Matching image sets via adaptive multi convex hull

Traditional nearest points methods use all the samples in an image set to construct a single convex or affine hull model for classification. However, strong artificial features and noisy data may be generated from combinations of training samples when significant intra-class variations and/or noise occur in the image set. Existing multi-model approaches extract local models by clustering each image set individually only once, with fixed clusters used for matching with various image sets. This may not be optimal for discrimination, as undesirable environmental conditions (eg. illumination and pose variations) may result in the two closest clusters representing different characteristics of an object (eg. frontal face being compared to non-frontal face). To address the above problem, we propose a novel approach to enhance nearest points based methods by integrating affine/convex hull classification with an adapted multi-model approach. We first extract multiple local convex hulls from a query image set via maximum margin clustering to diminish the artificial variations and constrain the noise in local convex hulls. We then propose adaptive reference clustering (ARC) to constrain the clustering of each gallery image set by forcing the clusters to have resemblance to the clusters in the query image set. By applying ARC, noisy clusters in the query set can be discarded. Experiments on Honda, MoBo and ETH-80 datasets show that the proposed method outperforms single model approaches and other recent techniques, such as Sparse Approximated Nearest Points, Mutual Subspace Method and Manifold Discriminant Analysis.

Automatic classification of human epithelial type 2 cell indirect immunofluorescence images using cell pyramid matching

This paper describes a novel system for automatic classification of images obtained from Anti-Nuclear Antibody (ANA) pathology tests on Human Epithelial type 2 (HEp-2) cells using the Indirect Immunofluorescence (IIF) protocol. The IIF protocol on HEp-2 cells has been the hallmark method to identify the presence of ANAs, due to its high sensitivity and the large range of antigens that can be detected. However, it suffers from numerous shortcomings, such as being subjective as well as time and labour intensive. Computer Aided Diagnostic (CAD) systems have been developed to address these problems, which automatically classify a HEp-2 cell image into one of its known patterns (eg. speckled, homogeneous). Most of the existing CAD systems use handpicked features to represent a HEp-2 cell image, which may only work in limited scenarios. We propose a novel automatic cell image classification method termed Cell Pyramid Matching (CPM), which is comprised of regional histograms of visual words coupled with the Multiple Kernel Learning framework. We present a study of several variations of generating histograms and show the efficacy of the system on two publicly available datasets: the ICPR HEp-2 cell classification contest dataset and the SNPHEp-2 dataset.

Chemo-enzymatic synthesis of unnatural amino acids

A general chemo-enzymatic process has been developed to prepare enantiomerically pure L- and D-amino acids in high yield by deracemisation of racemic starting materials. The method has been developed from initial academic studies to be a robust, scalable industrial process. Unnatural amino acids, in high optical purity, are a rapidly growing class of intermediates required for pharmaceuticals, agrochemicals and other fine chemical applications. However, no single method has proven sufficiently adaptable to prepare these compounds generally at large scale. Our approach uses an enantioselective oxidase biocatalyst and a non-selective chemical reducing agent to effect the stereoinversion of one enantiomer and can result in an enantiomeric excess of > 99 % from a starting racemate, and product yields over 90 %. The current approach compares very favourably to resolution methods which have a maximum single pass yield of 50 %. Efficient methods have been developed to adapt the biocatalyst used in this process towards new target compounds and to optimise key factors which improve the process efficiency and offer competitive economics at scale.

Recovering Functional Integrity in Pastures Dominated by an Invasive Grass Species

Exotic grasses have been introduced in countries worldwide for pasture improvement, soil stabilisation and ornamental purposes. Some of these introductions have proven successful, but many have not (Cook & Dias 2006). In Australia, the Commonwealth Plant Introduction Scheme was initiated in 1929, and over-time introduced more than 5000 species of grasses, legumes and other forage and browse plants (Cook & Dias 2006). Lonsdale (1994) suggested that, in tropical Australia, 13% of introductions have become a problem, with only 5% being considered useful for agriculture. Low (1997) suggested that 5 out of 18 of Australia's worst tropical environmental weeds were intentionally introduced as pasture grasses. The spread and dominance of invasive grass species that degrade the quality of pastures for production can impact significantly on the livelihoods of small proprietors. Although Livestock grazing contributes only a small percentage to the world's GDP (1.5%), maintaining the long-term stability of this industry is crucial because of the high social and environmental consequence of a collapse. One billion of the world's poor are dependent on livestock grazing for food and income with this industry occupying more than 25% of the world's land base (Steinfeld et al. 2006). The ling-term sustainability of livestock grazing is also crucial for the environment. A recent FAO report attributed livestock production as a major cause of five of the most serious environmental problems: global warming, land degredation, air and water pollution, and the loss of biodiversity (Steinfeld et al. 2006). For these reasons, finding more effective approaches that guide the sustainable management of pastures is urgently needed. In Australia more than 55% of land use is for livestock grazing by sheelp and/or cattle. This land use dominate in the semi-arid and arid regions where rainfall and soil conditions are marginal for production (Commonwealth of Australia 2004). Although the level of agriculture production by conglomerates is increasing, the majority of livestock grazing within Australia remains family owned and operated (Commonwealth of Australia 2004). The sustainability of production from a grazed pasture is dependent on its botanical composition (Kemp & Dowling 1991, Kemp et al. 1996). In a grazed pasture, the dominance of an invasive grass species can impact on the functional integrity of the ecosystem, including production and nutrient cycling; wwhich will in turn, affect the income of proprietors and the ability of the system to recover from disturbance and environmental change. In Australia, $0.3 billion is spent on weed control in livestock production, but despite this substantial investment $1.9 billion is still lost in yield as a result of weeds (Sinden et al. 2004). In this paper, we adaprt a framework proposed for the restoration of degraded rainforest communities (Lamb & Gilmour 2003, Lamb et al. 2005) to compare and contrast options for recovering function integrity (i.e. a diverse set of desirable plant species that maintain key ecological processes necessary for sustainable production and nutrient cycling) within pasture communities dominated by an invasive grass species. To do this, we uase a case-study of the invasion of Eragrostis curvula (Africal lovegrss; hereafter, Lovegrass), a serious concern in Australian agricultural communities (Parsons and Cuthbertson 1992). The spread and dominance of Lovegrass is a problem because its low palatability, low nutritional content and competitiveness affect the livelihood of graziers by reducing the diversity of other plant species. We conclude by suggesting modifications to this framework for pasture ecosystems to help increase the effiency of strategies to protect functional integrity and balance social/economic and biodiversity values.

Exploitation of the menshutkin reaction for the controlled assembly of halogen bonded architectures incorporating 1,2-diiodotetrafluorobenzene and 1,3,5-Triiodotrifluorobenzene

1,4-Diazabicyclo[2.2.2]octane (DABCO) forms well-defined co-crystals with 1,2-diiodotetrafluorobenzene (1,2-DITFB), [(1,2-DITFB)2DABCO], and 1,3,5-triiodotrifluorobenzene, [(1,3,5-TITFB)2DABCO]. Both systems exhibited lower-than-expected supramolecular connectivity, which inspired a search for polymorphs in alternative crystallization solvents. In dichloromethane solution, the Menshutkin reaction was found to occur, generating chloride anions and quaternary ammonium cations through the reaction between the solvent and DABCO. The controlled in situ production of chloride ions facilitated the crystallization of new halogen bonded networks, DABCO–CH2Cl[(1,2-DITFB)Cl] (zigzag X-bonded chains) and (DABCO–CH2Cl)3[(1,3,5-TITFB)2Cl3]·CHCl3 (2D pseudo-trigonal X-bonded nets displaying Borremean entanglement), propagating with charge-assisted C–I···Cl– halogen bonds. The method was found to be versatile, and substitution of DABCO with triethylamine (TEA) gave (TEA-CH2Cl)3[(1,2-DITFB)Cl3]·4(H2O) (mixed halogen bond hydrogen bond network with 2D supramolecular connectivity) and TEA-CH2Cl[(1,3,5-TITFB)Cl] (tightly packed planar trigonal nets). The co-crystals were typically produced in high yield and purity with relatively predictable supramolecular topology, particularly with respect to the connectivity of the iodobenzene molecules. The potential to use this synthetic methodology for crystal engineering of halogen bonded architectures is demonstrated and discussed.

Predicting structural disruption of proteins caused by crossover

We present a machine learning model that predicts a structural disruption score from a protein s primary structure. SCHEMA was introduced by Frances Arnold and colleagues as a method for determining putative recombination sites of a protein on the basis of the full (PDB) description of its structure. The present method provides an alternative to SCHEMA that is able to determine the same score from sequence data only. Circumventing the need for resolving the full structure enables the exploration of yet unresolved and even hypothetical sequences for protein design efforts. Deriving the SCHEMA score from a primary structure is achieved using a two step approach: first predicting a secondary structure from the sequence and then predicting the SCHEMA score from the predicted secondary structure. The correlation coefficient for the prediction is 0.88 and indicates the feasibility of replacing SCHEMA with little loss of precision.

Facile dearomatisation of porphyrins using palladium-catalysed hydrazination: the 5,15-diiminoporphodimethenes and their redox products

The synthesis, electronic absorption and 1H NMR spectra of a suite of novel porphyrinoids derived from meso-bromoporphyrins by palladium-catalysed aminations using ethyl and tert-butylcarbazates are reported. Instead of the expected carbazate-substituted porphyrins, a facile oxidative dearomatisation of the porphyrin ring occurs in high yield, especially for the nickel(II) complexes, resulting in high yields of 5,15-diiminoporphodimethenes (DIPDs). The analogous zinc(II) and free base DIPDs were also characterised, the former by X-ray crystallography. The oxidation and reduction reactions of DIPDs and their precursor carbazate porphyrins were studied. Density Functional Theory (DFT) was used to calculate the optimised geometries and frontier molecular orbitals of DIPD Ni8c and bis(azocarboxylate) 19c, and Time Dependent DFT calculations allowed the prediction of electronic absorption spectra, whose characteristics corresponded well with those of the observed solution spectra. In the latter case, the calculated low-energy absorptions were unlike those of a typical porphyrin, due to the near-degeneracy of the highest filled frontier orbitals, and the wide energy separation between the unfilled orbitals. This feature was present in the observed spectrum.

Ecosystem management along ephemeral rivers: Trading off socio-economic water supply and vegetation conservation under flood regime uncertainty

In ecosystems driven by water availability, plant community dynamics depend on complex interactions between vegetation, hydrology, and human water resources use. Along ephemeral rivers—where water availability is erratic—vegetation and people are particularly vulnerable to changes in each other's water use. Sensible management requires that water supply be maintained for people, while preserving ecosystem health. Meeting such requirements is challenging because of the unpredictable water availability. We applied information gap decision theory to an ecohydrological system model of the Kuiseb River environment in Namibia. Our aim was to identify the robustness of ecosystem and water management strategies to uncertainties in future flood regimes along ephemeral rivers. We evaluated the trade-offs between alternative performance criteria and their robustness to uncertainty to account for both (i) human demands for water supply and (ii) reducing the risk of species extinction caused by water mining. Increasing uncertainty of flood regime parameters reduced the performance under both objectives. Remarkably, the ecological objective (species coexistence) was more sensitive to uncertainty than the water supply objective. However, within each objective, the relative performance of different management strategies was insensitive to uncertainty. The ‘best’ management strategy was one that is tuned to the competitive species interactions in the Kuiseb environment. It regulates the biomass of the strongest competitor and, thus, at the same time decreases transpiration, thereby increasing groundwater storage and reducing pressure on less dominant species. This robust mutually acceptable strategy enables species persistence without markedly reducing the water supply for humans. This study emphasises the utility of ecohydrological models for resource management of water-controlled ecosystems. Although trade-offs were identified between alternative performance criteria and their robustness to uncertain future flood regimes, management strategies were identified that help to secure an ecologically sustainable water supply.

Robots, the internet and teaching history in the age of the NBN and the Australian curriculum

Have you ever wished you were Doctor Who and could pop yourself and your students into a Tardis and teleport them to an historical event or to meet a historical figure? We all know that unfortunately time travel is not (yet) possible, but maybe student and teacher teleportation just might be – sort of. Over the past few centuries and in lieu of time travel our communities have developed museums as a means of experiencing some of our history...

Medically-attended respiratory illnesses amongst pregnant women in Brisbane, Australia

There are limited community-based data on the burden of influenza and influenza-like illnesses during pregnancy to inform disease surveillance and control. We aimed to determine the incidence of medically-attended respiratory illnesses (MARI) in pregnant women and the proportion of women who are tested for respiratory pathogens at these visits. We conducted a nested retrospective cohort study of a non-random sample of women aged ≥18 years who had a live birth in maternity units in Brisbane, Queensland, from March 2012 to October 2014. The primary outcomes were self-reported doctor visits for MARI and laboratory investigations for respiratory pathogens. Descriptive analyses were performed. Among 1202 participants, 222 (18.5%, 95%CI 16.3%-20.7%) self-reported MARI during their pregnancy. Of those with an MARI, 20.3% (45/222) self-reported a laboratory test was performed. We were able to confirm with health service providers that 46.7% (21/45) of tests were undertaken, responses from providers were not received for the remainder. Whilst one in five women in this population reported a MARI in pregnancy, only 3.7% (45/1202) reported a clinical specimen had been arranged at the consultation and the ability to validate that self-report was problematic. As the focus on maternal immunisation increases, ascertainment of the aetiological agent causing MARI in this population will be required and efficient and reliable methods for obtaining those data at the community level need to be established.