Monte Carlo modelling of the spectral reflectance of the human eye

Learning to Recover Spectral Reflectance From RGB Images

This paper tackles spectral reflectance recovery (SRR) from RGB images. Since capturing ground-truth spectral reflectance and camera spectral sensitivity are challenging and costly, most existing approaches are trained on synthetic images and utilize the same parameters for all unseen testing images, which are suboptimal especially when the trained models are tested on real images because they never exploit the internal information of the testing images. To address this issue, we adopt a self-supervised meta-auxiliary learning (MAXL) strategy that fine-tunes the well-trained network parameters with each testing image to combine external with internal information. To the best of our knowledge, this is the first work that successfully adapts the MAXL strategy to this problem. Instead of relying on naive end-to-end training, we also propose a novel architecture that integrates the physical relationship between the spectral reflectance and the corresponding RGB images into the network based on our mathematical analysis. Besides, since the spectral reflectance of a scene is independent to its illumination while the corresponding RGB images are not, we recover the spectral reflectance of a scene from its RGB images captured under multiple illuminations to further reduce the unknown. Qualitative and quantitative evaluations demonstrate the effectiveness of our proposed network and of the MAXL. Our code and data are available at https://github.com/Dong-Huo/SRR-MAXL.

Sensitivity of visible range multi-wavelength algorithms for retinal tissue oximetry to acquisition parameters

This study examined the sensitivity of broadband spectroscopy algorithms for retinal tissue oximetry to spectral acquisition parameters. Monte Carlo simulations were conducted on a 4-layer retinal model to assess the impact of various parameters. The optimal spectral range for accurate measurements was determined to be 530 nm to 585 nm. Decreased spectral resolution below 4 nm significantly reduced accuracy. Using an acquisition area larger than the blood vessel resulted in an underestimation of oxygen saturation, especially for high values. A threshold was observed where increased light intensity had no significant impact on measurement variability. The study highlights the importance of informed parameter selection for accurately assessing retinal microcapillary oxygenation and studying local hemodynamics.

Intraocular reflectance of the ocular fundus and its impact on increased retinal hazard

PURPOSE

Inside the eye light can be reflected multiple times due to light-tissue interactions and the spherical geometry of the eye. Due to these optical properties, a defined retinal area is not only illuminated by direct light but also by indirect, reflected light from the inner side of the eyewall. During illumination for ophthalmic surgery, this could lead to an unintended increase in intraocular retinal irradiance, which was already discussed in previous studies but without a detailed consideration of spectral differences and a potential influence of pigmentation. In this study this effect is investigated wavelength-dependent to see if different wavelengths lead to different increase in irradiance, with a special focus on the raise in photochemical and thermal hazard to the retina. It is also examined whether this effect is dependent on the pigmentation of the eye.

METHODS

The reflectance properties of either less or highly pigmented porcine eyes are measured in the wavelength range between 350 and 1100nm with an integrating sphere and a spectrometer. With these reflectance spectra the wavelength-dependent Sphere Multiplier M of porcine eyes can be calculated, which represents the increase of radiance due to multiple reflections inside a sphere compared to a planar diffuser of the same size. Based on measurements of the emitted irradiance of ophthalmic illumination fibers the increase in photochemical and thermal retinal hazard due to these multiple reflections is calculated for eyes with small and high amounts of pigmentation.

RESULTS

The reflectance of the inner eyewall in the range between 350 and 1100nm is significantly higher for eyes with low pigmentation (between 4.90% and 37.44% reflectance) in comparison to eyes with a high amount of pigmentation (between 4.30% and 28.88% reflectance). The Sphere Multiplier for the inner side of the eyewall (sclera, choroid and retina) ranges between 1.13 and 1.59 and between 1.13 and 1.48 for eyes with low and high pigmentation, respectively, in the range between 350 and 1100nm. The reflectance, as well as the Sphere Multiplier, is strongly wavelength-dependent due to the absorption spectra of melanin and hemoglobin, which are located in the eye. With increasing wavelength, the reflection properties and the Sphere Multiplier also increases. With this, the photochemical retinal hazard of highly pigmented eyes increases by (14.11± 0.09)% and of lightly pigmented eyes by (16.75±0.35)% compared to if the reflection properties are not considered. The thermal retinal hazard increases by (14.30±0.07)% for highly pigmented eyes and by (19.65±0.17)% for low pigmented eyes.

CONCLUSION

This study demonstrates that the anatomy and pigmentation of the eye plays an important role for the reflectance properties of the eye and for the photochemical and thermal hazard to the retina.

Monte-Carlo simulation and tissue-phantom model for validation of ocular oximetry

Ocular oximetry, in which blood oxygen saturation is evaluated in retinal tissues, is a promising technique for the prevention, diagnosis and management of many diseases and conditions. However, the development of new tools for evaluating oxygen saturation in the eye fundus has often been limited by the lack of reference tools or techniques for such measurements. In this study, we describe a two-step validation method. The impact of scattering, blood volume fraction and lens yellowing on the oximetry model is investigated using a tissue phantom, while a Monte Carlo model of the light propagation in the eye fundus is used to study the effect of the fundus layered-structure. With this method, we were able to assess the performance of an ocular oximetry technique in the presence of confounding factors and to quantify the impact of the choroidal circulation on the accuracy of the measurements. The presented strategy will be useful to anyone involved in studies based on the eye fundus diffuse reflectance.

Determination of Micropulse Modes with Targeted Damage to the Retinal Pigment Epithelium Using Computer Modeling for the Development of Selective Individual Micropulse Retinal Therapy

PURPOSE

When using a serial laser system for selective impact on the retinal pigment epithelium (RPE), there is a challenge to determine the optimal range of micropulse parameters which result in targeted damage to the RPE. This study proposes a computer model that has identified the optimal parameters to be applied.

METHODS

This study was conducted on 18 patients who were diagnosed with acute central serous chorioretinopathy and transparent optical media, aged 35 to 46 years old, and type 2 and 3 on the Fitzpatrick scale. Testing of the micropulse mode was performed on the Navilas 577s laser system; 864 spots were analyzed in total. Considering the probability of damage visualization at different laser power, the computer simulation of tissue heating and protein denaturation was performed to determine the micropulse modes which resulted in selective damage to the RPE.

RESULTS

The computer model parameter ΔE = 3.34 × 10⁵ J/mol was determined from fitting the model predictions to the autofluorescence test results. The micropulse modes with a micropulse duration of 50-100 µs, duty cycle 2.4-4.8%, 10 ms-pulse envelope (5 micropulses), and spot diameter of 100 µm have efficiency and selectivity above 67% and correspond to the optimal therapeutic window for targeted RPE damage at a certain power. Increasing the micropulse duration, number of micropulses, and duty cycle leads to a decrease in the selective effect on the RPE and higher damage to adjacent tissues.

CONCLUSION

The concepts of efficiency and selectivity have been introduced to quantify the amount of damage caused. The optimal range of micropulse parameters which result in effective and selective damage on the RPE has been determined for the Navilas 577s laser system. The proposed method can be used for any other serial laser system. A comparison of the different micropulse modes, as well as the CW modes, has been performed.

Hyperspectral optical coherence tomography for in vivo visualization of melanin in the retinal pigment epithelium

Previous studies for melanin visualization in the retinal pigment epithelium (RPE) have exploited either its absorption properties (using photoacoustic tomography or photothermal optical coherence tomography [OCT]) or its depolarization properties (using polarization sensitive OCT). However, these methods are only suitable when the melanin concentration is sufficiently high. In this work, we present the concept of hyperspectral OCT for melanin visualization in the RPE when the concentration is low. Based on white light OCT, a hyperspectral stack of 27 wavelengths (440-700 nm) was created in post-processing for each depth-resolved image. Owing to the size and shape of the melanin granules in the RPE, the variations in backscattering coefficient as a function of wavelength could be identified-a result which is to be expected from Mie theory. This effect was successfully identified both in eumelanin-containing phantoms and in vivo in the low-concentration Brown Norway rat RPE.

Statistical Geometrical Features for Microaneurysm Detection

Automated microaneurysm (MA) detection is still an open challenge due to its small size and similarity with blood vessels. In this paper, we present a novel method which is simple, efficient, and real-time for segmenting and detecting MA in color fundus images (CFI). To do this, a novel set of features based on statistics of geometrical properties of connected regions, that can easily discriminate lesion and non-lesion pixels are used. For large-scale evaluation proposed method is validated on DIARETDB1, ROC, STARE, and MESSIDOR dataset. It proves robust with respect to different image characteristics and camera settings. The best performance was achieved on per-image evaluation on DIARETDB1 dataset with sensitivity of 88.09 at 92.65% specificity which is quite encouraging for clinical use.

Investigating the potential of Zernike polynomials to characterise spatial distribution of macular pigment

It has been postulated that particular patterns of macular pigment (MP) distribution may be associated with the risk for eye diseases such as age-related macular degeneration (AMD). This work investigates the potential of Zernike polynomials (ZP) to characterise the level and distribution of MP, and their suitability as a representation for analysis of the effects of age and AMD on MP patterns. As the case study, MP distribution maps computed using an experimental method based on fundus reflectance (MRIA) were obtained for ninety volunteers representing three groups: under-fifty without AMD, fifty and over without AMD, and fifty and over with AMD. ZP with 105 coefficients were fitted to the maps using least-squares optimisation and found to represent MP maps accurately (RMSE<10⁻¹). One-way MANOVA analysis carried out on ZP representations showed that the three subject groups have significantly different means (Wilk’s Lambda 0.125, p<0.0001). Linear discriminant analysis with leave-one-out scheme resulted in accuracy, sensitivity and specificity of classification according to, respectively, disease status regardless of age (81% all); disease status in the age-matched groups (87%, 88%, 86%); age irrespective of disease status (81%, 83%, 73%); and age for subjects without AMD (83%, 88%, 80%). Mean MP distributions computed from ZP coefficients for the three groups showed more elevated and more peaked MP for the healthy under-fifty group; more irregular and more elevated peripheral levels in over-fifty AMD group than in over-fifty non-AMD group; and moderate radial asymmetry in non-AMD over-50 group. The results suggest that ZP coefficients are capable of accurately representing MP in a way that captures certain spatial patterns of its distribution. Using the ZP representation MP maps could be classified according to both age and disease status with accuracy significantly greater than chance, with peak elevation, pattern irregularity and radial asymmetry identified as important features.

Scattering contribution to the double-pass PSF using Monte Carlo simulations

PURPOSE

Scattering in the eye occurs mainly at two sites: the eye's optical media and the deeper retinal layers. Although the two phenomena are often treated collectively, their spatial domain of contribution to the double-pass Point Spread Function (PSF) is different: the fundus effect is limited to the narrow and middle part of the PSF whereas scattering in the eye's optics extends also to wide angles. The objective of this work was to determine the domain of contribution at the double-pass PSF of light scattered in the ocular media and the ocular fundus, using simulated and experimental data for two different wavelengths and for two different pigmentations.

METHODS

Diffuse reflection was simulated using Monte Carlo simulations for a four-layer retinal fundus model. Four situations were simulated in total for two different choroidal absorptions at two different wavelengths. Light diffusion in the fundus was the only phenomenon considered in the model. The simulations were compared against experimental fundus reflection data obtained in a previous study.

RESULTS

The simulations showed that at 560 nm, diffusion in the fundus causes light to extend to a radius of 2°, independently of the choroidal pigmentation, whereas at 650 nm it extends to radii of 4.5° and 4° for low and high choroidal pigmentation respectively. Experimental data showed a similar behaviour at low angles where light diffusion in the fundus is dominant, but different at higher angles due to scattering in the ocular media.

CONCLUSION

The spatial contribution of light diffused in the ocular fundus to the PSF was found to be limited to narrower angles compared to that of scattering at the ocular media. The comparison of simulated and optical data showed that beyond 2° at 560 nm and 4-4.5° at 650 nm the only phenomenon contributing to the PSF is scattering in the ocular media, whereas the fundus contribution can be assumed as negligible.

Quantifying melanin concentration in retinal pigment epithelium using broadband photoacoustic microscopy

Melanin is the dominant light absorber in retinal pigment epithelium (RPE). The loss of RPE melanin is a sign of ocular senescence and is both a risk factor and a symptom of age-related macular degeneration (AMD). Quantifying the RPE melanin concentration provides insight into the pathological role of RPE in ocular aging and the onset and progression of AMD. The main challenge in accurate quantification of RPE melanin concentration is to distinguish this ten-micrometer-thick cell monolayer from the underlying choroid, which also contains melanin but carries different pathognomonic information. In this work, we investigated a three-dimensional photoacoustic microscopic (PAM) method with high axial resolution, empowered by broad acoustic detection bandwidth, to distinguish RPE from choroid and quantify melanin concentrations in the RPE ex vivo. We first conducted numerical simulation on photoacoustic generation in the RPE, which suggested that a PAM system with at least 100-MHz detection bandwidth provided sufficient axial resolution to distinguish the melanin in RPE from that in choroid. Based on simulation results, we integrated a transparent broadband micro-ring resonator (MRR) based detector in a homebuilt PAM system. We imaged ex vivo RPE-choroid complexes (RCCs) from both porcine and human eyes and quantified the absolute melanin concentrations in the RPE and choroid, respectively. In our study, the measured melanin concentrations were 14.7 mg/mL and 17.0 mg/mL in human and porcine RPE, and 12 mg/mL and 61 mg/mL in human and porcine choroid, respectively. This study suggests that broadband PAM is capable of quantifying the RPE melanin concentration from RCCs ex vivo.

Preliminary investigation of multispectral retinal tissue oximetry mapping using a hyperspectral retinal camera

Oximetry measurement of principal retinal vessels represents a first step towards understanding retinal metabolism, but the technique could be significantly enhanced by spectral imaging of the fundus outside of main vessels. In this study, a recently developed Hyperspectral Retinal Camera was used to measure relative oximetric (SatO2) and total hemoglobin (HbT) maps of the retina, outside of large vessels, in healthy volunteers at baseline (N = 7) and during systemic hypoxia (N = 11), as well as in patients with glaucoma (N = 2). Images of the retina, on a field of view of ∼30°, were acquired between 500 and 600 nm with 2 and 5 nm steps, in under 3 s. The reflectance spectrum from each pixel was fitted to a model having oxy- and deoxyhemoglobin as the main absorbers and scattering modeled by a power law, yielding estimates of relative SatO2 and HbT over the fundus. Average optic nerve head (ONH) saturation over 8 eyes was 68 ± 5%. During systemic hypoxia, mean ONH saturation decreased by 12.5% on average. Upon further development and validation, the relative SatO2 and HbT maps of microvasculature obtained with this imaging system could ultimately contribute to the diagnostic and management of diseases affecting the ONH and retina.

Optical eye simulator for laser dazzle events

An optical simulator of the human eye and its application to laser dazzle events are presented. The simulator combines optical design software (ZEMAX) with a scientific programming language (MATLAB) and allows the user to implement and analyze a dazzle scenario using practical, real-world parameters. Contrary to conventional analytical glare analysis, this work uses ray tracing and the scattering model and parameters for each optical element of the eye. The theoretical background of each such element is presented in relation to the model. The overall simulator's calibration, validation, and performance analysis are achieved by comparison with a simpler model based uponCIE disability glare data. Results demonstrate that this kind of advanced optical eye simulation can be used to represent laser dazzle and has the potential to extend the range of applicability of analytical models.

Monte Carlo investigation on quantifying the retinal pigment epithelium melanin concentration by photoacoustic ophthalmoscopy

The retinal pigment epithelium (RPE) melanin plays an important role in maintaining normal visual functions. A decrease in the RPE melanin concentration with aging is believed to be associated with several blinding diseases, including age-related macular degeneration. Quantifying the RPE melanin noninvasively is therefore important in evaluating the retinal health and aging conditions. Photoacoustic ophthalmoscopy (PAOM), as an optical absorption-based imaging technology, can potentially be applied to measure variations in the RPE melanin if the relationship between the detected photoacoustic (PA) signal amplitudes and the RPE melanin concentrations can be established. In this work, we tested the feasibility of using PA signals from retinal blood vessels as references to measure RPE melanin variation using Monte Carlo (MC) simulation. The influences from PAOM axial resolution, the depth and diameter of the retinal blood vessel, and the RPE thickness were examined. We proposed a calibration scheme by relating detected PA signals to the RPE melanin concentrations, and we found that the scheme is robust to these tested parameters. This study suggests that PAOM has the capability of quantitatively measuring the RPE melanin in vivo.

A method to assist in the diagnosis of early diabetic retinopathy: Image processing applied to detection of microaneurysms in fundus images

Diabetes increases the risk of developing any deterioration in the blood vessels that supply the retina, an ailment known as Diabetic Retinopathy (DR). Since this disease is asymptomatic, it can only be diagnosed by an ophthalmologist. However, the growth of the number of ophthalmologists is lower than the growth of the population with diabetes so that preventive and early diagnosis is difficult due to the lack of opportunity in terms of time and cost. Preliminary, affordable and accessible ophthalmological diagnosis will give the opportunity to perform routine preventive examinations, indicating the need to consult an ophthalmologist during a stage of non proliferation. During this stage, there is a lesion on the retina known as microaneurysm (MA), which is one of the first clinically observable lesions that indicate the disease. In recent years, different image processing algorithms, which allow the detection of the DR, have been developed; however, the issue is still open since acceptable levels of sensitivity and specificity have not yet been reached, preventing its use as a pre-diagnostic tool. Consequently, this work proposes a new approach for MA detection based on (1) reduction of non-uniform illumination; (2) normalization of image grayscale content to improve dependence of images from different contexts; (3) application of the bottom-hat transform to leave reddish regions intact while suppressing bright objects; (4) binarization of the image of interest with the result that objects corresponding to MAs, blood vessels, and other reddish objects (Regions of Interest-ROIs) are completely separated from the background; (5) application of the hit-or-miss Transformation on the binary image to remove blood vessels from the ROIs; (6) two features are extracted from a candidate to distinguish real MAs from FPs, where one feature discriminates round shaped candidates (MAs) from elongated shaped ones (vessels) through application of Principal Component Analysis (PCA); (7) the second feature is a count of the number of times that the radon transform of the candidate ROI, evaluated at the set of discrete angle values {0°, 1°, 2°, …, 180°}, is characterized by a valley between two peaks. The proposed approach is tested on the public databases DiaretDB1 and Retinopathy Online Challenge (ROC) competition. The proposed MA detection method achieves sensitivity, specificity and precision of 92.32%, 93.87% and 95.93% for the diaretDB1 database and 88.06%, 97.47% and 92.19% for the ROC database. Theory, results, challenges and performance related to the proposed MA detecting method are presented.

Monte Carlo simulation of retinal light absorption by infants

Retinal damage can occur in normal ambient lighting conditions. Infants are particularly vulnerable to retinal damage, and thousands of preterm infants sustain vision damage each year. The size of the ocular fundus affects retinal light absorption, but there is a lack of understanding of this effect for infants. In this work, retinal light absorption is simulated for different ocular fundus sizes, wavelengths, and pigment concentrations by using the Monte Carlo method. The results indicate that the neural retina light absorption per volume for infants can be two or more times that for adults.

Monte Carlo simulation of laser beams interaction with the human eye using Geant4

Background

Due to the unique characteristics of the eye, ophthalmologic diagnostic techniques often rely on the photons interaction with the retina to infer its internal structure. Although these techniques are widely used, the interpretation of the generated images is not always fully understood, as in scanning laser ophthalmoscopy dark field imaging. This limits the exploitation of its full potential as a diagnostic tool for deep abnormalities in the retina, as in the situation of drusen.

Methods

With the aim of better understanding the retinal diagnostic images, we have carried out computer simulations of incident laser beams interacting with different structures of the human eye, including a retina with and without drusen. We have used the Geant4 simulation toolkit, applying the optical package of the electromagnetic (EM) physics working group, to simulate the physical processes of reflection, refraction, absorption, and scattering of low energy photons (2 eV) in biological tissues. For each simulation it was used a single beam of orange light, with a Gaussian profile, that travels through all optical elements of the eye. The reflected beam characteristics were analyzed by virtual detectors in different locations, which collected information about the number and position of photons. The geometry and optical properties of all components of the eye were considered according to the published data.

Results

Simulation results put in evidence that the presence of drusen influences the profile of the reflected beams. It changes the mean free path of the photons, modifying its reflection pattern, which depends on the area illuminated by the incident beam. This result is also visible when the reflected beam is analyzed outside of the eye, when the profile has no longer a symmetrical Gaussian distribution. These results will support the retinal diagnostic images that will be obtained in a near future with a new developed ophthalmic apparatus.

Conclusions

The shape analysis of the reflected beams in retinal laser scanning techniques could increase its potential as a diagnostic examination tool for the deeper structures of the retina.

Accuracy of retinal oximetry: a Monte Carlo investigation

Retinal hemoglobin oxygen saturation (sO2) level is believed to be associated with the pathophysiology of several leading blinding diseases. Methods to properly measure retinal sO2 have been investigated for decades; however, the accuracy of retinal oximetry is still considered to be limited. The Monte Carlo simulation of photon transport in retina to examine how the accuracy of retinal oximetry is affected by local parameters is discussed. Fundus photography was simulated in a multilayer retinal model, in which a single vessel segment with 0.7 sO2 was embedded, at six optical wavelengths. Then, 200 million photons were traced in each simulation to ensure statistically stable results. The optical reflectance and energy deposit were recorded to measure sO2 using both the reflection method (existing retinal oximetry) and a new absorption method, photoacoustic ophthalmoscopy (PAOM). By varying the vessel diameter and melanin concentration in the retinal pigment epithelium, the relative error of sO2 measurement in the reflection method increased with increasing vessel diameter and melanin concentration; in comparison, the sO2 measurement was insensitive to these two parameters in PAOM. The results suggest that PAOM potentially can be a more accurate tool in quantifying retinal sO2.

Microfluidics based phantoms of superficial vascular network

Several new bio-photonic techniques aim to measure flow in the human vasculature non-destructively. Some of these tools, such as laser speckle imaging or Doppler optical coherence tomography, are now reaching the clinical stage. Therefore appropriate calibration and validation techniques dedicated to these particular measurements are therefore of paramount importance. In this paper we introduce a fast prototyping technique based on laser micromachining for the fabrication of dynamic flow phantoms. Micro-channels smaller than 20 µm in width can be formed in a variety of materials such as epoxies, plastics, and household tape. Vasculature geometries can be easily and quickly modified to accommodate a particular experimental scenario.

Drusen characterization with multimodal imaging

PURPOSE

To characterize the known appearance of cuticular drusen, subretinal drusenoid deposits (reticular pseudodrusen), and soft drusen as revealed by multimodal fundus imaging and to create an explanatory model that accounts for these observations.

METHODS

Reported color, fluorescein angiographic, autofluorescence, and spectral domain optical coherence tomography images of patients with cuticular drusen, soft drusen, and subretinal drusenoid deposits were reviewed, as were actual images from affected eyes. Representative histological sections were examined. The geometry, location, and imaging characteristics of these lesions were evaluated. A hypothesis based on the Beer-Lambert law of light absorption was generated to fit these observations.

RESULTS

Cuticular drusen appear as numerous, uniform, round, yellow-white punctate accumulations under the retinal pigment epithelium (RPE). Soft drusen are larger, yellow-white dome-shaped mounds of deposit under the RPE. Subretinal drusenoid deposits are polymorphous light-gray interconnected accumulations above the RPE. Based on the model, both cuticular and soft drusen appear yellow because of the removal of shorter wavelength light by a double pass through the RPE. Subretinal drusenoid deposits, which are located on the RPE, are not subjected to short-wavelength attenuation and therefore are more prominent when viewed with blue light. The location and morphology of extracellular material in relationship to the RPE, and associated changes to RPE morphology and pigmentation, appeared to be the primary determinants of druse appearance in different imaging modalities.

CONCLUSION

Although cuticular drusen, subretinal drusenoid deposits, and soft drusen are composed of common components, they are distinguishable by multimodal imaging because of differences in location, morphology, and optical filtering effects by drusenoid material and the RPE.

Ketamine-xylazine anesthesia causes hyperopic refractive shift in mice

Mice have increasingly been used as a model for studies of myopia. The key to successful use of mice for myopia research is the ability to obtain accurate measurements of refractive status of their eyes. In order to obtain accurate measurements of refractive errors in mice, the refraction needs to be performed along the optical axis of the eye. This represents a particular challenge, because mice are very difficult to immobilize. Recently, ketamine-xylazine anesthesia has been used to immobilize mice before measuring refractive errors, in combination with tropicamide ophthalmic solution to induce mydriasis. Although these drugs have increasingly been used while refracting mice, their effects on the refractive state of the mouse eye have not yet been investigated. Therefore, we have analyzed the effects of tropicamide eye drops and ketamine-xylazine anesthesia on refraction in P40 C57BL/6J mice. We have also explored two alternative methods to immobilize mice, i.e. the use of a restraining platform and pentobarbital anesthesia. We found that tropicamide caused a very small, but statistically significant, hyperopic shift in refraction. Pentobarbital did not have any substantial effect on refractive status, whereas ketamine-xylazine caused a large and highly significant hyperopic shift in refraction. We also found that the use of a restraining platform represents good alternative for immobilization of mice prior to refraction. Thus, our data suggest that ketamine-xylazine anesthesia should be avoided in studies of refractive development in mice and underscore the importance of providing appropriate experimental conditions when measuring refractive errors in mice.

Dynamic eye phantom for retinal oximetry measurements

Measurements of oxygen saturation and flow in the retina can yield information about eye health and the onset of eye pathologies such as diabetic retinopathy. Recently, we developed a multiaperture camera that uses the division of the retinal image into several wavelength-sensitive subimages to compute retinal oxygen saturation. The calibration of such instruments is particularly difficult due to the layered structure of the eye and the lack of alternative measurement techniques. For this purpose, we realize an in vitro model of the human eye composed of a lens, the retina vessel, and three layers: the choroid, the retinal pigmented epithelium, and the sclera. The retinal vessel is modeled with a microtube connected to a micropump and a hemoglobin reservoir in a closed circulatory system. Hemoglobin oxygenation in the vessel could be altered using a reversible fuel cell. The sclera is represented by a Spectralon slab. The optical properties of the other layers are mimicked using titanium dioxide as a scatterer, ink as an absorber, and epoxy as a supporting structure. The optical thickness of each layer of the eye phantom is matched to each respective eye layer.

Analysis of postnatal eye development in the mouse with high-resolution small animal magnetic resonance imaging

PURPOSE

Studies of myopia in mice have been complicated by the difficulty in obtaining accurate measurements of small changes observed in the growing mouse eye in vivo and the lack of data on refractive eye development. The purpose of this study was to carry out an in vivo high-resolution analysis of mouse eye growth and refractive development.

METHODS

High-resolution small animal magnetic resonance imaging and high-resolution infrared photorefraction were used to analyze refractive development in postnatal day (P)21 to P89 C57BL/6J mice.

RESULTS

The growth of the mouse eye decelerated after P40. The eye maintained a slightly prolate shape during growth. The anterior chamber growth exhibited a similar pattern, whereas the corneal radius of curvature (CRC) increased linearly. The growth rate of the lens remained constant until P89. The lens "overgrew" the eye at P40, resulting in a decline in vitreous chamber depth. Mice showed myopic refractive errors at a younger age (-13.2 +/- 2.0 D; mean +/- SD, P21). The refractive errors stabilized around emmetropic values by P32 and remained emmetropic until P40. Mice became progressively hyperopic with age (+1.2 +/- 1.7 D, P67; +3.6 +/- 2.3 D, P89).

CONCLUSIONS

Development of ocular components in the mouse is similar to that of the tree shrew but different from that of higher primates and humans. Primary differences can be attributed to the age-related changes of the crystalline lens and CRC. In spite of these differences, mice appear to be able to achieve and maintain emmetropic refractive status at P32 to P40.

Macular pigment, photopigments, and melanin: distributions in young subjects determined by four-wavelength reflectometry

We have developed an objective procedure, using a modified retinal camera, to determine macular pigment (MP) optical density distributions in the human retina. Using two multi-band filters, reflectance maps of the retinas of young subjects (<25 years old) were obtained at 460, 528, 610 and 670 nm, without pupil dilation. The log-transformed maps were combined linearly to yield optical density maps of MP, cone and rod photopigments, and melanin. MP optical density and heterochromatic flicker photometry results for 22 subjects were in reasonable agreement. Cone photopigments, like MP, showed similar, well-defined peaks at the fovea, whereas rod photopigment showed a minimum. Melanin was more broadly distributed.

Monte Carlo simulation of latanoprost induced iris darkening

The aim of this study was to provide numerical evidence that latanoprost induced iris darkening (LIID) can be caused by changes to the melanin granule size distribution in the anterior segment of the iris. Iridectomies from two patients were used, where both had undergone unilateral treatment with latanoprost and had exhibited LIID. The untreated eye provided the comparative control. Micrographs from the iris samples were analysed to determine the number and size of the mature melanin granules. Monte Carlo (MC) simulation of light propagation in the iris was performed to examine the changes in reflectance and absorption with varying particle size and density. The reflected intensity of light was obtained as a function of wavelength. CIE colour theory was employed in order to estimate a perceived colour from the reflectance data. MC simulations showed that the reflectance was reduced for the LIID irises compared to the control irises for both subjects and for all wavelengths of light. The MC simulated colours were in good agreement with the in vivo photography of the eye colour. Hence, we have demonstrated that increases in melanin granule size causes iris darkening, and can explain LIID.

Automatic detection of microaneurysms in color fundus images

This paper addresses the automatic detection of microaneurysms in color fundus images, which plays a key role in computer assisted diagnosis of diabetic retinopathy, a serious and frequent eye disease. The algorithm can be divided into four steps. The first step consists in image enhancement, shade correction and image normalization of the green channel. The second step aims at detecting candidates, i.e. all patterns possibly corresponding to MA, which is achieved by diameter closing and an automatic threshold scheme. Then, features are extracted, which are used in the last step to automatically classify candidates into real MA and other objects; the classification relies on kernel density estimation with variable bandwidth. A database of 21 annotated images has been used to train the algorithm. The algorithm was compared to manually obtained gradings of 94 images; sensitivity was 88.5% at an average number of 2.13 false positives per image.

Qualitative assessment of undetectable melanin distribution in lightly pigmented irides

A longstanding issue in biomedical research concerns the biophysical factors affecting the color diversity of the human irides and its connection with ocular diseases. Although the pigmentation and morphological characteristics of heavily and moderately pigmented irides have been extensively discussed in the literature, similar studies are scarce for lightly pigmented irides. We present computer experiments indicating that the spectral signature of these specimens may be directly affected by still undetectable melanin distributions in the outermost iridal layers. Our findings represent in silico evidence of this relationship which, in turn, may have implications in the investigation of the higher risk of death from metastatic ocular melanoma verified in individuals with light-colored irides.

Quantitative analysis of multi-spectral fundus images

We have developed a new technique for extracting histological parameters from multi-spectral images of the ocular fundus. The new method uses a Monte Carlo simulation of the reflectance of the fundus to model how the spectral reflectance of the tissue varies with differing tissue histology. The model is parameterised by the concentrations of the five main absorbers found in the fundus: retinal haemoglobins, choroidal haemoglobins, choroidal melanin, RPE melanin and macular pigment. These parameters are shown to give rise to distinct variations in the tissue colouration. We use the results of the Monte Carlo simulations to construct an inverse model which maps tissue colouration onto the model parameters. This allows the concentration and distribution of the five main absorbers to be determined from suitable multi-spectral images. We propose the use of "image quotients" to allow this information to be extracted from uncalibrated image data. The filters used to acquire the images are selected to ensure a one-to-one mapping between model parameters and image quotients. To recover five model parameters uniquely, images must be acquired in six distinct spectral bands. Theoretical investigations suggest that retinal haemoglobins and macular pigment can be recovered with RMS errors of less than 10%. We present parametric maps showing the variation of these parameters across the posterior pole of the fundus. The results are in agreement with known tissue histology for normal healthy subjects. We also present an early result which suggests that, with further development, the technique could be used to successfully detect retinal haemorrhages.

Model-based parameter recovery from uncalibrated optical images

We propose a novel method for quantitative interpretation of uncalibrated optical images which is derived explicitly from an analysis of the image formation model. Parameters characterising the tissue are recovered from images acquired using filters optimised to minimise the error. Preliminary results are shown for the skin, where the technique was successfully applied to aid the diagnosis and interpretation of non-melanocytic skin cancers and acne; and for the more challenging ocular fundus, for mapping of the pigment xanthophyll.

In vivo confocal imaging of the retina in animal models using scanning laser ophthalmoscopy

Scanning-laser ophthalmoscopy is a technique for confocal imaging of the eye in vivo. The use of lasers of different wavelengths allows to obtain information about specific tissues and layers due to their reflection and transmission characteristics. In addition, fluorescent dyes excitable in the blue and infrared range offer a unique access to the vascular structures associated with each layer. In animal models, a further enhancement in specificity can be obtained by GFP expression under control of tissue-specific promotors. Important fields of application are studies in retinal degenerations and the follow-up of therapeutic intervention.

Modelling and validation of spectral reflectance for the colon

The spectral reflectance of the colon is known to be affected by malignant and pre-malignant changes in the tissue. As part of long-term research on the derivation of diagnostically important parameters characterizing colon histology, we have investigated the effects of the normal histological variability on the remitted spectra. This paper presents a detailed optical model of the normal colon comprising mucosa, submucosa and the smooth muscle layer. Each layer is characterized by five variable histological parameters: the volume fraction of blood, the haemoglobin saturation, the size of the scattering particles, including collagen, the volume fraction of the scattering particles and the layer thickness, and three optical parameters: the anisotropy factor, the refractive index of the medium and the refractive index of the scattering particles. The paper specifies the parameter ranges corresponding to normal colon tissue, including some previously unpublished ones. Diffuse reflectance spectra were modelled using the Monte Carlo method. Validation of the model-generated spectra against measured spectra demonstrated that good correspondence was achieved between the two. The analysis of the effect of the individual histological parameters on the behaviour of the spectra has shown that the spectral variability originates mainly from changes in the mucosa. However, the submucosa and the muscle layer must be included in the model as they have a significant constant effect on the spectral reflectance above 600 nm. The nature of variations in the spectra also suggests that it may be possible to carry out model inversion and to recover parameters characterizing the colon from multi-spectral images. A preliminary study, in which the mucosal blood and collagen parameters were modified to reflect histopathological changes associated with colon cancer, has shown that the spectra predicted by our model resemble measured spectral reflectance of adenocarcinomas. This suggests that an extended model, which incorporates parameters corresponding to an abnormal colon, may be effective for differentiation between normal and cancerous tissues.

Absorption of the eye lens and macular pigment derived from the reflectance of cone photoreceptors

We measured the amplitude of the directional component of the bleached fundus reflectance, the so-called optical Stiles-Crawford effect, as a function of wavelength. The directional reflectance originates from within the outer segments of the photoreceptors. Thus only two anterior absorbers are of importance: macular pigment and the crystalline lens. Analysis of spectra obtained in pseudophakes established that the cone photoreceptors act as spectrally neutral reflectors. The reflectance spectra, expressed in density units, resembled the macular pigment density spectrum. Studying age effects in the lens of normal subjects resulted in a description of the optical density of the lens in terms of a "young" and an "aged" template. The young template represents the pigment O-beta-glucoside of 3-hydroxykynurenine, which dominates the light absorption in young eyes and decreases with age. The aged template represents the pigments accumulating in the lens with age. The total optical density increased with age, but it was lower in the wavelength region 500-650 nm than was previously assumed on the basis of psychophysical studies. Analysis of the spectra also provided precise individual estimates of the optical density of macular pigment. Finally, we observed a decrease in the photoreceptor reflectivity with age, possibly reflecting a degradation of the photoreceptors.

Spectral filter optimization for the recovery of parameters which describe human skin

This paper presents a method for finding spectral filters that minimize the error associated with histological parameters characterizing normal skin tissue. These parameters can be recovered from digital images of the skin using a physics-based model of skin coloration. The relationship between the image data and histological parameter values is defined as a mapping function from the image space to the parameter space. The accuracy of this function is determined by the choice of optical filters. An optimization criterion for finding the optimal filters is defined by combing methodology from differential geometry with statistical error analysis. It is shown that the magnitude of errors associated with the optimal filters is typically half of that for typical RGB filters on a three-parameter model of human skin coloration. Finally, other medical image applications are identified to which this generic methodology could be applied.

From colour to tissue histology: Physics-based interpretation of images of pigmented skin lesions

Through an understanding of the image formation process, diagnostically important facts about the internal structure and composition of pigmented skin lesions can be derived from their colour images. A physics-based model of tissue colouration provides a cross-reference between image colours and the underlying histological parameters. It is constructed by computing the spectral composition of light remitted from the skin given parameters specifying its structure and optical properties. The model is representative of all the normal human skin colours, irrespective of racial origin, age or gender. Abnormal skin colours do not conform to this model and thus can be detected. Once the model is constructed, for each pixel in a colour image its histological parameters are computed from the model. Represented as images, these 'parametric maps' show the concentration of dermal and epidermal melanin, blood and collagen thickness across the imaged skin as well as locations where abnormal colouration exists. In a clinical study the parametric maps were used by a clinician to detect the presence of malignant melanoma in a set of 348 pigmented lesions imaged using a commercial device, the SIAscope. Logistic regression identified the presence of melanin in the dermis, the abnormal distribution of blood within the lesion and the lesion size as the most diagnostically informative features. Classification based on these features showed 80.1% sensitivity and 82.7% specificity in melanoma detection.