted to / placed on the subject's head / face, the imaging unit 4 is calibrated in Step 310 to ensure that the neutral starting position of the imaging unit 4 is aligned with the predetermined initial reference position (i.e. the corneal apex). For example, the imaging unit 4 may be moved relative to the housing 2 (and relative to the subject's eye) in six degrees of freedom -horizontal or left-right (x-axis); vertical or up-down (y-axis); and anterior-posterior (z-axis) -until the imaging unit 4 is aligned with the corneal apex of the eye to be examined.

Subsequently, in Step 315, the imaging unit 4 then executes an imaging scan cycle. This comprises moving the imaging unit 4 relative to the predetermined corneal apex reference position, for example, by executing a pre-programmed series or sequence of movements, and scanning the subject's eye (e.g. capturing a video recording) whilst moving the imaging unit 4. In the case of retinal imaging for example, the pre-programmed series of movements carried out by the imaging unit 4 are calculated so as to ensure that the field-of-view of the imaging unit 4 will have covered substantially all of the subject's retina during the course of the scan. As previously mentioned, the portions of the eye that have been captured within the obtained images can subsequently be combined to build up an overall picture of the subject's retina (or indeed any other desired part of the subject's eye), since these images have been captured at various different positions and/or orientations relative to the subject's eye. Once the imaging scan cycle is complete, the captured images can be processed in Step 320 to obtain a final retinal image or set of retinal images for subsequent analysis.

During the image processing, individual image frames captured during the scan are processed (for example, in the case of a video recording, the individual image frames are isolated from the video feed), and the retina is then identified and segmented out or extracted from each image frame. For example, a U-net (a known neural network developed for biological image segmentation) could be used to provide the segmentation functionality. These segmented retinal images are then mosaicked, interlaced or stitched together in Step 325 -specifically, features in each image frame are tracked and features which overlap between frames are matched using homography -in order to obtain a final composite retinal image. This mosaicking process can be carried out using a publicly-available algorithm such as 'boofcv1.

The final image generated in this manner can then be output to the operator or medical practitioner carrying out the scan, who can then assess and analyse the image in order to perform a diagnosis or to facilitate a routine medical check-up.

A more detailed description of the specific device embodiment of Figures 2A and 2B will now be provided, with additional reference to Figures 4 to 6.

As shown in Figures 2A and 2B, the device housing 2 comprises a generally rectangular enclosure to which the imaging unit 4 is mounted, and substantially within which the imaging unit 4 is contained. Mien the device is mounted to the subject's head, the housing 2 extends across an upper portion of the subject's face, covering their eyes and at least a portion of their nose. The housing enclosure comprises an inner surface 2a that is arranged to face towards and rest against the subject's face when the device 1 is mounted to the subject's head; and top 2b and bottom 2c walls which extend perpendicularly outwards from the inner surface 2a away from the subject's head.

The housing design incorporates several features that are designed to improve the fit of the device to the subject's head. The housing inner surface 2a has a curved profile which generally mirrors the curvature of the subject's face, such that when the housing 2 is mounted to the subject's head, the inner surface 2a of the housing may rest substantially flush against the subject's face. The housing inner surface 2a also comprises a curved opening or slot 10 which is arranged to be generally in line with the plane of the user's eyes when the housing 2 is mounted to the subject's head; this opening 10 provides the imaging unit 4 with access to scan the subject's eye(s). The housing also comprises inbuilt cushioning that is designed to mould to the contours of the subject's face; as well as various recesses and cut-outs, for example, a cut-out 12 in the bottom wall 2c that is intended to receive the subject's nose.

The device 1 further comprises an adjustment unit 14 that is configured to couple the imaging unit 4 to the housing 2, and to move the imaging unit 4 relative to the housing 2, and thereby relative to the subject's eye(s). The adjustment unit 14 comprises a primary, rotational adjustment component 16 (shown in more detail in Figures 4A and 4B); and a secondary, translational adjustment component 18 (shown in more detail in Figures 5A and 5B). The primary rotational adjustment component 16 is configured to enable rotational movement of the imaging unit 4 relative to the housing 2; the secondary translational adjustment component 18 is configured to attach the imaging unit 4 to the primary adjustment component 16, and to enable translational movement of the imaging unit 4 relative to the primary adjustment component 16. The combination of rotary and translatory movement provided by the two adjustment components 16, 18 achieves the desired imaging functionality of the device 1: namely, enabling portions of the subject's eye to be captured during the course of an imaging cycle and hence, cumulatively, increases the portion of the subject's eye that can be imaged.

Specifically, in the illustrated embodiment, the primary adjustment component 16 comprises a rotary disk 20 or plate having a curved outer (anterior) surface 20a which, when the device 1 is mounted to the subject's head, faces outwards / away from the subject's eye and generally mirrors the curvature of the subject's eye. The curved outer disk surface 20a comprises an arcuate slot 22, opening or cut-out (more easily visible in Figure 4) extending radially outwards from the centre of the disk towards its circumference. The secondary adjustment component 18 engages with and protrudes from the arcuate slot 22, and is arranged for relative movement within and along the slot 22. In more detail, part of the secondary adjustment component 18 couples or is otherwise attached to the primary adjustment component 16, on the curved outer surface 20a of the rotary disk component 20.

The remainder of the secondary adjustment component, comprising the imaging unit, then protrudes through the arcuate slot 22 of the rotary disk component 20, extending in a posterior direction away from the curved outer surface 22a of the rotary disk component and towards the subject's eye. The movement of the rotary disk component is driven by a first motor 24 and results in rotational or orbital motion of the imaging unit 4 relative to a fixed central point (for example, the predetermined reference position corresponding to the corneal apex), thereby enabling 3600 imaging of the subject's eye. The radial translation of the secondary adjustment component 18 along the curved disk surface 20a and within the arcuate slot 22 is driven by a second motor 26 and results in radial motion of the imaging unit 4 relative to the fixed central point, such that increasing areas of the subject's eye are captured within the field-of-view of the sensor or image capture device 8 of the imaging unit 4 as the imaging unit 4 moves radially outwards, towards the periphery of the rotary disk component 20.

A sequence of motorised movements of the individual components of the adjustment unit 14, relative to a predetermined reference position, can therefore be pre-programmed to define one imaging cycle during which a dynamic scan of desired portions of the subject's eye is carried out. In the illustrated embodiment, the predetermined reference position is defined to be the corneal apex of the subject's eye. The imaging unit 4 is initially positioned at the radially inner-most point within the arcuate slot 22 of the rotary disk component 20, and calibration is carried out (as will be described in more detail subsequently with reference to Figure 7) to align this initial position of the imaging unit 4 with the corneal apex of the eye in question, and to define the neutral starting point of the imaging unit 4.

A first step of the imaging cycle is then carried out which comprises two sets of consecutive motorised movements: first, rotation of the rotary disk component 20 such that the imaging unit 4 undergoes a complete revolution around the corneal apex; second, the secondary adjustment component 18 is translated along the arcuate slot 22 by a set increment, such that the imaging unit 4 is moved radially outwards towards the periphery of the rotary disk component 20. These two sets of movements are then iteratively executed (in the above order) in subsequent steps of the imaging cycle. The increment between each successive radial movement of the secondary adjustment component 18 is predefined to ensure a sufficient overlap (e.g. around 10% overlap) in the field-of-view of two consecutive images, so as to enable the subsequent image mosaicking to be performed accurately. Translation of the imaging unit 4 to the maximum circumferential extent of the arcuate slot 22, followed by performance of a complete revolution in this position, marks the end of the imaging cycle.

The imaging unit 4 will then be returned to its neutral starting position.

As a result, the imaging unit 4 executes a spiral movement path of increasing radius about the corneal apex of the subject's eye in the course of an imaging cycle, whilst still ensuring that the central viewing axis of the imaging unit's sensor 8 is fixated on the pupil of the subject's eye. Throughout the imaging cycle, the imaging unit 4 will be active, and the image sensor! image capture device 8 in the imaging unit 4 will be scanning the subject's eye and capturing images. This image capture may be carried out substantially continuously -i.e. capturing a video feed comprising multiple consecutive image frames -or with larger temporal intervals between successive images (for example, configuring the camera to capture an image prior to executing each rotation). The captured images may then be recorded and saved for future analysis or reference (for example, downloaded from the device to a digital storage facility).

In summary, the imaging unit 4 is able to move in multiple degrees of freedom relative to the subject's eye and to the device housing 2. The imaging unit 4 is therefore able to independently visualise different regions of the fundus oculi during one cycle of operation, without requiring operator involvement. This greatly increases the flexibility in the ocular (and specifically, retinal) imaging that is achieved, and allows larger portions / sections of the subject's eye to be imaged (since the final output image corresponds to a combination of the individually-captured images). However, the portability and ease of operation of the device 1 is still maintained. Furthermore, the automated nature of the imaging cycle, as well as the range of movements that may be carried out by the imaging unit 4, improves the reproducibility of the imaging process and thereby also of the resulting images.

Now that the various device components, and how they operate to carry out a scan of the subject's eye, have been broadly described, further details of each of these components On particular, the imaging unit 4 and the adjustment unit 14, and how they interface with one another) will now be described with reference to Figures 4A & 4B and Figure 5A & 5B.

Figures 4A, 4B and 4C illustrate further details of the primary adjustment component 16 according to an embodiment of the present invention. In addition to the above-described rotary disk component 20, the primary adjustment component 16 comprises a guide component 28 which couples the rotary disk component 20 to the device housing 2, and is configured to drive and guide the rotational motion of the rotary disk component 20.

The rotary disk component 20 comprises a circumferential lip 30 that extends substantially continuously around the disk's periphery. Offset radially inwards from the circumferential lip and located on the inward-facing surface 20b (relative to the subject's face when the device is mounted on the subject's head) of the rotary disk component 20 is a curved array of teeth 32, which also extends in a circle around the curved disk surface 20b. When the rotary disk component 20 is coupled with the guide component 28, the array of teeth 32 on the rotary disk component 20 is arranged to interface with the first motor 24, which is integrated into the guide component 28. Engagement between this motor 24 and the array of teeth 32 enables rotational movement of the rotary disk component 20 to be driven.

The guide component 28 comprises a planar portion 34 configured in use to extend between the top 2b and bottom 2c walls of the device housing 2, and to retain the primary adjustment component 16 within the housing. A C-shaped curving protrusion 36 extends laterally from the planar portion 34 and is configured to engage with the rotary disk component 20. Specifically, in the illustrated embodiment, intermittently distributed, curved projections 38 are provided on a radially-outer portion of the toothed array 32 of the rotary disk component 20; these are shown in Figure 4A and in greater detail in Figure 4C. These projections 38, combined with the continuous lip 30 provided on the outer face and circumference of the rotary disk component 20, create a channel 40 having a width that is substantially equivalent to the thickness of the edge of the C-shaped protrusion 36. The channel 40 provided by the rotary disk component 20 is therefore configured to receive and interface with the edge of the guide component's C-shaped protrusion 36. As a result of the engagement between rotary disk 20 and guide 28 components, the rotational movement of the rotary disk component 20 is guided by the C-shaped protrusion 36 of the guide component 28.

Alternatively, in some embodiments, the C-shaped protrusion 36 of the guide component 28 comprises a guide channel (not shown) on its inner surface which extends along some or all of the curved arc of the C-shaped protrusion 36. In such embodiments, the lip 30 of the rotary disk component 20 may then be configured to be received in, and to rotate within, the guide channel provided by the guide component 28.

In addition, advantageously, the overall, combined profile of the guide component's C-shaped protrusion 36 is greater than a half-circle, which restricts undesirable lateral displacement of the rotary disk component 20 during rotation. Whilst a complete circular profile could be utilised instead, the illustrated design of this C-shaped protrusion 36, which also tapers towards its open edge, can also accommodate the subject's nose.

The rotary disk component 20 also comprises the second motor 26, located on its outer (relative to the subject's eye when in use) curved surface 20a, which is configured to interface and engage with complementary teeth formations 42 on the secondary adjustment component 18, so as to drive translational movement of the secondary adjustment component 18 (and its associated imaging unit 4) along the arcuate slot 22 of the rotary disk component 20.

We now turn to Figures 5A and 5B which illustrate additional details of the secondary adjustment component 18 according to an embodiment of the present invention. The secondary adjustment component 18 comprises a base portion 44 configured to interface with the rotary disk component 20, and a complementary cover portion 46 configured to engage with and at least partially house the imaging unit 4.

The base portion 44 has a concave upper surface 44a which mirrors the curvature of the outer surface 20a of the rotary disk component 20, such that the base portion 44 of the secondary adjustment component 18 rests substantially flush against the curved outer surface 20a of the rotary disk component 20 when the two components are coupled together. This enables accurate transfer of the orbital motion of the rotary disk component 20 to the associated imaging unit 4. The previously-mentioned array of teeth formations 42 extend outwards from a lateral surface of the base portion 44, and engage with the second motor 26 located on the curved outer surface 20a of the rotary disk component 20 to drive the translational movement of the secondary adjustment component 18 within the arcuate slot.

Alternatively, it would be possible for the motor 26 to be located on the inner surface 20b (relative to the subject's eye when in use) of the rotary disk component 20. In this case, the complementary teeth formations 42 would be located on a corresponding portion of the secondary adjustment component 18.

An elongate male formation or extension 48 protrudes from the upper surface 44a of the base portion 44, and is shaped and configured to interface with a complementary female formation or recess 50 provided within the cover portion 46. The male formation 48 and the cover portion 46 are both narrower in width than the base portion 44 -the male formation 48 and the cover portion 46 are therefore able to extend inwardly through the arcuate slot 22 of the rotary disk component 20 towards the inner surface 2a of the device housing 2. In the illustrated embodiment, the male extension 48 corresponds to an extruded member which is contiguous with the base portion. The length of the male formation 48 may also be varied as desired to alter the proportion of the male formation 48 that is contained within the recess 50 of the cover portion 46, and thereby vary the position of the imaging unit 4 relative to the subject's eye for optimal imaging. The male formation 48 is retained within the cover portion recess 50 via the provision of magnets 52 provided on opposing surfaces of the two portions, which enables for easy interchangeability of either of these components as desired. However, it will be appreciated by the skilled person that other forms of attachment (either permanent or non-permanent) may be utilised instead.

Figures 5A, 55 and 6 also highlight further details of the imaging unit 4 according to an embodiment of the present invention. The imaging unit 4 comprises an imaging unit housing 54 which contains the previously-mentioned light source 6, for example an Infrared (IR) or visible light (VL) source which illuminates a desired portion of the subject's eye that is to be imaged. The imaging unit 4 also comprises the previously-mentioned image sensor or image capture device 8 provided within the imaging unit housing, for example an IR or VL camera which captures images of the illuminated portion of the subject's eye. It will be appreciated that in some embodiments, the use of IR illumination can be advantageous, due to the insensitivity of the pupillary light reflex to IR light (i.e. the degree of pupillary constriction in response to illumination). Wider pupil diameters can be maintained with IR illumination than with similar intensities of VL illumination, irrespective of ambient light, thereby allowing for larger portions of the retina to be captured. The IR image acquired, with variances in the degree of greyscale, can be recoloured during image post-processing.

It will be appreciated that whilst IR and VL sources are described above, alternative light sources having a wide range of output wavelengths can be used instead, depending on the desired imaging depth and detail, and the type and amount of resulting information that is to be extracted from the images. This is because different wavelengths of light can achieve different tissue penetration, and thus when detected by the image sensor, will show results that have different clinical significance. For example, blue light wavelengths could be used to image the superficial retina; whilst green light wavelengths could penetrate deeper and image intermediate retinal layers.

In the embodiment illustrated in Figure 5B, the light source 6 is housed within a recess on a sidewall 56a of an upper portion 56 of the imaging unit housing 54. The image sensor 8 is mounted within a lower portion 58 of the imaging unit housing 54 using fasteners (e.g. screws). The upper portion 56 of the imaging unit housing 54 is attached to a cover plate 60 provided on the top of the lower portion 58 of the imaging unit housing 54 (using fasteners such as screws). The lower portion 58 of the imaging unit housing 54 is configured to be received within an upwardly-opening recess 62 provided in the top of the cover portion 46, such that the cover plate 60 extends across the top of the cover portion 46, and is secured to the cover portion 46 so as to attach the imaging unit housing 54 to the secondary adjustment component 18. In the illustrated embodiment, lateral extensions! tabs 60a, 60b are provided on the sides of the cover plate 60, and configured to be received within complementary recesses 64a, 64b provided on the tops of the cover portion sidewalls 64. Fasteners (such as screws) may be used to attach the two components to one another.

As shown in Figure 6, the (high-intensity) light source 6 is located perpendicular to a central axis of the image capture device 8. The imaging unit 4 comprises a diaphragm 66 having a small (less than around 1 mm) aperture to collimate the light and generate a focussed beam for maximal illumination of the retina. The imaging unit also comprises a beam splitter plate 68 placed after the diaphragm 66 and angled at around 45 degrees such that the collimated light is reflected by 90 degrees towards the visual field and anterior to (i.e. away from) the camera lens. This light is then reflected by 180 degrees from the subject's eye, passes through the beam splitter 68 and is captured by the image capture device (i.e. a camera) 8. Advantageously, the beam splitter 68 is partially reflective, and ideally is around 50% reflective, and 50% transparent -a sufficient degree of transparency is required as the beam splitter 68 is located in front of the camera 8, but this needs to be balanced with a sufficient degree of reflectivity to ensure visual clarity for image capture.

To minimise undesirable scatter from light that is transmitted through the beam splitter 68 and reflected from internal walls of the imaging unit housing, a light-absorbing wall 70 aligned at 45 degrees (but oriented at the opposite angle to the beam splitter) is placed opposite to the beam splitter 68. To accommodate for variations in subject! patient refractive error (for example, myopia and hypermetropia), adjustment of the focus of the camera 8 can be performed manually or via a computer vision guided autofocus system. To minimise corneal glare, circular polarisers (not shown), orientated perpendicular to each other, can be placed anterior to the light source 6 and anterior to the camera lens.

In some embodiments, other optical configurations could be utilised. For example, a single or circular configuration of LEDs co-axial to the central axis of the camera could be used instead of a beam splitter 68.

The calibration process for accurately aligning the imaging unit 4 relative to a predetermined reference position of the subject's eye (i.e. the corneal apex), and for determining the neutral starting position of the imaging unit 4, will now be described with reference to Figures 7A, B and C. Due to the natural variability in facial and cranial anatomy, optimal operation and thus visualisation of the retina (or other parts of the subject's eye) can only be achieved by maintaining the stability of the imaging unit 4, and when the initial starting point of the imaging unit (prior to executing an imaging cycle) is aligned orthogonally with the apex of the cornea and directly in front of the pupil. This facilitates the inclusion of the pupil within the full range of movement of the imaging unit 4, and is advantageous because (as previously mentioned) the pupil acts as a window to view the fundus oculi.

Calibration of the device, and the associated adjustment of the imaging unit 4 relative to the subject's eye, is therefore useful to ensure accurate alignment of the camera / image capture sensor 8 with the eye, and specifically with the predetermined reference position corresponding to the corneal apex, before any retinal imaging can be carried out. Such calibration adjustments can be achieved manually, or using (preferably) automated mechanised systems.

As shown in Figures 7A, 7B and 7C, calibration of the device 1 is carried out in three orthogonal planes to ensure optical alignment with the subject's pupil and corneal apex (i.e. calibration and adjustment in the x-and y-directions -horizontally and vertically), as well as to ensure that the camera 8 is located at an optimal distance from the corneal apex for steady imaging (i.e. calibration and adjustment in the z-direction -backwards and forwards).

First considering calibration in the x-direction (horizontally; or in the transverse plane across the subject's face), as shown in Figure 7A, the imaging unit 4 is adjusted laterally relative to the subject's nasal midline (and also relative to the device housing 2) to align the camera / sensor 8 in the imaging unit 4 horizontally with the corneal apex. Such x-direction calibration is carried out to take into account the Interpupillary distance (IPD) of the subject, since this distance varies between subjects. The design of the housing 2 is intended to accommodate a range of IPDs (e.g. between 58 to 70 mm) and x-direction calibration is then used to optimise the alignment for a given subject. In the above-described embodiment, linear guiderails 72 are provided on the top and bottom walls 2b, 2c of the device housing 2 (seen most clearly in Figures 2A and 2B), and complementary grooves or channels 74 are provided along corresponding top and bottom positions on the primary adjustment component 16 (and specifically on the guide component 28) -this is seen most clearly in Figures 4A and 4B. The linear guiderails 72 on the housing 2 are arranged to be received in and to engage with the grooves 74 of the primary adjustment component 16, such that the entire adjustment unit 14 can be slidingly translated linearly along the guiderails 72 relative to the housing 2. This movement may be carried out manually, or via a motorised rack and pinion setup, or pulley system; such motorised setups may also be automated and controlled via a control processor (not shown). The control processor may be provided within the device itself, for example, it may be configured to be located within the device housing, or atop or below the housing, and be operatively connected to the device components. For example, a raspberry pi computing processor may be provided which is capable of controlling the various motor units and coordinating the camera recording with the motor unit movement using multithreading.

Now considering calibration in the y-direction (vertically, elevation or depression along the coronal plane of the subject's face), as shown in Figure 7B, the imaging unit 4 (and in fact the entire device 1 itself) is adjusted vertically relative to the subject's eyes, to align the camera / sensor 8 in the imaging unit vertically with the corneal apex. Such y-direction calibration is carried out to take into account the subject's face shape and curvature, and to ensure that the device housing 2 is substantially parallel to the coronal plane of the subject's head. In the above-described embodiment, the y-direction adjustment of the device housing is carried out using adjustable fastening means 76 which also (at least partially) provide the functionality to mount the device 1 to the subject's head. Such adjustable fastening means 76 can include, for example, ratcheted braces or Velcro straps, which can be easily adjusted by tensioning or loosening the braces / straps. As with the x-direction calibration, these adjustments may also be carried out either manually, or via (automated) motorised interactions.

Finally, calibration of the imaging unit in the z-direction (i.e. adjusting the imaging unit 4 to move it towards or away from the subject's face; anterior / posterior along the sagittal plane of the subject's head) is also carried out, as shown in Figure 7C, to ensure that the camera 8 is located at the optimal distance from the subject's corneal apex to capture a steady, focused image. Such z-direction calibration is carried out to account for the fact that the distance of the camera 8 may not be consistent across varying facial dimensions (i.e. will vary between subjects). For example, the inherent z-direction offset of the imaging unit 4 from the subject's face will likely vary between subjects due to: the design of the imaging unit 4 itself, as well as the design of the adjustment unit 14 (for example, to permit internal motion of these device components relative to one another); and the design of the device housing 2 (for example, to incorporate cushioning within the device housing 2, or to accommodate the facial shape of the subject).

The optimal distance between the corneal apex and the camera is dictated by the geometrical constraints of the curved outer surface radius of the rotary disk component. This is shown in Figure 7C, where the optimal distance between the corneal apex and the curved surface of the rotary disk component is achieved in the left-hand image -i.e. this distance corresponds to the radius R of the rotary disk component's surface arc length -but not in the right-hand image.

Because the curved surface of the rotary disk component 20 guides the orbital motion of the imaging unit, if the rotary disk component 20 is too close or too distant from the corneal apex, the pivoting point of the imaging unit 4 will be offset from the corneal apex. As the imaging unit 4 incrementally translates radially outwards towards the perimeter of the rotary disk component 20, the central camera axis would no longer be positioned over the corneal apex; this would result in a visible wobble' of the image captured which is undesirable. Nevertheless, the calibration process can exploit the presence of this 'wobble' using a dynamic feedback mechanism: test rotations of the rotary disk component can be performed whilst the subject's eye is imaged (e.g. using a live camera feed); the degree of 'wobble' in the resulting images of the live feed can then be assessed, and z-direction adjustment can be performed substantially in real-time so as to minimise the Wobble'. When the Wobble' has been minimised, the camera feed should show a substantially stable image.

As with the x-and y-direction calibrations, the anterior-posterior (z-direction) adjustments of the imaging unit 4 can be performed either manually, or automatically via mechanised interactions; for example using a rack and pinion mechanism, or leadscrews.

For all of the above calibration processes, operator and/or computer visualisation using a live camera feed is utilised -i.e. the camera 8 would be actively capturing images of the subject's eye during the calibration process (e.g. via a live video feed that may be displayed to the operator via an operatively-connected display). The live camera feed would include an overlay graticule or indicator (indicating the central viewing point of the camera / image sensor) which is used to guide the calibration adjustments: optimal alignment of the imaging unit On the x-and y-directions) would hence be considered to be achieved when the overlay graticule coincides with the corneal apex.

Furthermore, automating the motorised calibration processes involves implementation of a dynamic feedback loop by a control processor(s) to carry out the respective adjustments -the visual input from the live camera feed can be processed and the offsets of the overlay graticule from the corneal apex in the x-, y-and z-directions can be determined. The appropriate adjustments that then need to be made to the imaging unit in order to reduce these offsets are then translated into instructions for driving the various motors. The maximum adjustments that may be made in each of the three calibration planes are also predefined, and/or are determined by sensors (such as end-stops, not shown) that are provided within the device housing 2. These are taken into account by the control processor (not shown) when determining the appropriate adjustments that are to be made.

The device housing 2 in the above embodiments is shown and described as having an open' structure -i.e. having no outer (anterior) wall such that the imaging unit 4 and other device sub-components are readily accessible from the anterior side of the device housing 2 when it is mounted to the subject's head. However, in some embodiments (such as that shown in Figure 8), the device housing may comprise an outer wall, also referred to as a back cover or cover plate 78. In other words, rather than having an 'open' structure in which the device components are readily accessible by the operator / medical practitioner, the device housing 2 has a 'closed' structure where the imaging unit 4 is substantially entirely enclosed within the device housing 2. This minimises the amount of ambient light entering the device housing 2, and hence also the total amount of ambient light exposed to the eye.

This advantageously provides an imaging environment in the headset akin to the 'dark room' required for ophthalmic examinations and to facilitate pupillary mydriasis.

Additionally, in some 'closed' housing embodiments, the back cover 78 can incorporate a low-intensity central light source (not shown). The presence of this light source provides a steady reference point, for the contralateral, non-imaging eye, upon which the subject may fixate and thereby maintain a forward gaze within the imaged eye; such a setup is therefore advantageous compared with an 'open' device housing design (having no back cover plate) where the subject is simply advised to fixate on an arbitrary distant object at the centre of their field of view and eye line height. In such embodiments, a wide-angle lens may also be included along the optical path between the light source and the subject's eye, to increase the apparent distance of the fixation point for comfortable viewing.

Alternatively or additionally, in some 'closed' housing embodiments, an array of low-intensity light sources may be positioned at specific locations on the back cover 78. An example of such 'closed housing' embodiments is illustrated in Figure 8A, in which it can be seen that a plurality of openings 80 are provided on the surface 78a of the back cover (internal to the housing); each opening 80 is configured to receive and retain one of the light sources in the array (although it will be appreciated that not all of the openings need to contain a light source). The back cover 78 is shown in greater detail in Figure 8B, in which it can be seen that the back cover 78 comprises an outer wall 78a in which the openings 80 are provided, as well as orthogonally-extending top 78b and bottom 78c walls. The top and bottom walls 78b, 78c of the back cover 78 are configured to be attached to the corresponding top 2b and bottom 2c walls of the main device housing 2 (e.g. using fasteners such as screws) to form a complete 'closed' housing. A back plate / shield 82 is provided which is attached to the outer surface of the back cover 78. Furthermore, the light sources in the array may be in operative communication with a control processor (not shown, but configured to be located within the device housing, or atop or below the housing), which is programmed to illuminate the light sources according to a specific predefined sequence, thereby providing individual targets for the subject to focus and fixate upon. This technique is referred to as 'dynamic pupillary fixation'. The light sources in the array may be arranged in a specific configuration to facilitate the imaging of the subject's eye. For example, consecutively-illuminated light sources may be arranged at predefined distances from one another, such that when the subject's eye focuses on each consecutive source, the portion of the subject's eye that is imaged by the imaging unit 4 will vary by an amount that changes the field-of-view captured but still enables sufficient (e.g. around 10%) overlap between fields of view in consecutive images to be achieved (to allow for subsequent image mosaicking). This dynamic pupillary fixation technique advantageously enables the proportion of the fundus oculi (or specific subcomponents thereof) captured in the overall field-of-view of the final composite image to be increased.

Many modifications may be made to the above examples without departing from the scope of the present invention as defined in the accompanying claims.

For example, the adjustment unit 14 used to move the imaging unit relative to the device housing may be altered to replace orbital motion with translational motion along, and pivotal motion with respect to, various orthogonal axes, as is shown in the embodiment illustrated in Figures 9A, 9B and 9C.

As shown in these figures, the previously-described adjustment unit 14 (comprising the rotary disk component 20, the guide component 28 and the secondary adjustment component 18) has been replaced by an alternative adjustment unit 14'. This alternative adjustment unit 14' comprises a mobile platform or stage 84 configured to support an imaging unit 4 and to move the imaging unit relative to the device housing 2. This mobile support platform 84 is mounted upon, and translatable along, a linear guide component 86 which, in the illustrated embodiment extends across substantially the entire horizontal extent of the device housing 2. The support platform 84 utilises a rack and pinion mechanism to execute this translational motion: a motor 88 (e.g. a rotary gear) provided in the support platform 84 that is configured to engage and interact with an array of teeth 90 provided on the linear guide component 86. Interaction between the motor 88 and teeth 90 drives translatory movement of the support platform 84 (and hence of the imaging unit) relative to the subject's eye in the x-direction. In the illustrated embodiment, as the linear guide component 86 extends substantially the entire length of the device housing 2, the mobile platform 84 and its associated imaging unit 4 can be translated a sufficient distance to image both eyes without requiring removal of the device 1 from the subject's head (achieving substantially seamless bilateral imaging).

In the illustrated embodiment, the linear guide component 86 also comprises a linear guide rail 92 which is configured to extend parallel to the array of teeth 90, and to interact with a bearing 94 provided in the base of the support platform 84 -the support platform 84 therefore slides along the linear guide rail 92 when carrying out its translatory motion, which reduces the associated friction.

The linear guide component 86 is also movably coupled to the device housing 2 using adjustable fasteners 96, whereby adjustment of the fasteners 96 varies the vertical position of the associated imaging unit relative to the device housing 2, and to the subject's eye. For example, a dual lead screw configuration may be provided at opposite ends of the linear guide component 86 (as shown in the illustrated embodiment). Simultaneous clockwise or counter-clockwise rotation of both lead screws would cause elevation or depression of the support platform 84 (and its associated imaging unit) relative to the subject's eye.

The above-described features therefore enable the position of the imaging unit 4, relative to the subject's eye, to be adjusted within the horizontal (x) and vertical (y) planes.

Furthermore, the imaging unit 4 is also pivotably coupled to the mobile support platform 84 to enable the imaging unit 4 to carry out pivotal pitch and yaw movements, in order to maintain alignment of the imaging axis with the corneal apex when the imaging unit is adjusted in the vertical and horizontal planes respectively. For example, to implement pitch adjustment, the support platform 84 (shown most clearly in Figure 90) comprises a vertically-extending arm 98 which supports a first motor (not shown) configured to engage with a support of the imaging unit 4, and to pivot the imaging unit 4 in an anterior-posterior direction relative to the subject's eye. In addition, the support platform 84 further comprises a second motor (not shown, but typically incorporated into the base of the platform 84) which is configured to implement yaw adjustment by pivoting or rotating the mobile platform 84 in a left-right direction (about a vertical axis) relative to the subject's eye.

The device embodiment 1' illustrated in Figures 9A to 90 is hence able to provide comparable results to the device embodiment 1 illustrated in Figure 2A -the adjustment unit shown in Figures 9A to 9C provides the functionality of moving the imaging unit 4 around relative to the subject's eye, so as to enable images of multiple, different (overlapping) portions of the subject's eye to be imaged and subsequently analysed.

Furthermore, it will be appreciated that the adjustment unit 14' of Figure 9 advantageously enables greater flexibility and control of the articulation and positioning of the imaging unit relative to the subject's eye. In a similar manner to that described in respect of the adjustment unit 14 of Figure 2 to 7, the movements that are carried out by the adjustment unit 14' of Figure 9 may also be executed automatically -a series of pre-programmed movements or combinations of movements of the mobile support platform 84 and the imaging unit 4 may be executed (optionally separated by appropriate 'hold intervals') during an imaging cycle. Ultimately, as mentioned above, a corresponding degree of mapping of the subject's eye to that achieved by the device embodiment 1 of Figures 2 to 7 is also achieved using the adjustment unit 14' and the device embodiment 1' of Figure 9.

Another alternative embodiment of a device 100 for ocular imaging is also envisaged and is illustrated in Figures 10A, 10B and 10C. In these figures, the device embodiment 1 of Figure 2 to 7 has been altered by replacing the generally rectangular device housing 2 with a roughly hemispherical device housing 102. In other words, the imaging unit 4 and adjustment unit 14 (originally described in the embodiment of Figures 2 to 7) have been isolated from the device housing 2 shown in Figure 2 and effectively constitute an imaging device in their own right. However, the device housing 102 of Figures 10A & 10B is configured to only cover one of the subject's eyes when in use, and can be mounted or held onto the subject's head in a similar manner to an eye patch or eye cover (e.g. using straps or other fasteners). In order to improve the fit of the device housing 102 against the subject's face, the boundary morphology 104 of the device housing 102 can be altered from its hemispherical curvature and configured to accommodate the shape of the subject's nose. In addition, at least part of the boundary of the device housing 102 (e.g. the portion that interfaces with the subject's nose) can be made of a compliant cushioned material, which may be moulded to suit the shape of the subject's nose. In addition to increasing the comfort with which the device housing 102 fits on the subject's face, the use of such material would make it possible for the housing design to be substantially hemispherical and symmetrical.

Such a housing design could then be used for bilateral imaging (i.e. the device could be used to image either of the subject's eyes, since the compliant cushioning could be moulded when the device is worn to fit the shapes of the left or right sides of the subject's nose). The imaging unit 4 of this embodiment is configured to move in a corresponding manner to the imaging unit in the embodiment of Figures 2 to 7 -namely a combination of orbital / rotational motion using the rotary disk component 20, and radial translation of the imaging unit 4 across the curved surface 20a of the rotary disk component 20 (using the secondary adjustment component 18), causing the imaging unit 4 to implement a radially increasing spiral movement during the imaging cycle.

As will be appreciated based on the design of this embodiment, adjustment / calibration of the imaging unit 4 relative to the subject's eye in the horizontal and vertical directions is envisaged to be carried out manually (although still with the assistance of the live video feed and overlay graficule described previously). With regard to z-direction calibration and adjustment, it is envisaged that the boundary morphology 104 of the device housing 102 may be deformable to assist with this by moulding to the user's facial anatomy. A range of different devices having different housing boundary morphology shapes may be used to accommodate different subjects' facial shapes.

It will be appreciated that, as the adjustment unit 14 and imaging unit 4 of the Figures 10A & 10B embodiment are substantially identical to the corresponding components in the embodiment of Figure 2, corresponding advantages to those described previously in respect of the Figure 2 configuration will also apply in respect of the Figures 10A & 10B configuration. In addition, as the embodiment of Figures 10A & 10B only covers a single eye of the subject, it has the advantage of being lightweight and requiring less material to construct.

Claims

CLAIMS1. A head-mounted device for ocular imaging of a subject, the device comprising: a housing configured to be attachable to the subject's head; an imaging unit movably coupled to the housing and configured to capture a plurality of images of an eye of the subject; wherein the imaging unit is configured to be moved relative to the housing about a predetermined reference position, such that at least some of the plurality of images capture different portions of the subject's eye.
2. The head-mounted device of claim 1, wherein the housing comprises a headset configured to be mounted on the subject's face and over at least one of the subject's eyes.
3. The head-mounted device of claim 1 or claim 2, further comprising an adjustment unit configured to movably couple the imaging unit to the housing and to move the imaging unit relative to the housing.
4. The head-mounted device of claim 3, wherein the adjustment unit is configured to cause the imaging unit to carry out spiral orbital motion about the predetermined reference position.
5. The head-mounted device of claim 3 or claim 4, wherein the adjustment unit comprises: a rotational component configured to rotate the imaging unit about the predetermined reference position; and a translational component configured to translate the imaging unit radially relative to the predetermined reference position.
6. The head-mounted device of claim 5, wherein the rotational component comprises a rotary disk having a curved outer surface to which the imaging unit is configured to be coupled.
7. The head-mounted device of claim 6, wherein the imaging unit is configured to be coupled to the rotational component via the translational component.
8. The head-mounted device of claim 6 or claim 7, wherein the curved outer surface of the rotary disk comprises an arcuate slot, and the translational component is configured to interface with and move relative to the slot.
9. The head-mounted device of any of claims 5 to 8, wherein the adjustment unit further comprises a guide component configured to interface with and guide the rotational motion of the rotational component.
10. The head-mounted device of any preceding claim, wherein the housing comprises a cover plate configured to close an open side of the housing anterior to the subject's face.
11. The head-mounted device of claim 10, further comprising at least one illumination source located on the cover plate and within the housing.
12. The head-mounted device of claim 10 or claim 11, further comprising an array of illumination sources located on the cover plate and within the housing, the array of illumination sources being arranged to be illuminated according to a predetermined 15 sequence.
13. The head-mounted device of any of claims 1 to 9, wherein the housing comprises a cover configured to be mounted over one of the subject's eyes.
14. The head-mounted device of claim 3, wherein the adjustment unit comprises a plurality of adjustment mechanisms, each of the plurality of adjustment mechanisms configured to cause the imaging unit to carry out one or more of the following: translational motion relative to the predetermined reference position; or pivotal motion about a central viewing axis passing through the predetermined reference position.
15. The head-mounted device of claim 14, wherein the plurality of adjustment mechanisms comprise any one or more of the following: a dual lead screw configuration; a pulley system; or a rack and pinion mechanism.
16. The head-mounted device of claim 14 or claim 15, wherein the adjustment unit comprises a complementary linear guiderail and bearing arrangement for translational movement of the imaging unit within the housing.
17. The head-mounted device of any preceding claim wherein the imaging unit comprises a light source for illumination of the subject's eye; and an image sensor configured to capture light reflected from the subject's eye.
18. The head-mounted device of claim 17, wherein the light source is Infrared light.
19. A method of imaging an eye of a subject using a head-mounted imaging device comprising a housing and an imaging unit movably coupled to the housing, the method comprising: attaching the housing of the device to the subject's head; moving the imaging unit relative to the housing and about a predetermined reference position; and while moving the imaging unit, capturing a plurality of images of the eye of the subject with the imaging unit, such that at least some of the plurality of captured images capture different portions of the subjects eye
20. The method of claim 19, wherein moving the imaging unit relative to the housing comprises moving the imaging unit along a spiral orbital path relative to the predetermined reference position.
21. The method of claim 20, wherein moving the imaging unit along the spiral orbital path comprises iteratively repeating the following steps: rotating the imaging unit about the predetermined reference position; and translating the imaging unit radially outwardly from the predetermined reference position by a set increment.
22. The method of any of claims 19 to 21, further comprising: prior to the moving and capturing steps, calibrating the imaging unit by adjusting the position of the imaging unit to correspond with the predetermined reference position.
23. The method of claim 22, wherein calibrating the imaging unit comprises adjusting the position of the imaging unit relative to the housing in any one or more of the following ways: (a) lateral translation of the imaging unit; (b) vertical translation of the imaging unit; (c) anterior-posterior translation of the imaging unit; or (d) rotation of the imaging unit.
24. The method of any of claims 19 to 23, further comprising: segmenting a specific portion of the subject's eye from each of the plurality of captured images; and combining the segmented images to generate a composite image of the specific portion of the subject's eye.
25. The method of claim 24, wherein combining the segmented images comprises: identifying corresponding features present in consecutive ones of the plurality of segmented images; and mosaicking the segmented images together by matching the corresponding features identified.