Automatic extraction of focused regions from images with low depth-of-field (DOF) is a problem without an efficient solution yet. The capability of extracting focused regions can help to bridge the semantic gap by integrating image regions which are meaningfully relevant and generally do not exhibit uniform visual characteristics. There exist two main difficulties for extracting focused regions from low DOF images using high-frequency based techniques: computational complexity and performance. A novel unsupervised segmentation approach based on ensemble clustering is proposed to extract the focused regions from low DOF images in two stages. The first stage is to cluster image blocks in a joint contrast-energy feature space into three constituent groups. To achieve this, we make use of a normal mixture-based model along with standard expectation-maximization (EM) algorithm at two consecutive levels of block size. To avoid the common problem of local optima experienced in many models, an ensemble EM clustering algorithm is proposed. As a result, relevant blocks, i.e., block-based region-of-interest (ROI), closely conforming to image objects are extracted. In stage two, two different approaches have been developed to extract pixel-based ROI. In the first approach, a binary saliency map is constructed from the relevant blocks at the pixel level, which is based on difference of Gaussian (DOG) and binarization methods. Then, a set of morphological operations is employed to create the pixel-based ROI from the map. Experimental results demonstrate that the proposed approach achieves an average segmentation performance of 91.3% and is computationally 3 times faster than the best existing approach. In the second approach, a minimal graph cut is constructed by using the max-flow method and also by using object/background seeds provided by the ensemble clustering algorithm. Experimental results demonstrate an average segmentation performance of 91.7% and approximately 50% reduction of the average computational time by the proposed colour based approach compared with existing unsupervised approaches.

Micromotion, attributable to the modulus mismatch between the brain and electrode materials, is a fundamental phenomenon contributing to electrode failure for invasive Brain-Machine Interfaces. Spike recording quality from conventional chronic electrode designs deteriorates over the weeks/months post-implantation, in terms of signal amplitude and single unit stability, due to glial cell activation by sustained mechanical trauma. Conventional electrode designs consist of a rigid straight shaft and sharp tip, which can augment mechanical trauma sustained due to micromotion. The sinusoidal probe has been fabricated to reduce micromotion related mechanical trauma. The electrode is microfabricated from flexible materials and has design measures such as a sinusoidal shaft, spheroid tip and a 3D polyimide ball anchor to restrict electrode movement relative to the surrounding brain tissue, thus theoretically minimising micromotion. The electrode was compared to standard microwire electrodes and was shown to have more stable chronic recordings in terms of SNR and LFP power. A longer chronic recording period was achieved with the sinusoidal probe for the first generation. Quantitative histology detecting microglia and astrocytes showed reduced neuronal tissue damage especially for the tip region between 6-24 months chronic indwelling period for the sinusoidal probe. This may be linked to the more stable chronic recordings. This is the first demonstration that electrode designs wholly incorporating micromotion- reducing measures may decrease the magnitude of gliosis, with possible chronic recording longevity enhancement.

Medical equipment and environmental surfaces can provide a source of infection if not disinfected properly and contribute to the transmission of healthcare associated infection (HCAI). To help combat the problem of HCAI photocatalytically active titanium dioxide (Ti02) coatings could be used. The coatings have a self-cleaning effect and when activated by light of appropriate wavelength generate reactive oxygen species that can inactivate pathogenic micro-organisms. More recent work in photocatalytic disinfection has centred on the use of doped-Ti02 coatings which could be activated by ambient indoor light without the need for UV sources. In this work a range of metal and non-metal doped Ti02 coatings were characterised and assessed for their anti-bacterial ability for use as potential coatings for environmental surfaces. All films showed the presence of anatase, the most photocatalytically active form of Ti02, however some of the films did not show the presence of the dopant, possibly due to concentration being below XPS detection limit. In disinfection testing, most of the films did not exhibit visible light activated photocatalytic disinfection towards E. coli. Aside from the Ag-Ti02 film which, after 6 h visible light irradiation, did show slightly better inactivation of bacteria when compared to the light control. The only film to reduce bacteria considerably was S-Ti02 after 1.5 h irradiation; however this film also inactivated bacteria in the dark therefore the activity of S-Ti02 was likely due to sulphur radicals and probably not a photocatalytic affect however further investigation into inactivation mechanism is needed. These findings suggest that doped-Ti02 films activated by visible light lacked a photocatalytic affect and would not be beneficial as self-disinfecting surfaces. Ti02 films were then created via sputter deposition. To gain anatase phase a range of power settings and gas settings were varied until finally increasing the gas pressure resulted in anatase formation instead of rutile. The films were tested for their bactericidal ability but showed no enhanced killing when compared to the UVA light. It was thought the planar surface of the sputtered coatings had a low surface area therefore' reducing their photocatalytic disinfecting ability. However the films produced did show potential photocatalytic activity due to an increase in hydrophilicity of the surface after UVA irradiation. Finally anatase Ti02 was sputtered onto Poly methyl methacrylate (PMMA) and assessed for UVC protection, increased hardness and disinfection properties under UVC irradiation. The coating prevented the degradation of the polymer under UVC irradiation and therefore would allow for rapid UVC disinfection of PMMA medical devices to prevent transmission of infection. UVC disinfection studies on the Ti02 coated PMMA showed complete inactivation of E. coli after 10 s and complete inactivation of Fusarium solani spores in 5 min. Hardness and Young's modulus testing showed that both parameters were slightly enhanced with the addition of the Ti02 coating which may increase wear resistance. Therefore, Ti02 coatings could be used as a protective coating for polymers used in medical devices and allow UVC disinfection of these devices to reduce the risk of infection.

Polymeric hydrogels are cross-linked three dimensional networks that can absorb and retain large amounts of water. They also demonstrate excellent biocompatibility, are relatively cheap and can be prepared using straightforward procedures. Photoinduced electron transfer (PET) is a mechanism which can prevent emission from a fluorescence sensor when a target analyte is absent from the sensor binding site. However, when the target analyte binds, the PET mechanism is cancelled and fluorescence is observed. Combining fluorescence sensors with polymeric hydrogels offers the potential for developing materials that respond to changes in their local environment by a modulation of their fluorescence signature. In this thesis polymeric fluorescent sensors based on the PET mechanism will incorporated within polymeric hydrogels and their response to target analytes evaluated. The first results chapter (chapter 2) discusses the synthesis, characterisation and photophysical evaluation of a Zn(ll) selective PET sensor. The sensor was then grafted onto a poly(allylamine) backbone along with a calibration fluorophore giving a ratiometric, polymeric Zn(ll) responsive sensor. The polymer was successfully incorporated into a Gantrez-based hydrogel formulation and shown to be capable of determining Zn(lI) levels ratiometrically. The second results chapter (chapter 3) discusses the use of a pair of commercially available dextran conjugated probes; fura-dextran and texas-redo dextran for their ratiometric response to Ca(ll) levels subject to varied conditions of dilution and Ca(ll) levels. These were shown to be capable of determining Ca(ll) levels from 0 - 12 mM under a range conditions ratio metrically. Attempts to incorporate this pair of probes into hydrogel formulations were undertaken, however these were unsuccessful. Further testing in alternative hydrogels should enable extension ofthis approach to within hydrogel-based systems. In the third results chapter (chapter 4), a similar type Zn(ll) responsive sensor as in re sults chapter 1 with methacrylate functionality was synthesised, characterised and it's photophysical properties evaluated. After redox initiated polymerisation at room temperature, the polymeric hydrogel sensor was fixed onto the bottom of 96 -well plates

Nanoparticles are known for their unique and exceptional properties and are widely used in biomedical, bioelectronics, pharmacology and environmental fields. This thesis looks at the synthesis of iron oxide nanoparticles (Fe₃O₄ and y-Fe₂O₃), iron oxide core-gold shell nanoparticles (γ-Fe₂O₃-Au), their characterization and application in the development of plasma immunoglobulin immunosensor electrode as a biomarker for Alzheimer’s disease (AD). Iron oxide coated with gold shell was used in the development of the immunosensor electrode in other to improve on the sensitivity, low plasma immunoglobulin concentration detection, stability and reproducibility of the immunosensor electrode for AD monitoring and detection. This is achieved by taking advantage of the increased surface area obtained from combining iron oxide core and gold shell nanoparticles for biomolecules binding, their biocompatibility that allows biomolecules to regenerate and the electron conducting properties of gold nanoparticles. Iron oxide nanoparticles in this thesis were synthesized through oxidative alkaline hydrolysis of ferrous salt and the synthesis parameters altered to see how it affects the nanoparticles properties, structure, purity and size. The synthesized iron oxide nanoparticles (Fe₃O₄) showed changes in shape and size and also the formation of side products or impurities mixed with Fe₃O₄ nanoparticles such as alkaganeite (β-FeOOH) nano-rods, α-Fe₂O₃ (hematite) and lepidocrocite (γ-FeOOH). Gold shell nanoparticles average size 24 nm was formed as a shell on γ-Fe₂O₃ nanoparticles (50 nm) through the iterative reduction of chloroauric acid with hydroxylamine. The iron oxide core-gold shell nanoparticles with average particles size 74 nm was used in the development of the immunosensor electrode. The immunosensor electrode and nanoparticles were characterized using physical and electrochemical techniques. The immunosensor electrode for the direct detection of plasma immunoglobulin as biomarker for AD is a novel work since no work has been carried out on direct electrochemical detection of plasma immunoglobulin for AD. The electrode was developed with a view to addressing the problems currently facing present Alzheimer’s disease biomarkers such as being too expensive, technically challenging, rarely available and inability to control repeated sampling for regular monitoring of AD. In developing the immunosensor electrode, depleted plasma immunoglobulin was used as the biomarker and polyclonal rabbit Anti-human IgA, IgG, IgM as the antibody specific for plasma conjugation. Gold electrode and γ-Fe₂O₃-Au nanoparticles were used as the immobilizing substrate for the immunosensor electrode. The immunosensor electrodes showed good response, sensitivity and reproducibility in differentiating plasma immunoglobulin from AD patients and control subjects up to the 8th (3.91 ppm) and 5th plasma concentration (31.25 ppm) for the modified gold electrode and γ-Fe₂O₃-Au electrode respectively. The electrodes had a better linear responds to plasma immunoglobulin at high concentration compared to enzyme linked immunosorbent assay (ELISA) technique.

Since its conception decades ago, Photoplethysmography (PPG) the non-invasive opto-electronic technique that measures arterial pulsations in-vivo has proven its worth by achieving and maintaining its rank as a compulsory standard of patient monitoring. However successful, conventional contact monitoring mode is not suitable in certain clinical and biomedical situations, e.g., in the case of skin damage, or when unconstrained movement is required. With the advance of computer and photonics technologies, there has been a resurgence of interest in PPG and one potential route to overcome the abovementioned issues has been increasingly explored, i.e., imaging photoplethysmography (iPPG). The emerging field of iPPG offers some nascent opportunities in effective and comprehensive interpretation of the physiological phenomena, indicating a promising alternative to conventional PPG. Heart and respiration rate, perfusion mapping, and pulse rate variability have been accessed using iPPG. To effectively and remotely access physiological information through this emerging technique, a number of key issues are still to be addressed. The engineering issues of iPPG, particularly the influence of motion artefacts on signal quality, are addressed in this thesis, where an engineering model based on the revised Beer-Lambert law was developed and used to describe opto-physiological phenomena relevant to iPPG. An iPPG setup consisting of both hardware and software elements was developed to investigate its reliability and reproducibility in the context of effective remote physiological assessment. Specifically, a first study was conducted for the acquisition of vital physiological signs under various exercise conditions, i.e. resting, light and heavy cardiovascular exercise, in ten healthy subjects. The physiological parameters derived from the images captured by the iPPG system exhibited functional characteristics comparable to conventional contact PPG, i.e., maximum heart rate difference was <3 bpm and a significant (p < 0.05) correlation between both measurements were also revealed. Using a method for attenuation of motion artefacts, the heart rate and respiration rate information was successfully assessed from different anatomical locations even in high-intensity physical exercise situations. This study thereby leads to a new avenue for noncontact sensing of vital signs and remote physiological assessment, showing clear and promising applications in clinical triage and sports training. A second study was conducted to remotely assess pulse rate variability (PRV), which has been considered a valuable indicator of autonomic nervous system (ANS) status. The PRV information was obtained using the iPPG setup to appraise the ANS in ten normal subjects. The performance of the iPPG system in accessing PRV was evaluated via comparison with the readings from a contact PPG sensor. Strong correlation and good agreement between these two techniques verify the effectiveness of iPPG in the remote monitoring of PRV, thereby promoting iPPG as a potential alternative to the interpretation of physiological dynamics related to the ANS. The outcomes revealed in the thesis could present the trend of a robust non-contact technique for cardiovascular monitoring and evaluation.

In wheelchair users who experience strong extensor spasms, high muscular forces exerted during the episode may lead to high contact forces between the child and the wheelchair. The forces may be physically powerful enough to cause pain and injury to the child, and can damage or break components of the wheelchair. Dynamic seating systems have been used in an attempt to reduce these contact forces. Such systems permit forward and backward movements as the occupant extends and retracts their body, consequently they are assumed to be beneficial to patients with strong extensor spasms. Questions about the magnitude and direction of the loads which these children can exert through a seating system have been raised. Additionally, the effectiveness of using dynamic components and the advantages of prolonged use remain unclear. The aim of this study was to quantify and compare the imparted forces on equivalent rigid and dynamic seating systems throughout activities of daily living. To achieve this, a mobile strain gauged seating system was developed which allowed the strain generated in the back and footrest components to be measured. At a certain instant of exerted force, the strain data was converted into force and moments acting on the backrest and footrests in three dimensions, assuming static equilibrium. The position of the resultant force on the backrest, termed the centre of pressure (COP), was also calculated. This project shows that the development of a fully mobile data acquisition system is achievable and practical. Results obtained from twelve children during their community based activity of daily living showed no significant differences in the mean and peak interface forces on the backrest between the rigid and dynamic systems. However, when using the dynamic backrest system, a significant decrease in force and bending moments were observed on the right footrest, the dominant side of most participants. Conversely, for the left footrest only the average bending moment about the transverse axis through the ankle showed a statistically significant decrease, with no significant difference demonstrated between the other variables for the two backrest systems. This work included a long-term case study using the dynamic backrest seating system. The data did not elicit any observable differences of changing in movement, probably due to the relative inactivity of the recruited volunteer. Further work and recruitment should focus on users who exhibit strong extensor spasms, as this work suggest that these are the population who may benefit the most from dynamic seating.

The squamo-columnar junction is an important landmark in the upper gastro-oesophageal anatomy, acid reflux occurring above this causes heartburn and may lead to oesophageal cancer if prolonged. Currently no medical device takes into account the position of the squamo-columnar junction despite its importance. The aim of this doctoral work is to develop and validate a device which is capable of monitoring the position of the squamocolumnar junction with respect to simultaneous acid reflux and pressure measurements. Using Hall effect sensors on a custom flexible circuit board, and a magnet attached to the squamo-columnar junction, a device was produced which when inserted into the oesophagus, measured the relative position of the squamocolumnar junction to manometry and pH-metry catheters. The accuracy of the measurement was at most 5 millimetres, often better; a better resolution than either the high resolution manometer or custom pH device. The device was validated in-vivo, demonstra ting a capability of measuring significant movement of the squamo-columnar junction during transient lower oesophageal sphincter relaxations. The accuracy of the high resolution manometer was tested, which showed significant drift, capable of causing misdiagnosis. A correction algorithm was produced which corrected linear drift, removing time dependant drift leading to significantly more accurate pressure readings. Catheter based upper gastro-oesophageal measurements have the potential to cause transient lower oesophageal sphincter relaxations by triggering mechanoreceptors in the pharynx; therefore a non-catheter based squamo-columnar junction locator was designed and tested. Using a larger magnet and significantly more sensitive Hall effect sensors and custom analogue circuitry, the squamo-columnar junction could be detected. The distance between the oesophagus and the skin is estimated to be between 8 and 9 centimetres for a healthy, non-obese male adult, the detection range for the non-catheter based squamo-columnar junction locator was 10.4 centimetres. The devices developed for this doctoral work has improved the field of gastroenterology research.

Molecular imprinted polymer (MIPs) is established as a technology for obtaining biological mimics using artificial materials. The aim of this study is to prepare monolithic non-covalent imprinted polymers for selective extraction of proteins from aqueous media, using a novel strategy of protein coated onto micro crystals (PCMCs) as a platform for protein imprinting. The MIPs were prepared by adding functional monomers methacrylic acid and 4-vinyl pyridine, a cross linker, ethylene glycol dimethacrylate, in a PCMCs/ acetonitrile suspension followed by the addition of the free radical initiator 2,2' azobisisobutyronitrile, in a free radical polymerisation. The MIPs developed exhibited selectivity towards the template but non selectivity towards non specific proteins was also observed. To reduce this, we used chemical additives and non specific blocking agents, however no difference was observed. In a further attempt to combat non selective protein binding, we added a monomer di (eth ylene glycol) vinyl ether (DEGVE) into the polymerisation process. Non specific protein binding was slightly reduced. To improve polymer functionality, we developed UV initiated polymers and these were compared with temperature initiated polymers, both polymers showed similar qualities. To gain understanding on the nature of the protein bound to the polymers, we looked at the template bound polymer activities. It was found that ~>93% of the protein bound to the imprinted polymer was active, while only ~75% of the protein bound to the control was active. Finally, the presence of DEGVE in the polymer matrix was investigated using a FTIR and micro-analysis; however, it was not clear whether this monomer was incorporated into the polymer matrix. Polymer physical characteristics, such as surface area and porosity, were examined using Brunauer, Emmett and Teller (BET), polymer swelling properties with a gravimetric technique and microscopic property with SEM. Most polymers demonstrated different physical characteristics.

Organic distributed feedback (DFB) lasers are a class of evanescent wave technology that can be used to measure changes in refractive index at the laser surface. These sensors are highly attractive for biosensing applications as they provide a sensitive platform for the label-free detection of a range of analytes, possibly in real-time, and they can be multiplexed for the detection of a suite of different analytes from a single test sample. The simple implementation of DFB lasers for sensing also means that they can be packaged into a compact sensing platform; this is especially true of DFB lasers incorporating an organic semiconductor as the gain layer where optical pumping may be performed with a compact source, such as a laser diode. In addition, organic semiconductor based DFB lasers have the potential for improved sensitivity relative to other organic DFB lasers (such as dye-doped) as the refractive index of organic semiconductors is generally higher, which leads to an increase in the interaction of the laser mode with the analyte binding region at the laser surface. In this thesis, the rst demonstration of an organic semiconductor (oligofluorene truxene (T3)) DFB laser for biosensing applications is described. Sensor development is focused on the ultimate aim of incorporating a T3 DFB laser into a compact and portable highthroughput sensing platform for the detection of cardiac biomarkers, Apolipoprotein B100, C-reactive protein and B-type natriuretic peptide in particular. Detection of these biomarkers is to be achieved via functionalisation of the T3 surface with oligonucleotide based probes. The structure of the T3 DFB laser is optimised experimentally and theoretically by tuning the gain layer thickness to maximise sensitivity to changes in refractive index at the laser surface, such as the binding of an analyte. The optimised laser sensor has a laser threshold of 30 µJ.cm⁻²/6 kW.cm⁻² (5 ns pulse duration) which makes optical pumping with a laser diode a possibility. The sensing potential of the DFB laser is shown via the detection of bulk solution refractive index changes and the addition of biomolecules to the laser surface, where a bulk sensitivity of 22 nm per refractive index unit is observed. The specific biosensing potential of the laser is highlighted through the functionalisation of the laser surface with biotin molecules and the subsequent detection of the complementary protein, avidin. The lowest limit of avidin detection achieved is 1µg.mL⁻¹; at this level of sensitivity, the current T3 laser is expected to be able to detect the larger and more abundant of two of the three cardiac biomarker targets, ApoB and CRP. The effects of structural changes to device sensitivity are modelled theoretically and demonstrate that detection of BNP may be achieved through the addition of a high-index cladding layer, a technique currently used for dye-doped DFB lasers. The first demonstration of a DFB laser used for reversible sensing is also presented in this thesis. Through the use of desthiobiotin, a biotin analogue, reversible avidin detection is performed. A reversible biosensor may be of particular interest for applications where a large number of repeated measurements are required, and may be prohibitive to the use of single-use, disposable sensors. Finally, functionalisation of the DFB laser with oligonucleotide probes is described. Several different techniques are explored for immobilisation of oligonucleotide probes on the T3 surface, with click chemistry and sulfhydryl linkage chemistries showing the most promise.