Design, Development and Realization of Quadratic Volterra Filters for Selected Applications

The basic concepts of digital signal processing are taught to the students in engineering and science. The focus of the course is on linear, time invariant systems. The question as to what happens when the system is governed by a quadratic or cubic equation remains unanswered in the vast majority of literature on signal processing. Light has been shed on this problem when John V Mathews and Giovanni L Sicuranza published the book Polynomial Signal Processing. This book opened up an unseen vista of polynomial systems for signal and image processing. The book presented the theory and implementations of both adaptive and non-adaptive FIR and IIR quadratic systems which offer improved performance than conventional linear systems. The theory of quadratic systems presents a pristine and virgin area of research that offers computationally intensive work. Once the area of research is selected, the next issue is the choice of the software tool to carry out the work. Conventional languages like C and C++ are easily eliminated as they are not interpreted and lack good quality plotting libraries. MATLAB is proved to be very slow and so do SCILAB and Octave. The search for a language for scientific computing that was as fast as C, but with a good quality plotting library, ended up in Python, a distant relative of LISP. It proved to be ideal for scientific computing. An account of the use of Python, its scientific computing package scipy and the plotting library pylab is given in the appendix Initially, work is focused on designing predictors that exploit the polynomial nonlinearities inherent in speech generation mechanisms. Soon, the work got diverted into medical image processing which offered more potential to exploit by the use of quadratic methods. The major focus in this area is on quadratic edge detection methods for retinal images and fingerprints as well as de-noising raw MRI signals

Binocular vision in insects: How mantids solve the correspondence problem

Praying mantids use binocular cues to judge whether their prey is in striking distance. When there are several moving targets within their binocular visual field, mantids need to solve the correspondence problem. They must select between the possible pairings of retinal images in the two eyes so that they can strike at a single real target. In this study, mantids were presented with two targets in various configurations, and the resulting fixating saccades that precede the strike were analyzed. The distributions of saccades show that mantids consistently prefer one out of several possible matches. Selection is in part guided by the position and the spatiotemporal features of the target image in each eye. Selection also depends upon the binocular disparity of the images, suggesting that insects can perform local binocular computations. The pairing rules ensure that mantids tend to aim at real targets and not at “ghost” targets arising from false matches.

Inhibition of sustained gamma oscillations (35-80 Hz) by fast transient responses in cat visual cortex.

Interactions between stimulus-induced oscillations (35-80 Hz) and stimulus-locked nonoscillatory responses were investigated in the visual cortex areas 17 and 18 of anaesthetized cats. A single square-wave luminance grating was used as a visual stimulus during simultaneous recordings from up to seven electrodes. The stimulus movement consisted of a superposition of a smooth movement with a sequence of dynamically changing accelerations. Responses of local groups of neurons at each electrode were studied on the basis of multiple unit activity and local slow field potentials (13-120 Hz). Oscillatory and stimulus-locked components were extracted from multiple unit activity and local slow field potentials and quantified by a combination of temporal and spectral correlation methods. We found fast stimulus-locked components primarily evoked by sudden stimulus accelerations, whereas oscillatory components (35-80 Hz) were induced during slow smooth movements. Oscillations were gradually reduced in amplitude and finally fully suppressed with increasing amplitudes of fast stimulus-locked components. It is argued that suppression of oscillations is necessary to prevent confusion during sequential processing of stationary and fast changing retinal images.

Telepaediatrics and diabetic retinopathy screening of young people with diabetes in Queensland

We have examined the feasibility of a telemedicine-enabled screening service for children and adolescents with diabetes in Queensland. There are approximately 1400 young people with diabetes in Queensland and only about two-thirds of them are screened in accordance with international guidelines. A regional retinal screening service was established using a non-mydriatic digital retinal camera. Seven centres volunteered to participate in the study. During a five-month pilot trial, 83 of the young people with diabetes who attend these centres underwent digital retinal screening (3.7%). Retinal images were sent via email to a paediatric ophthalmologist for review and results were returned via email. A copy of each participant's results was forwarded by mail to the referring diabetes doctor and the participant and family. The majority of the image files (96%) were rated as excellent or good. Only one participant was identified as having an abnormal result. Participants and their families expressed satisfaction with the digital retinal screening process.

Obstructive sleep apnoea is associated with sight threatening retinopathy and predicts the developement of preproliferative and proliferative retinopathy in patients with type 2 diabetes:a longitudinal analysis

troduct I on . An observational longitudinal study. P ur P ose . Assess the relationship between obstructive sleep apnoea (OSA) and DR cross-sectionally and longitudinally. M ethods . Adults with Type 2 diabetes mellitus (T2DM), who were re - cruited from a hospital-based diabetes clinic in the UK. Patients with pre-existing OSA, end-stage renal disease and non-diabetic retinopa - thy were excluded. OSA (apnoea hypopnea index ≥ 5 events/hour) was assessed by a single overnight home-based cardio-respiratory study (Alice PDX, Philips Respironics, USA). DR was assessed us - ing retinal images between 2007 and 2012. Sight threatening diabetic retinopathy (STDR) was defined as presence of pre-proliferative or proliferative DR, maculopathy or photocoagulation. Advanced DR was defined as pre-proliferative or proliferative DR. r esults . 199 patients were included (57.3% (n=114) men, 47.7% (n=95) White Europeans). STDR and OSA prevalence were 38.7% (n=77) and 62.8% respectively. A t b A sel I ne . STR prevalence was higher in patients with OSA (OSA+) compared to those without OSA (OSA-) [48.8% n=61 vs. 21.6% n=16, p<0.001]. After adjustment for confounders, OSA remained independently associated with STR (OR 3.7, 95% CI 1.6-8.9, p=0.006), maculopathy (OR 4.5, 95% CI 1.8-11.4, p=0.002) and advanced DR (OR 3.9, 95% CI 1.02-15.3, p=0.047). Mild and moderate to severe OSA were independently associated with STR and maculopathy and only moderate to severe OSA was associated with advanced DR following adjustment for con - founders. l ong I tud I n A lly . Over the follow-up period of (4.4±1 years), more OSA+ patients progressed from no or background DR to advanced DR (15.3% (n=17) vs. 3% (n=2), p=0.01). OSA was an independent pre - dictor of advanced DR development after adjustment for confounders (OR 6.6, 95% CI 1.2-35.1, p=0.03). OSA did not predict the develop - ment of maculopathy. c onclus I ons . OSA is independently associated with STR and predicts the development of preproliferative and proliferative DR. Intervention - al studies are needed to assess the impact of OSA treatment on DR.

Obstructive sleep apnoea predicts the development of preproliferative and proliferative retinopathy in patients with type 2 diabetes:a longitudinal analysis

Background and aims: Diabetic Retinopathy (DR) is a leading cause of blindness. OSA is associated with increased oxidative and nitrosative stress and endothelial dysfunction in patients with type 2 diabetes (T2DM). Hence, it is plausible that OSA can promote the development and progression of DR. Materials and methods: An observational longitudinal study in adults with T2DM. Patients with pre-existing OSA, end-stage renal disease and non-diabetic retinopathy were excluded. OSA (apnoea hypopnea index ≥ 5 events/hour) was assessed by a single overnight home-based cardio-respiratory monitoring (Alice PDX, etc.). DR was assesses using retinal images between 2007 and 2012. Sight threatening retinopathy (STR) was defined as pre-proliferative or proliferative DR, maculopathy or photocoagulation. Advanced DR was defined as pre-proliferative or proliferative DR. Results: 199 patients were included (57.3% men, 47.7% White Europeans). STR and OSA prevalence were 38.7 % and 62.8% respectively. STR preva-lence was higher in patients with OSA (OSA+) compared to those with-out (OSA-) [48.8% vs. 21.6%, p <0.001]. After adjustment for confounders, OSA remained independently associated with STR (OR 3.7, 95%CI 1.6-8.9, p=0.006, maculopathy (OR 4.5, 1.8-11.4, p=0.002) and advanced DR (OR 3.9, 1.02-15.3, p=0.047). Over 4.4±1 years, more OSA+ patients progressed from no or background DR to advanced DR (15.3% vs. 3%, p=0.01). OSA was an independent predictor of advanced DR development after adjustment (OR 6.6, 95%CI 1.2-35.1, p=0.03). OSA did not predict the development of maculopathy. Patients received continuous positive airway pressure treatment were less likely to develop advanced DR. Conclusion: OSA is independently associated with STR and predicts the development of preproliferative and proliferative DR. Interventional studies are needed to assess the impact of OSA treatment on DR.Supported by: NIHR (UK) and The UK Novo Nordisk Research Foundation.

Development of Swept Source Optical Coherence Tomography and Adaptive Optics Scanning Laser Ophthalmoscopy: Improved Imaging Speed and Handheld Applications

Optical coherence tomography (OCT) is a noninvasive three-dimensional interferometric imaging technique capable of achieving micrometer scale resolution. It is now a standard of care in ophthalmology, where it is used to improve the accuracy of early diagnosis, to better understand the source of pathophysiology, and to monitor disease progression and response to therapy. In particular, retinal imaging has been the most prevalent clinical application of OCT, but researchers and companies alike are developing OCT systems for cardiology, dermatology, dentistry, and many other medical and industrial applications.

Adaptive optics (AO) is a technique used to reduce monochromatic aberrations in optical instruments. It is used in astronomical telescopes, laser communications, high-power lasers, retinal imaging, optical fabrication and microscopy to improve system performance. Scanning laser ophthalmoscopy (SLO) is a noninvasive confocal imaging technique that produces high contrast two-dimensional retinal images. AO is combined with SLO (AOSLO) to compensate for the wavefront distortions caused by the optics of the eye, providing the ability to visualize the living retina with cellular resolution. AOSLO has shown great promise to advance the understanding of the etiology of retinal diseases on a cellular level.

Broadly, we endeavor to enhance the vision outcome of ophthalmic patients through improved diagnostics and personalized therapy. Toward this end, the objective of the work presented herein was the development of advanced techniques for increasing the imaging speed, reducing the form factor, and broadening the versatility of OCT and AOSLO. Despite our focus on applications in ophthalmology, the techniques developed could be applied to other medical and industrial applications. In this dissertation, a technique to quadruple the imaging speed of OCT was developed. This technique was demonstrated by imaging the retinas of healthy human subjects. A handheld, dual depth OCT system was developed. This system enabled sequential imaging of the anterior segment and retina of human eyes. Finally, handheld SLO/OCT systems were developed, culminating in the design of a handheld AOSLO system. This system has the potential to provide cellular level imaging of the human retina, resolving even the most densely packed foveal cones.

Automatic Glaucoma Detection Based on Optic Disc Segmentation and Texture Feature Extraction

The use of digital image processing techniques is prominent in medical settings for the automatic diagnosis of diseases. Glaucoma is the second leading cause of blindness in the world and it has no cure. Currently, there are treatments to prevent vision loss, but the disease must be detected in the early stages. Thus, the objective of this work is to develop an automatic detection method of Glaucoma in retinal images. The methodology used in the study were: acquisition of image database, Optic Disc segmentation, texture feature extraction in different color models and classification of images in glaucomatous or not. We obtained results of 93% accuracy

Automatic segmentation of seven retinal layers in SDOCT images congruent with expert manual segmentation.

Segmentation of anatomical and pathological structures in ophthalmic images is crucial for the diagnosis and study of ocular diseases. However, manual segmentation is often a time-consuming and subjective process. This paper presents an automatic approach for segmenting retinal layers in Spectral Domain Optical Coherence Tomography images using graph theory and dynamic programming. Results show that this method accurately segments eight retinal layer boundaries in normal adult eyes more closely to an expert grader as compared to a second expert grader.

Quantitative Analysis of Mouse Retinal Layers Using Automated Segmentation of Spectral Domain Optical Coherence Tomography Images.

PURPOSE Quantification of retinal layers using automated segmentation of optical coherence tomography (OCT) images allows for longitudinal studies of retinal and neurological disorders in mice. The purpose of this study was to compare the performance of automated retinal layer segmentation algorithms with data from manual segmentation in mice using the Spectralis OCT. METHODS Spectral domain OCT images from 55 mice from three different mouse strains were analyzed in total. The OCT scans from 22 C57Bl/6, 22 BALBc, and 11 C3A.Cg-Pde6b(+)Prph2(Rd2) /J mice were automatically segmented using three commercially available automated retinal segmentation algorithms and compared to manual segmentation. RESULTS Fully automated segmentation performed well in mice and showed coefficients of variation (CV) of below 5% for the total retinal volume. However, all three automated segmentation algorithms yielded much thicker total retinal thickness values compared to manual segmentation data (P < 0.0001) due to segmentation errors in the basement membrane. CONCLUSIONS Whereas the automated retinal segmentation algorithms performed well for the inner layers, the retinal pigmentation epithelium (RPE) was delineated within the sclera, leading to consistently thicker measurements of the photoreceptor layer and the total retina. TRANSLATIONAL RELEVANCE The introduction of spectral domain OCT allows for accurate imaging of the mouse retina. Exact quantification of retinal layer thicknesses in mice is important to study layers of interest under various pathological conditions.

Diurnal Variation of Retinal Thickness with Spectral Domain OCT

Purpose. To investigate whether diurnal variation occurs in retinal thickness measures derived from spectral domain optical coherence tomography (SD-OCT). Methods. Twelve healthy adult subjects had retinal thickness measured with SD-OCT every 2 h over a 10 h period. At each measurement session, three average B-scan images were derived from a series of multiple B-scans (each from a 5 mm horizontal raster scan along the fovea, containing 1500 A-scans/B-scan) and analyzed to determine the thickness of the total retina, as well as the thickness of the outer retinal layers. Average thickness values were calculated at the foveal center, at the 0.5 mm diameter foveal region, and for the temporal parafovea (1.5 mm from foveal center) and nasal parafovea (1.5 mm from foveal center). Results. Total retinal thickness did not exhibit significant diurnal variation in any of the considered retinal regions (p > 0.05). Evidence of significant diurnal variation was found in the thickness of the outer retinal layers (p < 0.05), with the most prominent changes observed in the photoreceptor layers at the foveal center. The photoreceptor inner and outer segment layer thickness exhibited mean amplitude (peak to trough) of daily change of 7 ± 3 μm at the foveal center. The peak in thickness was typically observed at the third measurement session (mean measurement time, 13:06). Conclusions. The total retinal thickness measured with SD-OCT does not exhibit evidence of significant variation over the course of the day. However, small but significant diurnal variation occurs in the thickness of the foveal outer retinal layers.

Retinal and choroidal thickness in myopic anisometropia

PURPOSE: To examine the foveal retinal thickness (RT) and subfoveal choroidal thickness (ChT) between the fellow eyes of myopic anisometropes. METHODS: Twenty-two young (mean age 23 ± 5 years), healthy myopic anisometropes (≥ 1 D spherical equivalent [SEq] anisometropia) without amblyopia or strabismus were recruited. Spectral domain optical coherence tomography (SD-OCT) was used to capture images of the retina and choroid. Customised software was used to register, align and average multiple foveal OCT B-Scan images from each subject in order to enhance image quality. Two independent masked observers then manually determined the RT and ChT at the centre of the fovea from each SD-OCT image, which were then averaged. Axial length was measured using optical low coherence biometry during relaxed accommodation. RESULTS: The mean absolute SEq anisometropia was 1.74 ± 0.95 D and the mean interocular difference in axial length was 0.58 ± 0.41 mm. There was a strong correlation between SEq anisometropia and the interocular difference in axial length (r = 0.90, p < 0.001). Measures of RT and ChT were highly correlated between the two observers (r = 0.99 and 0.97 respectively) and in close agreement (mean inter-observer difference: RT 1.3 ± 2.2 µm, ChT 1.5 ± 13.7 µm). There was no significant difference in RT between the more (218 ± 18 µm) and less myopic eyes (215 ± 18 µm) (p > 0.05). However, the mean subfoveal ChT was significantly thinner in the more myopic eye (252 ± 46 µm) compared to the fellow, less myopic eye (286 ± 58 µm) (p < 0.001). There was a moderate correlation between the interocular difference in ChT and the interocular difference in axial length (r = -0.50, p < 0.01). CONCLUSIONS: Foveal RT was similar between the fellow eyes of myopic anisometropes; however, the subfoveal choroid was significantly thinner in the more myopic (longer) eye of our anisometropic cohort. The interocular difference in ChT correlated with the magnitude of axial anisometropia.

Retinal and choroidal thickness in myopic anisometropia

Purpose: To compare the retinal thickness (RT) and choroidal thickness (ChT) between the fellow eyes of non-amblyopic myopic anisometropes. Methods: The eyes of 22 non-amblyopic myopic anisometropes (1 D spherical equivalent refraction [SER] anisometropia) were examined using spectral domain optical coherence tomography (SD-OCT). Customised software was used to register, align and average multiple foveal OCT B-Scan images from each subject in order to enhance image quality. Two independent masked observers manually determined the RT and ChT from each SD-OCT image up to 2.5 mm nasal and temporal to the fovea. Axial length (AXL) was measured using optical low coherence biometry during relaxed accommodation. Results: The mean SER anisometropia was 1.74 ± 0.95 D and the mean interocular AXL difference was 0.58 ± 0.41 mm. There was no significant difference in foveal RT between the fellow eyes (P > 0.05). Mean subfoveal ChT was significantly thinner in the more myopic eye (252 ± 46 μm compared to the fellow, less myopic eye (286 ± 58 μm) (P < 0.001). There was a moderate correlation between the interocular difference in subfoveal ChT and the interocular difference in AXL (r = -0.50, P < 0.01). Asian anisometropes displayed more regionally symmetrical (nasal-temporal)interocular differences in ChT profile compared to Caucasians. Conclusions: RT was similar between the fellow eyes of myopic anisometropes; however, the subfoveal choroid was significantly thinner in the more myopic (longer) eye of this anisometropic cohort. The interocular asymmetry in ChT correlated with the interocular difference in AXL.

Automated diagnosis of Age-related Macular Degeneration using greyscale features from digital fundus images

Age-related Macular Degeneration (AMD) is one of the major causes of vision loss and blindness in ageing population. Currently, there is no cure for AMD, however early detection and subsequent treatment may prevent the severe vision loss or slow the progression of the disease. AMD can be classified into two types: dry and wet AMDs. The people with macular degeneration are mostly affected by dry AMD. Early symptoms of AMD are formation of drusen and yellow pigmentation. These lesions are identified by manual inspection of fundus images by the ophthalmologists. It is a time consuming, tiresome process, and hence an automated diagnosis of AMD screening tool can aid clinicians in their diagnosis significantly. This study proposes an automated dry AMD detection system using various entropies (Shannon, Kapur, Renyi and Yager), Higher Order Spectra (HOS) bispectra features, Fractional Dimension (FD), and Gabor wavelet features extracted from greyscale fundus images. The features are ranked using t-test, Kullback–Lieber Divergence (KLD), Chernoff Bound and Bhattacharyya Distance (CBBD), Receiver Operating Characteristics (ROC) curve-based and Wilcoxon ranking methods in order to select optimum features and classified into normal and AMD classes using Naive Bayes (NB), k-Nearest Neighbour (k-NN), Probabilistic Neural Network (PNN), Decision Tree (DT) and Support Vector Machine (SVM) classifiers. The performance of the proposed system is evaluated using private (Kasturba Medical Hospital, Manipal, India), Automated Retinal Image Analysis (ARIA) and STructured Analysis of the Retina (STARE) datasets. The proposed system yielded the highest average classification accuracies of 90.19%, 95.07% and 95% with 42, 54 and 38 optimal ranked features using SVM classifier for private, ARIA and STARE datasets respectively. This automated AMD detection system can be used for mass fundus image screening and aid clinicians by making better use of their expertise on selected images that require further examination.