Hierarchically grouped 2D local features applied to edge contour localisation

One of the most significant research topics in computer vision is object detection. Most of the reported object detection results localise the detected object within a bounding box, but do not explicitly label the edge contours of the object. Since object contours provide a fundamental diagnostic of object shape, some researchers have initiated work on linear contour feature representations for object detection and localisation. However, linear contour feature-based localisation is highly dependent on the performance of linear contour detection within natural images, and this can be perturbed significantly by a cluttered background. In addition, the conventional approach to achieving rotation-invariant features is to rotate the feature receptive field to align with the local dominant orientation before computing the feature representation. Grid resampling after rotation adds extra computational cost and increases the total time consumption for computing the feature descriptor. Though it is not an expensive process if using current computers, it is appreciated that if each step of the implementation is faster to compute especially when the number of local features is increasing and the application is implemented on resource limited ”smart devices”, such as mobile phones, in real-time. Motivated by the above issues, a 2D object localisation system is proposed in this thesis that matches features of edge contour points, which is an alternative method that takes advantage of the shape information for object localisation. This is inspired by edge contour points comprising the basic components of shape contours. In addition, edge point detection is usually simpler to achieve than linear edge contour detection. Therefore, the proposed localization system could avoid the need for linear contour detection and reduce the pathological disruption from the image background. Moreover, since natural images usually comprise many more edge contour points than interest points (i.e. corner points), we also propose new methods to generate rotation-invariant local feature descriptors without pre-rotating the feature receptive field to improve the computational efficiency of the whole system. In detail, the 2D object localisation system is achieved by matching edge contour points features in a constrained search area based on the initial pose-estimate produced by a prior object detection process. The local feature descriptor obtains rotation invariance by making use of rotational symmetry of the hexagonal structure. Therefore, a set of local feature descriptors is proposed based on the hierarchically hexagonal grouping structure. Ultimately, the 2D object localisation system achieves a very promising performance based on matching the proposed features of edge contour points with the mean correct labelling rate of the edge contour points 0.8654 and the mean false labelling rate 0.0314 applied on the data from Amsterdam Library of Object Images (ALOI). Furthermore, the proposed descriptors are evaluated by comparing to the state-of-the-art descriptors and achieve competitive performances in terms of pose estimate with around half-pixel pose error.

An investigation into the role of interleukin-6 in linking systemic and local inflammatory responses in patients with colorectal cancer

Colorectal cancer is the second most common cause of cancer death in the UK. It is accepted that both tumour and host factors are important determinants of disease progression and survival. While systemic and local inflammatory responses are increasingly recognized to be of particular importance the understanding of the mechanisms linking these important inflammatory processes remains unclear. This thesis examines the prognostic importance of measures of systemic and local inflammation and proposes a hypothesis for a link between tumour necrosis, systemic and local inflammatory responses in patients with colorectal cancer. Chapter 3 reports the comparison of the prognostic value of longitudinal measurements of systemic inflammation in patients undergoing curative resection of colorectal cancer. In Chapter 3 the results demonstrate that there was no significant overall change in either mGPS or NLR from pre- to post- operatively. This study highlighted the associations between pre- and post- operative mGPS and NLR and T-stage (p<0.001), TNM stage (p<0.005) and cancer-specific survival. The relationships between pre-operative measurements were examined using multivariate analysis. For pre-operative measurement both mGPS and NLR were associated with cancer-specific survival while when post-operative measures were examined only mGPS was specifically associated with cancer-specific survival (HR 4.81, CI 2.13-10.83, P<0.001). Chapter 4 examines the prognostic value of the Klintrup-Makinen scoring method and the existing limitations with regard to its clinical utility. An automated scoring method using commercially available image analysis software was developed and compared with manual scoring of tumour inflammatory infiltrates. This study demonstrated that both manual K-M scoring (p<0.001) and automated K-M scoring (p<0.05) had prognostic value in patients who had undergone potentially curative resection of colorectal cancer, and that the novel automated method may provide an objective method of assessment of tumour inflammatory infiltrates using routinely stained haematoxylin and eosin sections of tumour samples. In chapter 5 a hypothesis was proposed that Interleukin-6 may link tumour necrosis and systemic and local inflammatory responses in patients with colorectal cancer. This chapter examined the basis for this hypothesis, which is presented in figure 5.1. In addition, in chapter 5 the importance of this potential link is examined. In chapter 6, the hypothesis outlined in chapter 5 was examined in a cohort of patients who had undergone attempted curative resection of colorectal cancer. This study examined the inter-relationships between circulating mediators, in particular IL-6, tumour necrosis and systemic and local inflammatory responses. This results of this study demonstrated that IL-6 was associated with tumour necrosis (<0.001) and mGPS (<0.001) independent of T-stage. Thus adding weight to the hypothesis that elevated circulating concentrations of IL-6 may play a role in modulating both the systemic and local inflammatory responses in patients with cancer. Chapter 7 further develops the hypothesis that IL-6 signalling may be important in modulating systemic and local inflammatory responses in patients with colorectal cancer. Further, in chapter 7 the basis for the role of trans-signalling in this signaling pathway was examined. In this study, we reported that neither expression of the soluble IL-6 receptor or soluble gp130 were associated with systemic or local inflammatory responses. As a result the possible reasons for these findings were explored and future work suggested. A prospective database of patients undergoing attempted curative resection of colorectal cancer in Glasgow Royal Infirmary was used throughout this thesis. This database was created and is maintained regularly by successive research fellows at the Royal Infirmary. The work presented in this thesis highlights the importance of the host response in the form of systemic and local inflammation in patients with colorectal cancer and proposes a link between these responses and tumour necrosis. In addition, this work adds weight to the body of evidence suggesting that assessment of these host responses may improve stratification to treatment for patients with colorectal cancer. Further, this work proposes a mechanistic link, between tumour necrosis, systemic and local inflammatory responses through Interleukin-6, that merits further investigation.

Transmission electron tomography: quality assessment and enhancement for three-dimensional imaging of nanostructures

Nanotechnology has revolutionised humanity's capability in building microscopic systems by manipulating materials on a molecular and atomic scale. Nan-osystems are becoming increasingly smaller and more complex from the chemical perspective which increases the demand for microscopic characterisation techniques. Among others, transmission electron microscopy (TEM) is an indispensable tool that is increasingly used to study the structures of nanosystems down to the molecular and atomic scale. However, despite the effectivity of this tool, it can only provide 2-dimensional projection (shadow) images of the 3D structure, leaving the 3-dimensional information hidden which can lead to incomplete or erroneous characterization. One very promising inspection method is Electron Tomography (ET), which is rapidly becoming an important tool to explore the 3D nano-world. ET provides (sub-)nanometer resolution in all three dimensions of the sample under investigation. However, the fidelity of the ET tomogram that is achieved by current ET reconstruction procedures remains a major challenge. This thesis addresses the assessment and advancement of electron tomographic methods to enable high-fidelity three-dimensional investigations. A quality assessment investigation was conducted to provide a quality quantitative analysis of the main established ET reconstruction algorithms and to study the influence of the experimental conditions on the quality of the reconstructed ET tomogram. Regular shaped nanoparticles were used as a ground-truth for this study. It is concluded that the fidelity of the post-reconstruction quantitative analysis and segmentation is limited, mainly by the fidelity of the reconstructed ET tomogram. This motivates the development of an improved tomographic reconstruction process. In this thesis, a novel ET method was proposed, named dictionary learning electron tomography (DLET). DLET is based on the recent mathematical theorem of compressed sensing (CS) which employs the sparsity of ET tomograms to enable accurate reconstruction from undersampled (S)TEM tilt series. DLET learns the sparsifying transform (dictionary) in an adaptive way and reconstructs the tomogram simultaneously from highly undersampled tilt series. In this method, the sparsity is applied on overlapping image patches favouring local structures. Furthermore, the dictionary is adapted to the specific tomogram instance, thereby favouring better sparsity and consequently higher quality reconstructions. The reconstruction algorithm is based on an alternating procedure that learns the sparsifying dictionary and employs it to remove artifacts and noise in one step, and then restores the tomogram data in the other step. Simulation and real ET experiments of several morphologies are performed with a variety of setups. Reconstruction results validate its efficiency in both noiseless and noisy cases and show that it yields an improved reconstruction quality with fast convergence. The proposed method enables the recovery of high-fidelity information without the need to worry about what sparsifying transform to select or whether the images used strictly follow the pre-conditions of a certain transform (e.g. strictly piecewise constant for Total Variation minimisation). This can also avoid artifacts that can be introduced by specific sparsifying transforms (e.g. the staircase artifacts the may result when using Total Variation minimisation). Moreover, this thesis shows how reliable elementally sensitive tomography using EELS is possible with the aid of both appropriate use of Dual electron energy loss spectroscopy (DualEELS) and the DLET compressed sensing algorithm to make the best use of the limited data volume and signal to noise inherent in core-loss electron energy loss spectroscopy (EELS) from nanoparticles of an industrially important material. Taken together, the results presented in this thesis demonstrates how high-fidelity ET reconstructions can be achieved using a compressed sensing approach.

Wavelengths in bioconvection patterns

Unicellular bottom-heavy swimming microorganisms are usually denser than the fluid in which they swim. In shallow suspensions, the bottom heaviness results in a gravitational torque that orients the cells to swim vertically upwards in the absence of fluid flow. Swimming cells thus accumulate at the upper surface to form a concentrated layer of cells. When the cell concentration is high enough, the layer overturns to form bioconvection patterns. Thin concentrated plumes of cells descend rapidly and cells return to the upper surface in wide, slowly moving upwelling plumes. When there is fluid flow, a second viscous torque is exerted on the swimming cells. The balance between the local shear flow viscous and the gravitational torques determines the cells' swimming direction, (gyrotaxis). In this thesis, the wavelengths of bioconvection patterns are studied experimentally as well as theoretically as follow; First, in aquasystem it is rare to find one species lives individually and when they swim they can form complex patterns. Thus, a protocol for controlled experiments to mix two species of swimming algal cells of \emph{C. rienhardtii} and \emph{C. augustae} is systematically described and images of bioconvection patterns are captured. A method for analysing images using wavelets and extracting the local dominant wavelength in spatially varying patterns is developed. The variation of the patterns as a function of the total concentration and the relative concentration between two species is analysed. Second, the linear stability theory of bioconvection for a suspension of two mixed species is studied. The dispersion relationship is computed using Fourier modes in order to calculate the neutral curves as a function of wavenumbers $k$ and $m$. The neutral curves are plotted to compare the instability onset of the suspension of the two mixed species with the instability onset of each species individually. This study could help us to understand which species contributes the most in the process of pattern formation. Finally, predicting the most unstable wavelength was studied previously around a steady state equilibrium situation. Since assuming steady state equilibrium contradicts with reality, the pattern formation in a layer of finite depth of an evolving basic state is studied using the nonnormal modes approach. The nonnormal modes procedure identifies the optimal initial perturbation that can be obtained for a given time $t$ as well as a given set of parameters and wavenumber $k$. Then, we measure the size of the optimal perturbation as it grows with time considering a range of wavenumbers for the same set of parameters to be able to extract the most unstable wavelength.

Microwave-assisted synthesis and local analyses of positive insertion electrodes for Li-ion batteries

Efficient energy storage holds the key to reducing waste energy and enabling the use of advanced handheld electronic devices, hydrid electric vehicles and residential energy storage. Recently, Li-ion batteries have been identified and employed as energy storage devices due to their high gravimetric and volumetric energy densities, in comparison to previous technologies. However, more research is required to enhance the efficiency of Li-ion batteries by discovering electrodes with larger electrochemical discharge capacities, while maintaining electrochemical stability. The aims of this study are to develop new microwave-assisted synthesis routes to nanostructured insertion cathodes, which harbor a greater affinity for lithium extraction and insertion than bulk materials. Subsequent to this, state-of-the-art synchrotron based techniques have been employed to understand structural and dynamic behaviour of nanostructured cathode materials during battery cell operation. In this study, microwave-assisted routes to a-LiFePO4, VO2(B), V3O7, H2V3O8 and V4O6(OH)4 have all been developed. Muon spin relaxation has shown that the presence of b-LiFePO4 has a detrimental effect on the lithium diffusion properties of a-LiFePO4, in agreement with first principles calculations. For the first time, a-LiFePO4 nanostructures have been obtained by employing a deep eutectic solvent reaction media showing near theoretical capacity (162 mAh g–1). Studies on VO2(B) have shown that the discharge capacity obtained is linked to the synthesis method. Electrochemical studies of H2V3O8 nanowires have shown outstanding discharge capacities (323 mAh g–1 at 100 mA g–1) and rate capability (180 mAh g–1 at 1 A g–1). The electrochemcial properties of V4O6(OH)4 have been investigated for the first time and show a promising discharge capacity of (180 mAh g–1). Lastly, in situ X-ray absorption spectroscopy has been utilised to track the evolution of the oxidation states in a-LiFePO4, VO2(B) and H2V3O8, and has shown these can all be observed dynamically.