Ocular aberrometer characterization system and method

The ophthalmic aberrometry system addresses measurement inaccuracies by using a wavefront sensor and logical device with a composite analysis engine to correct ocular misalignments and system aberrations, ensuring precise and real-time aberration measurements for improved surgical outcomes.

JP7592076B2Active Publication Date: 2024-11-29ALCON INC
View PDF 10 Cites 0 Cited by

Patent Information

Application Number
JP2022519543
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Filing Date
2020-10-01
Publication Date
2024-11-29
Estimated Expiration
2040-10-01

AI Technical Summary

Technical Problem

Conventional ophthalmic aberrometers are susceptible to eye movements and optical aberrations, leading to inaccurate surgical planning and suboptimal surgical or visual outcomes due to measurement errors.

Method used

An ophthalmic aberrometry system with a wavefront sensor and logical device that includes a composite analysis engine, capable of identifying and correcting ocular misalignments and system aberrations using wavefront sensor data, eye tracker data, and machine learning algorithms to provide real-time, accurate aberration measurements.

Benefits of technology

The system significantly reduces measurement errors by compensating for eye movements and system aberrations, providing reliable and accurate ocular aberrometric measurements in real-time, enhancing surgical planning and visual outcomes.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 0007592076000001
    Figure 0007592076000001
  • Figure 0007592076000002
    Figure 0007592076000002
  • Figure 0007592076000003
    Figure 0007592076000003
Patent Text Reader

Abstract

Techniques are disclosed for systems and methods for providing improved ocular aberration measurement. The ocular aberration measurement system includes a wavefront sensor configured to provide wavefront sensor data related to an optical target monitored by the ocular aberration measurement system and a logic device configured to communicate with the wavefront sensor. The logic device is configured to identify a composite analysis engine for the ocular aberration measurement system based at least in part on an aberrometer model and / or an eye model associated with the ocular aberration measurement system, the aberrometer model and eye model being based at least in part on wavefront sensor data provided by the wavefront sensor. The logic device is also configured to generate a compact analysis engine for the ocular aberration measurement system based at least in part on the identified composite analysis engine.
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] One or more embodiments of the present disclosure relate generally to ocular aberration measurement, and more particularly to systems and methods for improving ocular aberration measurement, for example, in clinical or intraoperative settings. [Background technology]

[0002] Ocular surgery may involve reshaping the cornea and / or ocular surface, insertion and / or exchange of intraocular devices and / or artificial intraocular lenses (IOLs), and / or other surgical manipulation of the active optical components of the eye. To achieve optimal postoperative visual outcomes, good preoperative clinical evaluation and surgical planning, as well as intraoperative monitoring of the execution of the surgical plan, are important.

[0003] Ocular aberration measurements performed by an ocular aberrometer are a common method typically used to characterize the eye preoperatively, monitor the progress of surgery, and evaluate the success of surgery. Conventional ocular aberrometers often suffer from a variety of measurement errors related to ocular movements and / or misalignment and optical aberrations of the aberrometer itself, which can lead to inaccurate surgical planning and suboptimal surgical or visual outcomes for the patient. Summary of the Invention [Problem to be solved by the invention]

[0004] Thus, there is a need in the art for systems and methods that improve clinical and / or intraoperative ocular aberration measurements that lead to optimized surgical or visual outcomes for patients. [Means for solving the problem]

[0005] Techniques are disclosed for systems and methods for providing improved ocular aberration measurements. According to one or more embodiments, an ocular aberration measurement system may include a wavefront sensor configured to provide wavefront sensor data related to an optical target monitored by the ocular aberration measurement system, and a logic device configured to communicate with the wavefront sensor. The logic device may be configured to identify a composite analysis engine for the ocular aberration measurement system based at least in part on an aberrometer model and / or an eye model associated with the ocular aberration measurement system, the aberrometer model and the eye model being based at least in part on wavefront sensor data provided by the wavefront sensor. The logic device may also be configured to generate a compact analysis engine for the ocular aberration measurement system based at least in part on the identified composite analysis engine.

[0006] In other embodiments, the method may include identifying a composite analysis engine for the ocular aberrometry system based at least in part on an aberrometer model and / or an eye model associated with the ocular aberrometry system, the aberrometer model and the eye model being based at least in part on wavefront sensor data provided by a wavefront sensor of the ocular aberrometry system, and generating a compact analysis engine for the ocular aberrometry system based at least in part on the identified composite analysis engine.

[0007] According to some embodiments, a non-transitory machine-readable medium may include a plurality of machine-readable instructions adapted, when executed by one or more processors, to cause the one or more processors to perform a method. The method may include identifying a composite analysis engine for the ocular aberrometry system based at least in part on an aberrometer model and / or an eye model associated with the ocular aberrometry system, the aberrometer model and the eye model being based at least in part on wavefront sensor data provided by a wavefront sensor of the ocular aberrometry system, and generating a compact analysis engine for the ocular aberrometry system based at least in part on the identified composite analysis engine.

[0008] In a further embodiment, the ocular aberration measurement system may include a wavefront sensor configured to provide wavefront sensor data related to an optical target monitored by the ocular aberration measurement system, and a logic device configured to communicate with the wavefront sensor. The logic device may be configured to receive ocular aberration measurement output data including at least the wavefront sensor data provided by the wavefront sensor, determine an estimated ocular misalignment corresponding to a relative position and / or orientation of the optical target monitored by the ocular aberration measurement system based at least in part on the received ocular aberration measurement output data, and generate user feedback corresponding to the received ocular aberration measurement output data based at least in part on the estimated ocular misalignment.

[0009] In other embodiments, a method may include receiving ocular aberration measurement output data from an ocular aberration measurement system including a wavefront sensor, the ocular aberration measurement output data including at least wavefront sensor data related to an optical target monitored by the ocular aberration measurement system, determining an estimated ocular misalignment corresponding to a relative position and / or orientation of the optical target monitored by the ocular aberration measurement system based at least in part on the received ocular aberration measurement output data, and generating user feedback corresponding to the received ocular aberration measurement output data based at least in part on the estimated ocular misalignment.

[0010] According to some embodiments, a non-transitory machine-readable medium may include a plurality of machine-readable instructions adapted, when executed by one or more processors, to cause the one or more processors to perform a method. The method may include receiving ocular aberration measurement output data from an ocular aberration measurement system comprising a wavefront sensor, the ocular aberration measurement output data including at least wavefront sensor data related to an optical target monitored by the ocular aberration measurement system, determining an estimated ocular misalignment corresponding to a relative position and / or orientation of the optical target monitored by the ocular aberration measurement system based at least in part on the received ocular aberration measurement output data, and generating user feedback corresponding to the received ocular aberration measurement output data based at least in part on the estimated ocular misalignment.

[0011] The scope of the present invention is defined by the claims, which are incorporated by reference into this section. A more complete understanding of the embodiments of the present invention will be afforded to those skilled in the art, and additional advantages thereof will be realized, by consideration of the following detailed description of one or more embodiments. Reference is first made to the accompanying drawings, which will be briefly described. [Brief description of the drawings]

[0012] [Figure 1] FIG. 1 shows a block diagram of an ocular aberration measurement system according to one embodiment of the present disclosure. [Figure 2A-B] 2A-B show block diagrams of an aberrometry characterization target for an ocular aberration measurement system according to one embodiment of the present disclosure. [Diagram 3] FIG. 3 shows a block diagram of an ocular aberration measurement system according to one embodiment of the present disclosure. [Figure 4] FIG. 4 illustrates a flow diagram of a process for characterizing an ocular aberration measurement system according to one embodiment of the present disclosure. [Diagram 5] FIG. 5 illustrates a flow diagram of a process for operating an ocular aberration measurement system according to one embodiment of the present disclosure. [Figure 6]FIG. 6 shows a diagram of a multi-layer neural network according to one embodiment of the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0013] Embodiments of the present invention and their advantages are best understood by reference to the following detailed description, it being appreciated that like reference numerals are used to identify like elements shown in one or more of the figures.

[0014] According to various embodiments of the present disclosure, ocular aberration measurement systems and methods provide substantially real-time measurement and monitoring of aberrations in a patient's eye, reducing system and measurement errors inherent to conventional systems. For example, as a patient's eye fixates during an ocular aberration measurement or examination, it naturally drifts. This fixation drift includes changes in the apex, tilt, torsional rotation, and x, y, and z positions of the eye. These misalignments introduce errors into the calculation of the wavefront characteristic aberrations of the eye. If the average of these misalignments is close to zero, averaging the wavefront aberration measurements from individual frames of the examination image stream may be sufficient to eliminate most of the errors caused by misalignment. If it is known that the average misalignment for the test sequence is not close to zero, or if it cannot be confirmed that the misalignment has a mean close to zero, the measurement error or noise can be reduced using various strategies described herein, including one or a combination of averaging wavefront sensor data (e.g., represented by Zernike polynomial expansions) whose misalignment does not exceed a preset threshold, correcting errors based on misalignment in the wavefront sensor data using a composite analysis method (described herein), identifying the fixation state of the eye (e.g., for each wavefront measurement) based on a cluster analysis applied to the series of wavefront measurements, and / or based on eye tracker data. The estimated misalignments and / or their effect on the wavefront measurements may be reduced to an eye misalignment metric and provided to a user of the ocular aberration measurement system as a display view with a corresponding graphic, or may be used to cause the ocular aberration measurement system to ignore images from the test image sequence, or to stop and / or resume the test. Thus, embodiments provide substantially real-time monitoring feedback while providing more reliable and accurate aberration measurements than conventional systems, such as by reducing variability in clinical and intraoperative ocular aberrometry tests due to eye movement during test image sequences.

[0015] In further embodiments of the present disclosure, ocular aberration measurement systems and methods provide platforms and techniques to accurately characterize system aberrations and correct wavefront sensor data that would otherwise be corrupted by misalignments associated with the patient's eye. For example, to account for errors that may be caused by system aberrations or generated during eye movement, the ocular aberration measurement systems described herein may employ one or more of two unique forms of calibration for accurate higher order aberration (HOA) analysis: a system characterization process, and a combined analysis training process.

[0016] For the system characterization process, the ocular aberrometry system may be used to measure a series of reference interferograms (e.g., in the form of wavefront sensor data) generated by a model target configured to present substantially a single type of variable aberration (e.g., defocus aberration) to the ocular aberrometry system with substantially zero misalignment. The reference interferograms may be used to characterize and / or quantify any system aberrations of a particular ocular aberrometry system, which may be used to correct the wavefront sensor data provided by the wavefront sensor of that particular ocular aberrometry system, such as by removing it prior to subsequent analysis.

[0017] For the composite analysis training process, the ocular aberrometry system may be used to capture a set of wavefront measurements generated by a model target configured to present a selection of different types and varying aberration strengths (e.g., for Zernike expansion coefficients up to the sixth order) to the ocular aberrometry system, along with variable misalignments in the tip, tilt, torsional rotation, and x, y, and z positions of the aberrating elements of the model target. The set of wavefront measurements may be used to train and / or refine a composite analysis engine, as described herein, performed by the ocular aberrometry system and configured to generate substantially accurate estimated misalignments and / or corrected wavefront sensor data, for example, based on uncorrected wavefront sensor data or a combination of uncorrected wavefront sensor data and eye tracker data. Thus, the embodiments provide more reliable and accurate aberration measurements than conventional systems, such as by improving the precision and accuracy of aberration measurements due to reduced errors resulting from system aberrations and off-axis or oblique aberration measurements. Moreover, the embodiments provide a more robust (e.g., reliable and rapid) aberration measurement system by increasing the range of misalignments that may be accurately compensated for or corrected and therefore included in a particular inspection (e.g., present and / or detected within a given image of an inspection image sequence).

[0018] Figure 1 illustrates a block diagram of an ocular aberration measurement system 100, according to one embodiment of the present disclosure. In the embodiment illustrated in Figure 1, the ocular aberration measurement system 100 may be implemented to provide substantially real-time (e.g., 30 Hz updates) monitoring of an optical target 102 (e.g., a patient's eye) while continuously compensating for common characterization errors, such as patient movement, system optical aberrations, thermal variations, vibrations, and / or other characterization errors, that would otherwise degrade the ocular aberration measurements provided by the ocular aberration measurement system 100.

[0019] As shown in FIG. 1, the ocular aberration measurement system 100 includes a beacon 110 that generates a probe beam 111 that is used to illuminate an optical target 102 for the wavefront sensor 120 and / or other elements of the ocular aberration measurement system 100. The ocular aberration measurement system 100 may also include various other sensor-specific beacons and / or light sources, such as a light emitting diode (LED) array 132 that is used to illuminate the optical target 102 for the eye tracker 130, and an OCT beacon 123 that is used to generate an OCT probe beam 124 to illuminate the optical target 102 for the OCT sensor 122. Beam splitters 112-116 are used to provide the probe beam 111 to the optical target 102 and to generate associated sensor beams 113, 115, 117 that are provided by the optical target 102 (e.g., a portion of the probe beam 111, the OCT probe beam 124, and light generated by the LED array 132 that is reflected by the optical target 102). The beacon 110, the OCT beacon 123, the LED array 132, and each sensor element of the ocular aberration measurement system 100 may be controlled by a controller 140 (e.g., via communication links 141-144), which may also act as an interface between the beacon 110, the OCT beacon 123, the LED array 132, and each sensor element of the ocular aberration measurement system 100 and other elements of the ocular aberration measurement system 100, including a user interface 146, a server 150, a distributed server 154, and other modules 148 (e.g., accessed via any of the communication links 145, 149, and 155), as shown.

[0020] In typical operation, controller 140 initializes one or more of wavefront sensor 120, optional OCT sensor 122, and optional eye tracker 130, controls beacon 110, OCT beacon 123, and / or LED array 132 to illuminate optical target 102, and receives ocular aberrometry output data (e.g., wavefront sensor data, eye tracker data, OCT sensor data) from various sensor elements of ocular aberrometry system 100. Controller 140 may process the ocular aberrometry output data itself (e.g., to detect or correct misalignment and / or extract aberrometry parameters from the wavefront sensor data) or may provide the ocular aberrometry output data to server 150 and / or distributed server system 154 (e.g., via network 152) for processing as described herein. Controller 140 and / or server 150 may be configured to receive user input at user interface 146 (e.g., to control operation of ocular aberration measurement system 100) and / or generate user feedback for display to a user via a display on user interface 146, such as a display view of ocular aberration measurement output data and / or characteristics of ocular aberration measurement output data, as described herein. Controller 140, server 150, and / or distributed server system 154 may be configured to store, process, and / or otherwise manipulate data related to operation and / or characterization of ocular aberration measurement system 100, including, for example, machine learning and / or training of a complex analysis engine (e.g., a neural network-based classification and / or regression engine), to characterize system aberrations of ocular aberration measurement system 100, detect misalignment relative to optical target 102, and / or correct wavefront sensor data and / or associated aberration classification coefficients, as described herein.In various embodiments, ocular aberration measurement system 100 may be configured to provide substantially real-time monitoring and user feedback (e.g., 30 Hz or higher frequency updates) of the optical aberration measurements of optical target 102 while continuously compensating for common characterization errors, as described herein, making ocular aberration measurement system 100 particularly well suited for clinical and intra-operative examinations.

[0021] The beacon 110 may be implemented with a laser source (e.g., producing substantially coherent light) and / or a superluminescent diode (e.g., producing relatively low coherence light, “SLD”) that may be controlled by the controller 140 to generate a probe beam 111 primarily for use with the wavefront sensor 120. The OCT beacon 123 may be implemented with a laser source and / or a superluminescent diode (e.g., producing relatively low coherence light, particularly suitable for the OCT sensor 122) that may be controlled by the controller 140 to generate an OCT probe beam 124 primarily for use with the OCT sensor 122. In various embodiments, the OCT beacon may be integrated with the OCT sensor 122 as shown, for example, integrated with the beacon 110, and / or implemented as its own standalone beacon, similar to the beacon 110 (e.g., with an appropriate arrangement of beam splitters). The LED array 132 may be implemented by a shaped or patterned array of LEDs that may be controlled by the controller 140 to illuminate the target 102 primarily for use with the eye tracker 130. The beam splitters 112-116 may be implemented by any of a number of optical components (e.g., pellicle beam splitters, mirrors) configured to aim and / or pass the probe beam 111 at the optical target 102 and redirect at least a portion of the probe beam 111 and / or a source beam generated by the optical target 102 (e.g., reflected portions of the probe beam 111 or 124 and / or light emitted from the LED array 132) towards various sensor elements of the ocular aberration measurement system 100 to form sensor beams 113-117 (e.g., sensor beams). The optical target 102 may be, for example, a patient's eye, or may be implemented by a single pass (probe beam 111 is off and optical target 102 generates its own illumination) or double pass (e.g., normal operation with probe beam 111 on) model target, for example, as described herein.

[0022] The wavefront sensor 120 may be implemented as any one or combination of devices or device architectures configured to measure aberrations of an optical wavefront, such as an optical wavefront of the sensor beam 117 generated by at least reflection of the probe beam 111 from the optical target 102, and the wavefront sensor 120 may be configured to provide associated wavefront sensor data. For example, the wavefront sensor 120 may be implemented as any one or combination of a Shack-Hartmann wavefront sensor, a phase-shifting Schlieren wavefront sensor, a wavefront curvature sensor, a pyramidal wavefront sensor, a common-path interferometer, a multi-directional shearing interferometer, a Ronchi tester, a shearing interferometer, and / or other wavefront sensors that may be configured for use in ophthalmology. The wavefront sensor data provided by the wavefront sensor 120 may be represented in a variety of forms, including Zernike coefficients, e.g., Fourier, cosine, or Hartley transforms, or Taylor polynomials in cylindrical or Cartesian coordinates, or an interferogram.

[0023] The OCT sensor 122 may be implemented with any one or combination of devices or device architectures configured to capture micrometer and / or sub-micrometer resolution two-dimensional and three-dimensional images from within an optically scattering medium, such as the optical target 102, using relatively low coherence light and low coherence interferometry, and provide associated OCT sensor data. For example, the OCT sensor 122 may be implemented as any one or combination of OCT sensor architectures that may be configured for use in ophthalmology. The eye tracker 130 may be implemented with any one or combination of devices or device architectures configured to track the orientation and / or position of the optical target 102 and / or features of the optical target 102 (e.g., retina, pupil, iris, cornea, lens), including a conventional eye tracker or a fundus camera, and provide associated eye tracker data. In some embodiments, the eye tracker 130 may be configured to capture images of one or more types of Purkinje reflections associated with the target 102.

[0024] Controller 140 may be implemented as any suitable logic device (e.g., a processing device, a microcontroller, a processor, an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), a memory storage device, a memory reader, or other device or combination of devices) that may be adapted to execute, store, and / or receive suitable instructions, such as, for example, software instructions implementing control loops or processes for controlling various operations of ocular aberration measurement system 100 and / or elements of ocular aberration measurement system 100. Such software instructions may also implement methods for processing sensor signals, identifying sensor information, providing user feedback (e.g., through user interface 146), querying devices for operating parameters, selecting operating parameters for devices, or performing any of the various operations described herein (e.g., operations performed by logic devices of the various devices of ocular aberration measurement system 100).

[0025] Additionally, a machine-readable medium may be provided for storing non-transitory instructions for loading into and execution by controller 140. In these and other embodiments, controller 140 may be implemented with other components, such as volatile memory, non-volatile memory, one or more interfaces, and / or various analog and / or digital components for interfacing with devices of ocular aberration measurement system 100, as appropriate. For example, controller 140 may be adapted to store, e.g., over time, sensor signals, sensor information, composite analysis parameters, calibration parameters, sets of calibration points, training data, reference data, and / or other operating parameters, and provide such stored data to other elements of ocular aberration measurement system 100. In some embodiments, controller 140 may be integrated with user interface 146.

[0026] The user interface 146 may be implemented as one or more of a display, a touch screen, a keyboard, a mouse, a joystick, a knob, a virtual reality headset, and / or any other device capable of accepting user input and / or providing feedback to a user. In various embodiments, the user interface 146 may be adapted to provide user input to other devices of the ocular aberration measurement system 100, such as the controller 140. The user interface 146 may also be implemented by one or more logic devices, which may be adapted to execute instructions, such as software instructions, that implement any of the various processes and / or methods described herein. For example, the user interface 146 may be adapted to form a communication link, send and / or receive communications (e.g., sensor data, control signals, user input, and / or other information), identify parameters for one or more operations, and / or perform various other processes and / or methods described herein.

[0027] In some embodiments, the user interface 146 may be adapted to accept user input to, for example, form a communication link (e.g., to the server 150 and / or the distributed server system 154), select certain parameters for operation of the ocular aberration measurement system 100, select a method for processing the sensor data, adjust the position and / or orientation of the articulated model target, and / or otherwise facilitate operation of the ocular aberration measurement system 100 and devices within the ocular aberration measurement system 100. Once the user interface 146 accepts the user input, the user input may be transmitted to other devices of the system 100 via one or more communication links. In an embodiment, the user interface 146 may be adapted to display time series of various sensor data and / or other parameters as part of a display that includes graphs or maps of such data and / or parameters. In some embodiments, the user interface 146 may be adapted to accept user input that, for example, modifies a control loop or process parameter of the controller 140 or a control loop or process parameter of any other element of the ocular aberration measurement system 100.

[0028] Other module 148 may include any one or combination of sensors and / or devices configured to facilitate operation of ocular aberration measurement system 100. For example, other module 148 may include a temperature sensor configured to measure one or more temperatures associated with the operation of one or more elements of ocular aberration measurement system 100, a humidity sensor configured to measure ambient humidity around ocular aberration measurement system 100, a vibration sensor configured to measure vibration amplitude and / or presence associated with the operation of ocular aberration measurement system 100, a patient sensor configured to measure the posture, movement, or other characteristics of a patient-supplied optical target 102, and / or other sensors capable of providing sensor data that facilitates operation of ocular aberration measurement system 100 and / or helps correct common system errors inherent to the operation of ocular aberration measurement system 100. In additional embodiments, other module 148 may include additional lighting and camera systems similar to the combination of LED array 132 and eye tracker 130 configured to capture images of one or more types of Purkinje reflexes associated with target 102.

[0029] The server 150 may be implemented, for example, similarly to the controller 140 and may include various elements of a personal computer or server computer used to store, process, and / or otherwise manipulate relatively large data sets related to one or more patients, including, for example, relatively large sets of training data, as described herein, for training neural networks or implementing other types of machine learning. The distributed server system 154 may be implemented, for example, as a distributed combination of multiple embodiments of the controller 140 and / or the server 150 and may include networking and storage devices and capabilities configured to facilitate the storage, processing, and / or other manipulation of relatively large data sets, including, for example, relatively large sets of training data, as described herein, for training neural networks or implementing other types of machine learning in a distributed manner. The network 152 may be implemented as one or more of a wired and / or wireless network, a local area network, a wide area network, the Internet, a cellular network, and / or according to other network protocols and / or topologies.

[0030] 2A-B show block diagrams of aberration measurement characterization targets 202A-B for the ocular aberration measurement system 100, according to one embodiment of the present disclosure. In the embodiment shown in FIG. 2A, the aberration measurement characterization target 202A may be implemented as a single-pass model target configured to characterize optical aberrations associated with the ocular aberration measurement system 100. As shown in FIG. 2A, the aberration measurement characterization target / single-pass model target 202A includes a laser or SLD source 260 that generates a source beam 211 through a lens system 262 and is oriented along the optical axis of the optical aberration system 100 (e.g., aligned with the probe beam 111 exiting the beam splitter 116). In various embodiments, the source 260 may be configured to generate, for example, a diverging spherical wavefront with controllable vergence power, a converging spherical wavefront with controllable vergence power, or a plane wave with zero optical power. Lens system 262 is coupled to a linear motion actuator 264 via mount 265, which enables lens system 262 to be moved along its optical axis to vary the defocus aberration of single-pass model target 202A to generate a plurality of reference interferograms and corresponding wavefront sensor data (e.g., provided by wavefront sensor 120). Such reference interferograms and / or corresponding wavefront sensor data may be aggregated and stored, for example, as an aberrometer model 360 of FIG. 3, and used to correct system aberrations associated with ocular aberration measurement system 100, as described herein.

[0031] In some embodiments, lens system 262 may be implemented as a National Institute of Standards and Technology (NIST) traceable lens, and linear motion actuator 264 may be implemented as a relatively high precision actuator stage configured to position lens system 262 at a set of positions spaced from light source 260 to generate source beam 211 having a known and predefined defocus power (e.g., defocus aberration), such as from −12 to 6 diopters or from 0 to ±5 diopters, in steps of 5.0 D, or in higher resolution steps, according to the range of defocus aberrations typically experienced by patients monitored by ocular aberration measurement system 100. The resulting aberrometer model 360 may be used to compensate for various system aberrations, including those due to shot noise, thermal variations, and vibration.

[0032] As shown in Figure 2B, the aberrometry characterization target / double-pass model target 202B includes an interchangeable ocular aberration model 270 that is removably coupled to a six degree of freedom (6DOF) motion actuator 272 via one or more mounts 273, where the 6DOF motion actuator 272 is configured to vary the position and / or orientation of the interchangeable ocular aberration model 270 to generate a plurality of selected (e.g., known) misalignments (e.g., relative to an optical axis of the ocular aberration measurement system 100) and a corresponding plurality of sets of wavefront sensor data. Such misalignments and corresponding wavefront sensor data may be aggregated and stored, for example, as eye model 370 of Figure 3, and used to train a combined or compact analysis engine to detect misalignments associated with the ophthalmic target 102 and correct the corresponding wavefront sensor data, as described herein.

[0033] In some embodiments, the interchangeable ocular aberration model 270 may form one element of a set of interchangeable ocular aberration models, each formed with a precise amount of predefined ocular aberration (e.g., represented by precise predefined Zernike coefficient amplitudes, such as amplitudes expressed in microns for Zernike expansions up to the sixth order). In one embodiment, such an interchangeable ocular aberration model may be cut on a contact lens lathe using transparent poly(methyl methacrylate) (PMMA). More generally, such an interchangeable ocular aberration model may be measured by a third party profiler with traceability to NIST for ground truth comparison with measurements made by the ocular aberration measurement system 100. In various embodiments, the 6DOF motion actuator 272 may be configured to provide micrometer resolution positioning of the interchangeable ocular aberration model 270 along the x-, y-, and z-axes, as well as microradian orientation of the interchangeable ocular aberration model 270 about the θy (tip), θx (tilt), and θz (twist) directions, as shown by representative coordinate frames 280A-B.

[0034] In general operation, each interchangeable ocular aberration model 270 in a set (e.g., a set of 10 or more) may be, in turn, attached to a 6DOF motion actuator 272, and the controller 140 may be configured to control the 6DOF motion actuator 272 to position and / or orient the interchangeable ocular aberration model 270 at a set of relative positions and / or orientations (e.g., relative to the optical axis of the ocular aberration measurement system 100) within a range of misalignments typically experienced by patients monitored by the ocular aberration measurement system 100. In one embodiment, the set of misalignments may include approximately 40,000 different misalignments. In various embodiments, the combined set of misalignments and a corresponding set of wavefront sensor data (e.g., provided by the wavefront sensor 120) may form a supervised data set (e.g., eye model 370) that may be used to identify a composite or compact analysis engine, as described herein. Such analysis engine may be used to compensate for misalignments of the optical target 102, which may also be measured by the eye tracker 130. More generally, the combination of characterizations performed by the aberration measurement characterization targets 202A-B may be used to compensate or correct both system aberrations and errors in the wavefront sensor data caused by misalignment of the optical target 102.

[0035] Figure 3 illustrates a block diagram of an ocular aberration measurement system 300, according to one embodiment of the present disclosure. In the embodiment illustrated in Figure 3, the ocular aberration measurement system 300 may be configured to use characterization data generated by the ocular aberration measurement system 100 via the aberration measurement characterization targets 202A-B to generate a composite analysis engine 350 and / or a compact analysis engine 340, which may be used during operation of the ocular aberration measurement system 100 to provide substantially real-time monitoring and user feedback (e.g., frequency updates of 30 Hz or greater) of the optical aberration measurements of the optical targets 102 while continuously compensating for common characterization errors, as described herein.

[0036] 3, ocular aberration measurement system 300 is similar to ocular aberration measurement system 100, but with additional details regarding various data structures and executable program instructions used in the operation of ocular aberration measurement system 100 or 300. For example, controller 140 is shown implemented with compact analysis engine 340, and server 150 and distributed server system 154 are each shown implemented with or storing one or more of aberrometer model 360, eye model 370, training data 392, supervised learning engine 390, combined analysis / neural network engine 350, and compact analysis engine 340. Although dashed lines generally indicate optional storage and / or implementation of certain elements, in various embodiments, controller 140, server 150, and distributed server system 154 may each implement or store any of the identified elements and / or additional elements, as described herein.

[0037] In general, the aberrometer model 360 may be generated by aggregating sensor data associated with the use of the single pass model target 202A to characterize the ocular aberration measurement system 100, the eye model 370 may be generated by aggregating sensor data associated with the use of the double pass model target 202B to characterize a parameter space associated with the optical target 102, and the training data 392 may be generated by incorporating the aberrometer model 360 and / or the eye model 370 and / or by generating and aggregating a simulated set of training data as described herein with respect to the elements of Figure 6. The supervised learning engine 350 may be implemented as a static learning engine and / or according to a procedurally generated learning engine (e.g., a genetic algorithm updatable learning engine) configured to generate the composite analysis engine 350 using the training data 392. The composite analysis engine 350 may be implemented, for example, as a deep neural network and / or with other composite analysis methodologies, including various other neural network architectures or composite analysis methodologies, including dense K-nearest neighbor (k-NN) databases for classification and / or regression, as described herein. The compact analysis engine 340 may be implemented as the composite analysis engine 350 in a compact form, such as a form more suitable for relatively low resource but high performance execution by the controller 140, and thus may be implemented as a deep neural network and / or other composite analysis methodologies. In certain embodiments, the compact analysis engine 340 may be implemented as a neural network having fewer hidden layers and / or neurons / layers than the composite analysis engine 350.

[0038] FIG. 4 illustrates a flow diagram of a process 400 for characterizing the ocular aberration measurement system 100 and / or 300 according to an embodiment of the present disclosure. It should be appreciated that any step, sub-step, sub-process, or block of the process 400 may be performed in a different order or configuration than the embodiment illustrated by FIG. 4. For example, in other embodiments, one or more blocks may be removed from the process or added thereto. Additionally, block inputs, block outputs, various sensor signals, sensor information, calibration parameters, and / or other operating parameters may be stored in one or more memories before moving to subsequent portions of the corresponding process. Although the process 400 is described with reference to the systems, processes, control loops, and images described with reference to FIGS. 1-3, the process 400 may be performed by other systems that differ from those systems, processes, control loops, and images, for example, including a different selection of electronic devices, sensors, assemblies, mobile structures, and / or mobile structure attributes.

[0039] In block 402, an aberrometer model associated with the ocular aberration measurement system is generated. For example, the controller 140, the server 150, and / or the distributed server system 154 may be configured to control the light source 260 of the single-pass model target 202A positioned as the optical target 102 monitored by the ocular aberration measurement system 100 to generate the light source beam 211 through the lens system 262 to illuminate the wavefront sensor 120 and / or other elements of the ocular aberration measurement system 100. The controller 140 may be configured to vary the defocus aberration of the single-pass model target 202A according to a plurality of selected defocus refractive powers to generate, for example, a plurality of sets of wavefront sensor data provided by the wavefront sensor 120. The controller 140, the server 150, and / or the distributed server system 154 may be configured to determine system aberrations associated with the ocular aberration measurement system 100 based at least in part on the plurality of sets of wavefront sensor data provided by the wavefront sensor 120. The associated set of system aberrations and / or wavefront sensor data may be stored (eg, on server 150 and / or distributed server system 154) as an aberrometer model 360.

[0040] In block 404, an eye model associated with the ocular aberration measurement system is generated. For example, the controller 140, the server 150, and / or the distributed server system 154 may be configured to control the beacon 110 of the ocular aberration measurement system 100 to generate a probe beam 111 to illuminate a double-pass model target 202B positioned as an optical target 102 monitored by the ocular aberration measurement system 100, which in turn illuminates one or more of the wavefront sensor 120, the eye tracker 130, the OCT sensor 122, and / or other elements of the ocular aberration measurement system 100 (e.g., via reflection of the probe beam 111). The controller 140 may be configured to vary the position and / or orientation of the exchangeable ocular aberration model 270 of the double-pass model target 202B relative to the optical axis 111 of the ocular aberration measurement system 100 according to a plurality of selected misalignments to generate a corresponding plurality of sets of wavefront sensor data provided by, for example, the wavefront sensor 120. The selected misalignments and / or corresponding sets of wavefront sensor data may be stored (e.g., on the server 150 and / or the distributed server system 154) as an eye model 370. Similar techniques may be used to incorporate eye tracker data from the eye tracker 130 and OCT sensor data from the OCT sensor 122 into the eye model 370.

[0041] At block 406, a composite analytical engine is identified. For example, the controller 140, the server 150, and / or the distributed server system 154 may be configured to identify the composite analytical engine 350 based at least in part on the aberrometer model 360 generated at block 402 and / or the eye model 370 generated at block 404. In some embodiments, the controller 140, the server 150, and / or the distributed server system 154 may be configured to form a deep neural network 600 including an input layer 620, an output layer 640, and at least one hidden layer 630-639 coupled between the input layer 620 and the output layer 640, each of which includes a plurality of neurons. The controller 140, server 150, and / or distributed server system 154 may be configured to train at least a trainable weighting matrix W associated with each neuron in the input layer, output layer, and hidden layer of the neural network 600 via the supervised learning engine 390 using the misalignments of the eye model 370 as ground truth output data and the corresponding sets of wavefront sensor data of the eye model 370 as training input data, as described herein. The resulting deep neural network may be stored and used as the composite analysis engine 350. In other embodiments, the controller 140, server 150, and / or distributed server system 154 may be configured to generate a plurality of corrected sets of wavefront sensor data corresponding to a plurality of selected misalignments of the eye model 370 based at least in part on system aberrations associated with the ocular aberrometry system in the aberrometer model 360, prior to forming the neural network 600 and training one or more composite analysis parameters of the neural network 600 using the supervised learning engine 390 to identify the composite analysis engine 350, as described herein.

[0042] At block 408, a compact analysis engine is generated. For example, the controller 140, the server 150, and / or the distributed server system 154 may be configured to form a compact neural network 600 comprising an input layer 620, an output layer 640, and a single hidden layer 630 coupled between the input layer 620 and the output layer 640, and generate a weighting matrix W associated with each neuron of the input, output, and / or hidden layer of the compact neural network 600 based at least in part on one or more composite analysis parameters associated with the multiple hidden layers 630-639 of the composite analysis engine 350. Upon generation, the compact analysis engine 340 may be stored in or otherwise integrated with or implemented by the controller 140, which may be used to generate substantially real-time (e.g., 30 frames / second) user feedback (e.g., display views including various graphics), as well as reliable and accurate monitoring of ocular misalignment, ocular aberrations, and / or other characteristics of the optical target 102, as described herein.

[0043] FIG. 5 illustrates a flow diagram of a process 500 for operating the ocular aberration measurement system 100 and / or 300 according to an embodiment of the present disclosure. It should be appreciated that any step, sub-step, sub-process, or block of the process 500 may be performed in a different order or configuration than the embodiment illustrated by FIG. 5. For example, in other embodiments, one or more blocks may be removed from the process or added thereto. Additionally, block inputs, block outputs, various sensor signals, sensor information, calibration parameters, and / or other operating parameters may be stored in one or more memories before moving to subsequent portions of the corresponding process. Although the process 500 is described with reference to the systems, processes, control loops, and images described with reference to FIGS. 1-3, the process 500 may be performed by other systems that differ from those systems, processes, control loops, and images, for example, including a different selection of electronic devices, sensors, assemblies, mobile structures, and / or mobile structure attributes.

[0044] At block 502, ocular aberration measurement output is received. For example, controller 140, server 150, and / or distributed server system 154 may be configured to receive ocular aberration measurement output data that includes at least wavefront sensor data provided by wavefront sensor 120. More generally, the ocular aberration measurement output data may include any one or more of wavefront sensor data provided by wavefront sensor 120, OCT sensor data provided by OCT sensor 122, eye tracker data provided by eye tracker 130, and / or other output data provided by ocular aberration measurement system 100, as described herein.

[0045] At block 504, an estimated ocular misalignment is determined. For example, the controller 140, the server 150, and / or the distributed server system 154 may be configured to determine an estimated ocular misalignment corresponding to a relative position and / or orientation of the optical target 102 (e.g., a patient's eye) monitored by the ocular aberration measurement system 100 based at least in part on the ocular aberration measurement output data received at block 502. In some embodiments, the controller 140, the server 150, and / or the distributed server system 154 may be configured to determine an estimated ocular misalignment based at least in part on the eye tracker sensor data received at block 502. In other embodiments, the wavefront sensor data received at block 502 includes a time series of wavefront sensor measurements, and the controller 140, the server 150, and / or the distributed server system 154 may be configured to determine an estimated ocular misalignment by determining, for each wavefront sensor measurement, a wavefront estimated ocular misalignment corresponding to the wavefront sensor measurement.

[0046] For example, in one embodiment, the controller 140, the server 150, and / or the distributed server system 154 may be configured to determine the estimated eye position misalignment by: for each wavefront sensor measurement, determining a corresponding estimated relative position and / or orientation of the optical target 102 (e.g., an estimated eye position of the optical target 102); identifying one or more clusters of the estimated relative positions and / or orientations of the optical targets based at least in part on one or more preset or adaptive cluster thresholds; determining a fixation position based at least in part on a geometric center of a largest one of the one or more identified clusters; and for each wavefront sensor measurement, determining an estimated eye position misalignment based at least in part on a difference between the fixation position and the estimated relative position and / or orientation of the optical target corresponding to the wavefront sensor measurement (e.g., the difference between the fixation position and the estimated eye position of the optical target 102 corresponding to a time series of wavefront sensor measurements). In related embodiments in which the ocular aberration measurement system 100 includes an eye tracker 130, the controller 140, the server 150, and / or the distributed server system 154 may be configured to determine, for each wavefront sensor measurement, a corresponding fixation state of the optical target based at least in part on the fixation threshold parameters and the eye tracker sensor data corresponding to the wavefront sensor measurement, and determine an estimated ocular misalignment by filtering out a subset of the wavefront sensor measurements before determining a corresponding estimated relative position and / or orientation of the optical target 102 based at least in part on the identified corresponding fixation state. For example, such a technique may exclude wavefront sensor measurements obtained while the optical target 102 is not detected to be in a fixation state, as detected by the eye tracker 130.

[0047] In block 506, corrected ocular aberration measurement output data is identified. For example, the controller 140, the server 150, and / or the distributed server system 154 may be configured to identify the corrected ocular aberration measurement output data based at least in part on the estimated ocular misalignment identified in block 504 and / or the wavefront sensor data received in block 502. In some embodiments in which the wavefront sensor data comprises a time series of wavefront sensor measurements, the controller 140, the server 150, and / or the distributed server system 154 may be configured to identify the corrected ocular aberration measurement output data by identifying, for each wavefront sensor measurement, an estimated ocular misalignment associated with the wavefront sensor measurement, and generating an average wavefront sensor measurement as the corrected wavefront sensor measurement based at least in part on each wavefront sensor measurement having an associated estimated ocular misalignment that is less than or equal to a preset maximum allowable misalignment. Such a preset maximum allowable misalignment may be selected or set by a manufacturer and / or a user of the ocular aberration measurement system 100.

[0048] In other embodiments, the controller 140, the server 150, and / or the distributed server system 154 may be configured to determine the corrected ocular aberration measurement output data by: determining, for each wavefront sensor measurement, an estimated ocular misalignment associated with the wavefront sensor measurement, for each wavefront sensor measurement having an associated estimated ocular misalignment that is less than or equal to a preset maximum allowable misalignment, determining a corrected wavefront sensor measurement based at least in part on the wavefront sensor measurement and / or the associated estimated ocular misalignment, and generating an average wavefront sensor measurement based at least in part on the corrected wavefront sensor measurement. For example, the controller 140, the server 150, and / or the distributed server system 154 may be configured to determine the corrected wavefront sensor measurement by applying the combined analysis engine 350 or the compact analysis engine 340 of the ocular aberration measurement system 100 to each wavefront sensor measurement to generate a corresponding wavefront estimated ocular misalignment and / or a corrected wavefront sensor measurement, as described herein.

[0049] In block 508, user feedback is generated. For example, the controller 140, the server 150, and / or the distributed server system 154 may be configured to generate user feedback corresponding to the ocular aberration measurement output data received in block 502 based at least in part on the estimated ocular misalignment determined in block 504. In some embodiments, the controller 140, the server 150, and / or the distributed server system 154 may be configured to generate user feedback by determining an ocular misalignment metric based at least in part on the estimated ocular misalignment and reporting the ocular misalignment metric via the user interface 146 of the ocular aberration measurement system 100. In other embodiments, the controller 140, the server 150, and / or the distributed server system 154 may be configured to generate user feedback based at least in part on the estimated ocular misalignment determined in block 504 and / or the corrected ocular aberration measurement output data determined in block 506. In various embodiments, such user feedback may include a substantially real-time display view of ocular misalignment metrics based at least in part on the estimated ocular misalignments identified in block 504, a substantially real-time display view of an ocular aberration map based at least in part on the corrected ocular aberration measurement output data identified in block 504, and / or an audible and / or visual alert indicating that at least one of the estimated ocular misalignments is greater than a pre-set maximum allowable misalignment, as described herein.

[0050] Embodiments of the present disclosure, as described herein, can thus provide substantially real-time (e.g., 30 frames / second) user feedback (e.g., display views including various graphics) and reliable and accurate monitoring of ocular misalignment, ocular aberrations, and / or other characteristics of the optical target 102. Such embodiments may be used to assist in various types of clinical and intra-operative eye examinations and help provide improved surgical outcomes.

[0051] FIG. 6 illustrates a diagram of a multi-layer or “deep” neural network (DNN) 600, according to an embodiment of the present disclosure. In some embodiments, neural network 600 may be representative of a neural network used to implement each of one or more models and / or analytical engines described with respect to systems 100 and / or 300. Neural network 600 processes input data 610 using an input layer 620. In various embodiments, input data 610 may correspond to aberrometry output data and / or training data provided to one or more models and / or analytical engines to generate and / or train one or more models and / or analytical models, as described herein. In some embodiments, input layer 620 may include multiple neurons or nodes that are used to condition input data 610 by scaling, biasing, filtering, range limiting, and / or otherwise conditioning input data 610 for processing by the remainder of neural network 600. In other embodiments, the input layer 620 may be configured to echo the input data 610 (e.g., if the input data 610 has already been appropriately scaled, biased, filtered, range limited, and / or otherwise adjusted). Each neuron in the input layer 620 generates an output that is provided to a neuron / node in the hidden layer 630. The hidden layer 630 includes a number of neurons / nodes that process the output from the input layer 620. In some implementations, each of the neurons in the hidden layer 630 generates an output that is then propagated through one or more additional hidden layers terminating in the hidden layer 639. The hidden layer 639 includes a number of neurons / nodes that process the output from the previous hidden layer. In the embodiment shown in FIG. 6, the output of the hidden layer 639 is provided to the output layer 640. In various embodiments, the output layer 640 includes one or more neurons / nodes that may be used to condition the output from the hidden layer 639 by scaling, biasing, filtering, range limiting, and / or otherwise conditioning the output from the hidden layer 639 to form the output data 650.In alternative embodiments, neural network 600 may be implemented according to different neural network or other processing architectures, including neural networks with only one hidden layer, neural networks with recurrent layers, and / or various other neural network architectures or composite analysis methodologies, including K-nearest neighbor (k-NN) databases for classification and / or regression.

[0052] In some implementations, the input layer 620, the hidden layers 631-639, and / or the output layer 640 each include one or more neurons. In one implementation, the input layer 620, the hidden layers 631-639, and / or the output layer 640 each may include the same or different numbers of neurons. In certain embodiments, the neural network 600 may include a total of about six layers with up to 2-4000 neurons in each layer. In various implementations, each such constituent neuron may be configured to receive a combination of its inputs x (e.g., a weighted sum generated using a trainable weighting matrix / vector W), an optional trainable bias b, and apply an activation function f to generate an output a, such that the combination follows the formula a=f(Wx+b). The activation function f may be implemented, for example, as a modified linear unit activation function, or as any one or combination of activation functions with upper and / or lower bounds, a log-sigmoid function, a hyperbolic tangent function, and / or according to other activation function forms. Each neuron in such a network may be configured to operate according to the same or different activation functions and / or different types of activation functions, as described herein. In certain embodiments corresponding to regression applications, only neurons in the output layer 640 may be configured to apply such linear activation functions to generate their respective outputs.

[0053] In various embodiments, the neural network 600 may be trained using supervised learning (e.g., implemented as a supervised learning engine 390), such as by systematically providing selected sets of training data (e.g., training data 392) to the neural network 600, where each set of training data includes a set of input training data and a corresponding set of ground truth (e.g., expected) output data (e.g., a combination of the aberrometer model 360 and the eye model 370), and then identifying and / or otherwise comparing differences between the resulting output data 650 (e.g., training output data provided by the neural network 600) and the ground truth output data (e.g., a "training error"). In some embodiments, the training error may be fed back to the neural network 600 to adjust various trainable weights, biases, and / or other complex analysis parameters of the neural network 600. In some embodiments, such training errors may be provided as feedback to neural network 600 using one or various backpropagation methods, including, for example, stochastic gradient descent and / or other backpropagation methods. In one or more examples, relatively large groups of selected sets of training data may be presented to neural network 600 multiple times until an overall loss function (e.g., mean squared error based on the difference between each set of training data) converges to or below a preset maximum acceptable loss threshold.

[0054] In additional embodiments, the supervised learning engine 390 may be configured to include semi-supervised learning, weakly supervised learning, active learning, structured prediction, and / or other generalized machine learning methods to aid in training the composite analysis parameters of the neural network 600 (e.g., generating the composite analysis engine 350) as described herein. For example, the supervised learning engine 390 may be configured to generate simulated sets of training data, each set including simulated input training data and a corresponding set of simulated ground truth output data, and to perform supervised learning based at least in part on the simulated sets of training data. Each of the simulated sets of input training data and ground truth data may be generated by modifying the input training data (e.g., to adjust aberration parameters and / or misalignment associated with the simulated double-path model target) and interpolating the unsimulated ground truth data to generate the corresponding simulated ground truth data. In various embodiments, the supervised learning engine 390 may be configured to simulate one or millions of sets of simulated training data, each of which deviates at least slightly from a set of training data corresponding to the eye model 370, and train the neural network 600 according to the one or millions of such sets of simulated training data.

[0055] Where applicable, various embodiments provided by the present disclosure may be implemented using hardware, software, or a combination of hardware and software. Also, where applicable, various hardware and / or software components described herein may be combined into composite components comprising software, hardware, and / or both without departing from the spirit of the present disclosure. Where applicable, various hardware and / or software components described herein may be separated into sub-components comprising software, hardware, or both without departing from the spirit of the present disclosure. Furthermore, where applicable, it is contemplated that software components may be implemented as hardware components and vice versa.

[0056] Software according to the present disclosure, such as non-transitory instructions, program code, and / or data, can be stored on one or more non-transitory machine-readable media. It is also contemplated that the software specified herein can be implemented using one or more networked and / or other general-purpose or special-purpose computers and / or computer systems. Where applicable, the order of the various steps described herein can be changed, combined into composite steps, and / or separated into sub-steps to provide the features described herein.

[0057] The above described embodiments are illustrative of the present invention, but are not intended to be limiting, and it will be appreciated that many modifications and variations are possible in accordance with the principles of the present invention. Accordingly, the scope of the present invention is defined solely by the following claims.

Claims

1. 1. An ocular aberration measurement system, comprising: a wavefront sensor configured to provide wavefront sensor data related to an optical target monitored by the ocular aberration measurement system; a logic device configured to communicate with the wavefront sensor, determining a combined analysis engine for the ocular aberrometry system based on an aberrometer model and an eye model associated with the ocular aberrometry system, the aberrometer model and the eye model being determined based on wavefront sensor data provided by the wavefront sensor; and a logic device configured to generate a compact analytical engine for the ocular aberration measurement system based on the identified composite analytical engine; The logic device is generating the aberrometer model associated with the ocular aberration measurement system; configured to identify the composite analysis engine based on the generated aberrometer model; Identifying the composite analysis engine includes: generating a plurality of corrected sets of wavefront sensor data corresponding to a plurality of selected misalignments of the eye model based on system aberrations associated with the ocular aberration measurement system in the aberrometer model; forming a neural network comprising an input layer of neurons, an output layer of neurons, and a plurality of hidden layers of neurons coupled between the input layer and the output layer; training at least one trainable weighting matrix associated with each neuron in the input layer, the output layer, and the hidden layer of the neural network via a supervised learning engine using the selected misalignments of the eye model as ground truth output data and a corresponding plurality of corrected sets of the wavefront sensor data of the eye model as training input data.

2. The logic device is generating the eye model associated with the ocular aberration measurement system; and determining the composite analysis engine based on the generated eye model.

2. An ocular aberration measurement system according to claim 1.

3. Generating the eye model comprises: illuminating at least the wavefront sensor of the ocular aberration measurement system via a double-pass model target positioned as the optical target monitored by the ocular aberration measurement system; Varying a position and / or orientation of the exchangeable ocular aberration model of the double pass model target relative to an optical axis of the ocular aberration measurement system according to a plurality of selected positional displacements to generate a corresponding plurality of sets of wavefront sensor data provided by the wavefront sensor; and storing the selected misalignments and corresponding sets of wavefront sensor data as the eye model.

3. An ocular aberration measurement system according to claim 2.

4. The double pass model target is said interchangeable ocular aberration model; a six degree of freedom motion actuator removably coupled to the interchangeable ocular aberration model and configured to change the position and / or orientation of the interchangeable ocular aberration model to generate the selected misalignments and corresponding sets of wavefront sensor data.

4. An ocular aberration measurement system according to claim 3.

5. Identifying the composite analysis engine includes: forming a neural network comprising an input layer of neurons, an output layer of neurons, and at least one hidden layer of neurons coupled between the input layer and the output layer; training at least one trainable weighting matrix associated with each neuron in the input layer, the output layer, and the hidden layer of the neural network via a supervised learning engine using the misalignments of the eye model as ground truth output data and a corresponding set of wavefront sensor data of the eye model as training input data.

3. An ocular aberration measurement system according to claim 2.

6. Generating the aberrometer model includes: illuminating at least the wavefront sensor of the ocular aberration measurement system via a single-pass model target positioned as the optical target monitored by the ocular aberration measurement system; Varying the defocus aberration of the single-pass model target according to a plurality of selected defocus optical powers to generate a plurality of sets of wavefront sensor data provided by the wavefront sensor; determining system aberrations associated with the ocular aberration measurement system based on a plurality of sets of wavefront sensor data; storing a plurality of sets of the system aberrations and / or the wavefront sensor data as the aberrometer model.

2. An ocular aberration measurement system according to claim 1.

7. The single pass model target is a plane wave retinal source comprising a laser or a superluminescent diode configured to generate a source beam oriented along an optical axis of the ocular aberration measurement system; a lens system coupled to a linear motion actuator configured to vary the defocus aberration of the single-pass model target to generate a plurality of reference interferograms and corresponding wavefront sensor data.

7. An ocular aberration measurement system according to claim 6.

8. generating the compact parsing engine forming a compact neural network comprising an input layer of neurons, an output layer of neurons, and at least one hidden layer of neurons coupled between the input layer and the output layer; generating a weighting matrix associated with each neuron of the input layer, the output layer, and / or the hidden layer of the compact neural network based on one or more composite analysis parameters associated with a plurality of hidden layers of the composite analysis engine; 2. An ocular aberration measurement system according to claim 1.

9. 1. A method comprising: identifying a combined analysis engine for an ocular aberrometry system based on an aberrometer model and an eye model associated with the ocular aberrometry system, the aberrometer model and the eye model being based on wavefront sensor data provided by a wavefront sensor of the ocular aberrometry system; generating a compact analytical engine for the ocular aberration measurement system based on the identified composite analytical engine; Furthermore, generating the aberrometer model associated with the ocular aberration measurement system; identifying the composite analytical engine based on the generated aberrometer model; Identifying the composite analysis engine includes: generating a plurality of corrected sets of wavefront sensor data corresponding to a plurality of selected misalignments of the eye model based on system aberrations associated with the ocular aberration measurement system in the aberrometer model; forming a neural network comprising an input layer of neurons, an output layer of neurons, and a plurality of hidden layers of neurons coupled between the input layer and the output layer; training at least one trainable weighting matrix associated with each neuron in the input layer, the output layer, and the hidden layer of the neural network via a supervised learning engine using the selected misalignments of the eye model as ground truth output data and a corresponding plurality of corrected sets of the wavefront sensor data of the eye model as training input data.

10. Furthermore, generating the eye model associated with the ocular aberration measurement system; and identifying the composite analysis engine based on the generated eye model.

10. The method of claim 9.

11. Generating the eye model comprises: illuminating at least the wavefront sensor of the ocular aberration measurement system via a double-pass model target positioned as an optical target monitored by the ocular aberration measurement system; Varying a position and / or orientation of the exchangeable ocular aberration model of the double pass model target relative to an optical axis of the ocular aberration measurement system according to a plurality of selected positional displacements to generate a corresponding plurality of sets of wavefront sensor data provided by the wavefront sensor; and storing the selected misalignments and corresponding sets of wavefront sensor data as the eye model. The method of claim 10.

12. The double pass model target is the interchangeable ocular aberration model; a six degree of freedom motion actuator removably coupled to the interchangeable ocular aberration model and configured to change the position and / or orientation of the interchangeable ocular aberration model to generate the selected misalignments and corresponding sets of wavefront sensor data. The method of claim 11.

13. Identifying the composite analysis engine includes: forming a neural network comprising an input layer of neurons, an output layer of neurons, and at least one hidden layer of neurons coupled between the input layer and the output layer; training at least one trainable weighting matrix associated with each neuron in the input layer, the output layer, and the hidden layer of the neural network via a supervised learning engine using the misalignments of the eye model as ground truth output data and a corresponding set of wavefront sensor data of the eye model as training input data. The method of claim 10.

14. Generating the aberrometer model includes: illuminating at least the wavefront sensor of the ocular aberration measurement system via a single-pass model target positioned as an optical target monitored by the ocular aberration measurement system; Varying the defocus aberration of the single-pass model target according to a plurality of selected defocus optical powers to generate a plurality of sets of wavefront sensor data provided by the wavefront sensor; determining system aberrations associated with the ocular aberration measurement system based on a plurality of sets of wavefront sensor data; storing a plurality of sets of the system aberrations and / or the wavefront sensor data as the aberrometer model.

10. The method of claim 9.

15. The single pass model target is a plane wave retinal source comprising a laser or a superluminescent diode configured to generate a source beam oriented along an optical axis of the ocular aberration measurement system; a lens system coupled to a linear motion actuator configured to vary the defocus aberration of the single-pass model target to generate a plurality of reference interferograms and corresponding wavefront sensor data. The method of claim 14.

16. generating the compact parsing engine forming a compact neural network comprising an input layer of neurons, an output layer of neurons, and at least one hidden layer of neurons coupled between the input layer and the output layer; generating a weighting matrix associated with each neuron of the input layer, the output layer, and / or the hidden layer of the compact neural network based on one or more composite analysis parameters associated with a plurality of hidden layers of the composite analysis engine; 10. The method of claim 9.

Citation Information

Patent Citations

  • Measuring device for wavefront aberration

    JP1994235619A

  • Refractive power measuring apparatus

    JP2005103103A

  • Apparatus and associated method for calibrating wavefront sensors

    JP2005512627A

  • Eye refracting power measuring apparatus

    JP2007007277A

  • System and method for analyzing wavefront aberration

    JP2007531559A