Using a video-based eye tracker to analyse the binocular near-reflex dynamics response.

PURPOSE: This study presents a novel video-based eye-tracking system for analysing the dynamics of the binocular near-reflex response. The system enables the simultaneous measurement of convergence, divergence and pupillary size during accommodation and disaccommodation to aid the comprehensive understanding of the three-component near-reflex. METHODS: A high-speed (90 Hz) video-based eye tracker was used to capture changes in eye gaze and pupil radius in 15 participants in response to altering stimulus conditions. An offline analysis involved separating the gaze vector components and pupil radius, which were fitted to a hyperbolic tangent function to characterise the dynamics of the near-reflex process. RESULTS: Significant differences in the temporal parameters of the pupil radius were observed between the near-to-far and far-to-near vision changes, with faster miosis compared with mydriasis. Additionally, differences in response times were found between gaze angle components, with longer convergence times compared to changes in the vertical direction (saccades). The steady-state values of the gaze components and pupil radius were in line with theoretical expectations and previous reports. CONCLUSIONS: The proposed system provides a non-invasive, portable and cost-effective method for evaluating near-reflex dynamics under natural viewing conditions using a video-based eye tracker. The sampling rate ensures the accurate assessment of vergence eye movements and pupillary dynamics. By simultaneously measuring eye convergence, divergence and pupil size, the system offers a more comprehensive assessment of the near-reflex response. This makes it a valuable tool for clinical diagnosis, research studies and investigating the effects of near work on the visual system.

Prediction of Subjective Refraction From Anterior Corneal Surface, Eye Lengths, and Age Using Machine Learning Algorithms.

Purpose: To develop a machine learning regression model of subjective refractive prescription from minimum ocular biometry and corneal topography features. Methods: Anterior corneal surface parameters (Zernike coefficients and keratometry), axial length, anterior chamber depth, and age were posed as features to predict subjective refractions. Measurements from 355 eyes were split into training (75%) and test (25%) sets. Different machine learning regression algorithms were trained by 10-fold cross-validation, optimized, and tested. A neighborhood component analysis provided features' normalized weights in predictions. Results: Gaussian process regression algorithms provided the best models with mean absolute errors of around 1.00 diopters (D) in the spherical component and 0.15 D in the astigmatic components. Conclusions: The normalized weights showed that subjective refraction can be predicted by only keratometry, age, and axial length. Increasing the topographic description detail of the anterior corneal surface implied by a high-order Zernike decomposition versus adjustment to a spherocylindrical surface is not reflected as improved subjective refraction prediction, which is poor, mainly in the spherical component. However, the highest achievable accuracy differs by only 0.75 D from that of other works with a more exhaustive eye refractive elements description. Although the chosen parameters may have not been the most efficient, applying machine learning and big data to predict subjective refraction can be risky and impractical when evaluating a particular subject at statistical extremes. Translational Relevance: This work evaluates subjective refraction prediction by machine learning from the anterior corneal surface and ocular biometry. It shows the minimum biometric information required and the highest achievable accuracy. Objetivo: El desarrollo de un modelo de regresión de aprendizaje automático prescripción refractiva subjetiva a partir de las características mínimas de la biometría ocular y la superficie corneal. Métodos: Los parámetros de la superficie corneal anterior (coeficientes de Zernike y queratometría), además de longitudes axiales y de cámara anterior, edades y las refracciones subjetivas no ciclopléjicas de 355 ojos se dividieron en un conjunto de entrenamiento (75%) y otro de test (25%) y se entrenaron diferentes algoritmos de regresión de aprendizaje automático mediante validación cruzada 10 veces, se optimizaron y se probaron sobre el conjunto test. Resultados: Los algoritmos de regresión del proceso gaussiano proporcionaron los mejores modelos con un error absoluto medio fue de alrededor de 1.00 D en el componente esférico y de 0.25 D en los componentes astigmáticos. Conclusiones: Los pesos normalizados mostraron que la refracción subjetiva puede predecirse utilizando únicamente la queratometría, la edad y la longitud axial como características. El aumento del detalle de la descripción topográfica de la superficie corneal anterior que supone una descomposición de Zernike de alto orden frente al ajuste a una superficie esferocilíndrica realizado por queratometría no se refleja en una mejora de la predicción de la refracción subjetiva, que es pobre, en cualquier caso, principalmente en el componente esférico. Sin embargo, la máxima precisión alcanzada difiere en sólo 0,75 D de la de otros trabajos con una descripción más exhaustiva de los elementos refractivos del ojo. De todos modos, el aprendizaje automático y los datos masivos aplicados a la predicción de la refracción subjetiva pueden ser arriesgados y poco prácticos cuando se evalúa a un sujeto concreto en los extremos estadísticos, aunque los parámetros elegidos puedan no haber sido los más ineficaces. Relevancia Traslativa: El trabajo evalúa la predicción de la refracción subjetiva mediante aprendizaje automático a partir de la superficie corneal anterior y la biometría ocular, mostrando la mínima información biométrica requerida y la máxima precisión alcanzable.

Comparative analysis of spontaneous blinking and the corneal reflex.

Ocular surface health, the cognitive status, psychological health or human neurological disorders, among others, can be assessed by studying eye blinking, which can be differentiated in spontaneous, reflex and voluntary. Its diagnostic potential has provided a great number of works that evaluate their characteristics and variations depending on the subject's condition (sex, tiredness, health, …). The objective of this study was to analyse the differences in blinking kinematics of spontaneous and reflex blinks, distinguishing between direct and consensual reflexes, using a self-developed, non-invasive and image processing-based method. A video-oculography system is proposed using an air jet driven by a syringe to induce reflex and a high-speed camera to record the blinking of both eyes. The light intensity diffused by the eye changes during blinking and peaks when the eyelid closes. Sixty-second sequences were recorded of 25 subjects blinking. Intensity curves were off-line fitted to an exponentially modified Gaussian (EMG) function, whose σ, µ and τ parameters were analysed. A two-way analysis of variance (ANOVA) of these parameters was conducted to test the influence of the subject, the eye and blink type. In the closing phase, direct and consensual corneal reflexes are faster than spontaneous blinking, but there was no significant difference between them, nor between right and left eyes. In the opening phase, the direct corneal reflex was the slowest and significant differences appeared between right and left eyes.

Corneal Stability following Hyperopic LASIK with Advanced Laser Ablation Profiles Analyzed by a Light Propagation Study.

PURPOSE: To assess anterior corneal surface stability 12 months following hyperopic LASIK correction with a light propagation algorithm. SETTING: Vissum Instituto Oftalmológico de Alicante, Universidad Miguel Hernández, Alicante, Spain. METHODS: This retrospective consecutive observational study includes 37 eyes of 37 patients treated with 6th-generation excimer laser platform (Schwind Amaris). Hyperopic LASIK was performed in all of them by the same surgeon (JLA) and completed 12-month follow-up. Corneal topography was analyzed with a light propagation algorithm, to assess the stability of the corneal outcomes along one year of follow-up. RESULTS: Between three and twelve months postoperatively, an objective corneal power (OCP) regression of 0.39 D and 0.41 D was found for 6 mm and 9 mm central corneal zone, respectively. Subjective outcomes at the end of the follow-up period were as follows: 65% of eyes had spherical equivalent within ±0.50 D. 70% of eyes had an uncorrected distance visual acuity 20/20 or better. 86% of eyes had the same or better corrected distance visual acuity. In terms of stability, 0.14 D of regression was found. No statistically significant differences were found for all the study parameters evaluated at different postoperative moments over the 12-month period. CONCLUSIONS: Light propagation analysis confirms corneal surface stability following modern hyperopic LASIK with a 6th-generation excimer laser technology over a 12-month period.

Blinking characterization from high speed video records. Application to biometric authentication.

The evaluation of eye blinking has been used for the diagnosis of neurological disorders and fatigue. Despite the extensive literature, no objective method has been found to analyze its kinematic and dynamic behavior. A non-contact technique based on the high-speed recording of the light reflected by the eyelid in the blinking process and the off-line processing of the sequence is presented. It allows for objectively determining the start and end of a blink, besides obtaining different physical magnitudes: position, speed, eyelid acceleration as well as the power, work and mechanical impulse developed by the muscles involved in the physiological process. The parameters derived from these magnitudes provide a unique set of features that can be used to biometric authentication. This possibility has been tested with a limited number of subjects with a correct identification rate of up to 99.7%, thus showing the potential application of the method.

Realistic limits for subpixel movement detection.

Object tracking with subpixel accuracy is of fundamental importance in many fields since it provides optimal performance at relatively low cost. Although there are many theoretical proposals that lead to resolution increments of several orders of magnitude, in practice this resolution is limited by the imaging systems. In this paper we propose and demonstrate through simple numerical models a realistic limit for subpixel accuracy. The final result is that maximum achievable resolution enhancement is connected with the dynamic range of the image, i.e., the detection limit is 1/2∧(nr.bits). The results here presented may aid in proper design of superresolution experiments in microscopy, surveillance, defense, and other fields.

Method for targetless tracking subpixel in-plane movements.

We present a targetless motion tracking method for detecting planar movements with subpixel accuracy. This method is based on the computation and tracking of the intersection of two nonparallel straight-line segments in the image of a moving object in a scene. The method is simple and easy to implement because no complex structures have to be detected. It has been tested and validated using a lab experiment consisting of a vibrating object that was recorded with a high-speed camera working at 1000 fps. We managed to track displacements with an accuracy of hundredths of pixel or even of thousandths of pixel in the case of tracking harmonic vibrations. The method is widely applicable because it can be used for distance measuring amplitude and frequency of vibrations with a vision system.

A high-resolution binocular video-oculography system: assessment of pupillary light reflex and detection of an early incomplete blink and an upward eye movement.

BACKGROUND: The pupillary light reflex characterizes the direct and consensual response of the eye to the perceived brightness of a stimulus. It has been used as indicator of both neurological and optic nerve pathologies. As with other eye reflexes, this reflex constitutes an almost instantaneous movement and is linked to activation of the same midbrain area. The latency of the pupillary light reflex is around 200 ms, although the literature also indicates that the fastest eye reflexes last 20 ms. Therefore, a system with sufficiently high spatial and temporal resolutions is required for accurate assessment. In this study, we analyzed the pupillary light reflex to determine whether any small discrepancy exists between the direct and consensual responses, and to ascertain whether any other eye reflex occurs before the pupillary light reflex. METHODS: We constructed a binocular video-oculography system two high-speed cameras that simultaneously focused on both eyes. This was then employed to assess the direct and consensual responses of each eye using our own algorithm based on Circular Hough Transform to detect and track the pupil. Time parameters describing the pupillary light reflex were obtained from the radius time-variation. Eight healthy subjects (4 women, 4 men, aged 24-45) participated in this experiment. RESULTS: Our system, which has a resolution of 15 microns and 4 ms, obtained time parameters describing the pupillary light reflex that were similar to those reported in previous studies, with no significant differences between direct and consensual reflexes. Moreover, it revealed an incomplete reflex blink and an upward eye movement at around 100 ms that may correspond to Bell's phenomenon. CONCLUSIONS: Direct and consensual pupillary responses do not any significant temporal differences. The system and method described here could prove useful for further assessment of pupillary and blink reflexes. The resolution obtained revealed the existence reported here of an early incomplete blink and an upward eye movement.

Vibration frequency measurement using a local multithreshold technique.

In this paper, we demonstrate the use of a video camera for measuring the frequency of small-amplitude vibration movements. The method is based on image acquisition and multilevel thresholding and it only requires a video camera with high enough acquisition rate, not being necessary the use of targets or auxiliary laser beams. Our proposal is accurate and robust. We demonstrate the technique with a pocket camera recording low-resolution videos with AVI-JPEG compression and measuring different objects that vibrate in parallel or perpendicular direction to the optical sensor. Despite the low resolution and the noise, we are able to measure the main vibration modes of a tuning fork, a loudspeaker and a bridge. Results are successfully compared with design parameters and measurements with alternative devices.

Image processing for safety assessment in civil engineering.

Behavior analysis of construction safety systems is of fundamental importance to avoid accidental injuries. Traditionally, measurements of dynamic actions in civil engineering have been done through accelerometers, but high-speed cameras and image processing techniques can play an important role in this area. Here, we propose using morphological image filtering and Hough transform on high-speed video sequence as tools for dynamic measurements on that field. The presented method is applied to obtain the trajectory and acceleration of a cylindrical ballast falling from a building and trapped by a thread net. Results show that safety recommendations given in construction codes can be potentially dangerous for workers.

Open-access operating algorithms for commercial videokeratographer and improvement of corneal sampling.

We present an algorithm to process images of reflected Placido rings captured by a commercial videokeratoscope. Raw data are obtained with no Cartesian-to-polar-coordinate conversion, thus avoiding interpolation and associated numerical artifacts. The method provides a characteristic equation for the device and is able to process around 6 times more corneal data than the commercial software. Our proposal allows complete control over the whole process from the capture of corneal images until the computation of curvature radii.

Resolution limits to object tracking with subpixel accuracy.

Subpixel methods increase the accuracy and efficiency of image detectors, processing units, and algorithms and provide very cost-effective systems for object tracking. Published methods achieve resolution increases up to three orders of magnitude. In this Letter, we demonstrate that this limit can be theoretically improved by several orders of magnitude, permitting micropixel and submicropixel accuracies. The necessary condition for movement detection is that one single pixel changes its status. We show that an appropriate target design increases the probability of a pixel change for arbitrarily small shifts, thus increasing the detection accuracy of a tracking system. The proposal does not impose severe restriction on the target nor on the sensor, thus allowing easy experimental implementation.

Measurement of wide frequency range structural microvibrations with a pocket digital camera and sub-pixel techniques.

Analysis of vibrations and displacements is a hot topic in structural engineering. Although there is a wide variety of methods for vibration analysis, direct measurement of displacements in the mid and high frequency range is not well solved and accurate devices tend to be very expensive. Low-cost systems can be achieved by applying adequate image processing algorithms. In this paper, we propose the use of a commercial pocket digital camera, which is able to register more than 420 frames per second (fps) at low resolution, for accurate measuring of small vibrations and displacements. The method is based on tracking elliptical targets with sub-pixel accuracy. Our proposal is demonstrated at a 10 m distance with a spatial resolution of 0.15 mm. A practical application over a simple structure is given, and the main parameters of an attenuated movement of a steel column after an impulsive impact are determined with a spatial accuracy of 4 µm.

Noninvasive measurement of eye retraction during blinking.

We present a noninvasive technique for high-speed measuring of eye retraction and eyelid position during blinking. The anterior chamber of the eye is illuminated by the slit lamp of a biomicroscope and eye dynamics during a blinking sequence are captured with a high-speed camera working at 500 frames per second. Digital image processing allows quantitative analysis of cornea and eyelid positions during the closing and opening phases of the blinking process. Our method allows simultaneous measuring of corneal retraction, duration of down and up phases, total blinking duration, and average and peak speeds of the eyelids in both phases, thus providing a complete analysis of the blink's transversal motions.

Optical surface reconstruction technique through combination of zonal and modal fitting.

Videokeratometers and Scheimpflug cameras permit accurate estimation of corneal surfaces. From height data it is possible to adjust analytical surfaces that will be later used for aberration calculation. Zernike polynomials are often used as adjusting polynomials, but they have shown to be not precise when describing highly irregular surfaces. We propose a combined zonal and modal method that allows an accurate reconstruction of corneal surfaces from height data, diminishing the influence of smooth areas over irregular zones and vice versa. The surface fitting error is decreased in the considered cases, mainly in the central region, which is more important optically. Therefore, the method can be established as an accurate resampling technique.

Corneal primary aberrations compensation by oblique light incidence.

The eye is not a centered system. The line of sight connects the fovea with the center of the pupil and is usually tilted in the temporal direction. Thus, off-axis optical aberrations, mainly coma and oblique astigmatism, are introduced at the fovea. Tabernero et al. [J. Opt. Soc. Am. A 24(10), 3274-3283 (2007)] showed that a horizontal tilt of the crystalline lens generates a horizontal coma aberration that is compensated by the oblique light incidence on the eye. Here we suggest that corneal astigmatism may also play a role in compensation of oblique aberrations, and we propose a simple model to analyze such a possibility. A theoretical Kooijman eye model with a slight ( approximately 0.6 D) with-the-rule astigmatism is analyzed. Light rays at different incidence angles to the optical axis are considered, and the corresponding point spread functions (PSFs) at the retina are calculated. A quality criterion is used to determine the incidence angle that provides the narrowest and highest PSF energy peak. We show that the best image is obtained for a tilted incidence angle compatible with mean values of the angle kappa. This suggests that angle kappa, lens tilt, and corneal astigmatism may combine to provide a passive compensation mechanism to minimize aberrations on the fovea.

Correlation between the dioptric power, astigmatism and surface shape of the anterior and posterior corneal surfaces.

A knowledge of the shape of the cornea is of major importance for the planning and monitoring of surgery, and for the correct diagnosis of corneal diseases. Many authors have studied the geometry of the second corneal surface in the central region and it has been stated that there is a high correlation between the central radii of curvature and asphericities of the two corneal surfaces. In this work we extend this study to a larger, central, 6 mm diameter of the cornea. Surface height data, obtained with an Oculus Pentacam from 42 eyes of 21 subjects, were analysed to yield surface power vectors. Corneal heights of both surfaces were also decomposed into low-order Zernike polynomials and the correlations between each of the power vectors and low-order Zernike coefficients for the two surfaces were studied. There was not only a strong correlation between spherical powers and Zernike defocus coefficients, but also between the astigmatic components. The correspondence between the astigmatism in both surfaces found here can be of the utmost importance in planning optical surgery, since perfect spherical ablation of the first surface does not assure total correction of corneal astigmatism.

Adaptive sampling in convergent beams.

Numerical calculation of convergent Fresnel patterns through fast Fourier transform usually requires a large number of samples to fulfill the Nyquist sampling condition around the focus. From polynomial decomposition of the wavefront it is possible to determine which polynomial orders are the main contributors to the number of samples. This information can be used to properly modify the initial wavefront and relax the Nyquist condition thus giving a more efficient numerical algorithm.