A flexible hierarchical approach for facial age estimation based on multiple features

Age estimation from facial images is increasingly receiving attention to solve age-based access control, age-adaptive targeted marketing, amongst other applications. Since even humans can be induced in error due to the complex biological processes involved, finding a robust method remains a research challenge today. In this paper, we propose a new framework for the integration of Active Appearance Models (AAM), Local Binary Patterns (LBP), Gabor wavelets (GW) and Local Phase Quantization (LPQ) in order to obtain a highly discriminative feature representation which is able to model shape, appearance, wrinkles and skin spots. In addition, this paper proposes a novel flexible hierarchical age estimation approach consisting of a multi-class Support Vector Machine (SVM) to classify a subject into an age group followed by a Support Vector Regression (SVR) to estimate a specific age. The errors that may happen in the classification step, caused by the hard boundaries between age classes, are compensated in the specific age estimation by a flexible overlapping of the age ranges. The performance of the proposed approach was evaluated on FG-NET Aging and MORPH Album 2 datasets and a mean absolute error (MAE) of 4.50 and 5.86 years was achieved respectively. The robustness of the proposed approach was also evaluated on a merge of both datasets and a MAE of 5.20 years was achieved. Furthermore, we have also compared the age estimation made by humans with the proposed approach and it has shown that the machine outperforms humans. The proposed approach is competitive with current state-of-the-art and it provides an additional robustness to blur, lighting and expression variance brought about by the local phase features.

Diet, cell signalling, and tumourigenesis in multiple intestinal neoplasia mice

The incidence of colon cancer is high in Western societies, and in Finland it is among the three most common cancer types in both females and males. Environmental factors, including diet, affect colon cancer development. During the last few years, a vast amount of new, functional foods have been introduced to the consumers. Several products are already available that are marketed as promoting intestinal health. To be able to reliably call a dietary compound a chemopreventive substance it is of fundamental importance to understand the mechanism by which it affects tumour formation and the integrity of the epithelial cells. In this thesis, three different dietary compounds were studied in an experimental model of colon cancer. Inulin is a non-digestible fibre found naturally in chicory roots, artichokes and onions, amongst others. Nowadays it is widely used as an added dietary fibre in several food products. Conjugated linoleic acid (CLA) is a conjugated form of the fatty acid linoleic acid. CLA is formed by bacterial fermentation of linoleic acid in the rumen of cows and other ruminants. Concomitantly, it can naturally be found in milk and meat of ruminants. White currant is a colourless berry low in phenolic compounds that are believed to prevent cancer formation. Contrary to what was expected, inulin and the conjugated linoleic acid isomer trans-10, cis-12, were tumour growth promoting dietary constituents when fed to Min mice. Both diets decreased the NF-kappaB levels in the mucosa, but physiological adenoma development did not affect NF-kappaB. Diet altered beta-catenin and p53 signalling in the adenomas, confirming their involvement in adenoma growth. White currant, on the other hand, was chemopreventive, despite its low contents of phenolic compounds. The chemopreventive effect was accompanied by increased p53 levels in the mucosa, and decreased beta-catenin and NF-kappaB levels in the adenoma. This could explain the reduced adenoma number and size. The results underline the importance of carefully testing new dietary compounds in different settings to reliably confirm their health benefits. In this study two compounds that are consumed and believed to add to our health proved to be cancer promotive. A berry with low phenolic contents, on the other hand, was chemopreventive.

On Bit Steering in the Minimization of the Control Memory of Microprogrammed Processors

The microcommands constituting the microprogram of the control memory of a microprogrammed processor can be partitioned into a number of disjoint sets. Some of these sets are then encoded to minimize the word width of the ROM storing the microprogram. A further reduction in the width of the ROM words can be achieved by a technique known as bit steering where one or more bits are shared by two or more sets of microcommands. These sets are called the steerable sets. This correspondence presents a simple method for the detection and encoding of steerable sets. It has been shown that the concurrency matrix of two steerable sets exhibits definite patterns of clusters which can be easily recognized. A relation "connection" has been defined which helps in the detection of three-set steerability. Once steerable sets are identified, their encoding becomes a straightforward procedure following the location of the identifying clusters on the concurrency matrix or matrices.

Multiple liquid-crystals and NMR

The use of more than one liquid crystal solvents to determine molecular structure and conformation is discussed. Liquid crystals of similar and opposite signs of diamagnetic anisotropies are considered separately since they lead to different novel applications. Advantages of such experiments over those employing single solvents are pointed out with illustrative examples.

Porphyrins with multiple crown ether voids: novel systems for cation complexation studies

Porphyrins appended with crown ether, benzo-15-crown-5, at the methine positions have been synthesized and characterized. The fully and partially substituted porphyrins and their metallo (Co, Cu, and Zn) derivatives describe one or more ether cavities in the periphery that are capable of recognizing spherical cations. The ability of these macrocycles to complex cations (Na+, K+, Mg2+, Ca2+, Ba2', and NH4+) is investigated by use of visible, 'H NMR, ESR, and emission spectral studies. The tetrasubstituted crown porphyrin (TCP) exhibits very high selectivity for K+. The cations (K', Ba2+, and NH4+) that require two crown ether cavities for complexation promote dimerization of the porphyrins. The ESR study of the cation-induced porphyrin dimers reveals axial symmetry with the porphyrin planes separated by -4.2 A. This distance increases from the fully substituted to partially substituted porphyrins. The cations (K', Ba2+, and NH4') quench efficiently the fluorescence of the free base porphyrins and their metallo derivatives. The quenching process is attributed to the steric geometry of the dimers.

Natural Abundant Solid State NMR Studies in Designed Tripeptides for Differentiation of Multiple Conformers

Solid state NMR (SSNMR) experiments on heteronuclei in natural abundance are described for three synthetically designed tripeptides Piv-(L)Pro_(L)Pro-(L)Phe-OMe (1), Piv-(D)Pro_(L)Pro_(L)Phe-OMe (2), and Piv-(D)Pro_(L)Pro_(L)Phe-NHMe (3). These peptides exist in different conformation as shown by solution state NMR and single crystal X-ray analysis (Chatterjee et al., Chem Eur J 2008, 14, 6192). In this study, SSNMR has been used to probe the conformations of these peptides in their powder form. The C-13 spectrum of peptide (1) showed doubling of resonances corresponding to cis/cis form, unlike in solution where the similar doubling is attributed to cis/trans form. This has been confirmed by the chemical shift differences of C-beta and C-gamma carbon of Proline in peptide (1) both in solution and SSNMR. Peptide (2) and (3) provided single set of resonances which represented all transform across the di-Proline segment. The results are In agreement with the X-ray analysis. Solid state N-15 resonances, especially from Proline residues provided additional information, which is normally not observable in solution state NMR. H-1 chemical shifts are also obtained from a two-dimensional heteronuclear correlation experiment between H-1-C-13. The results confirm the utility of NMR as a useful tool for identifying different conformers in peptides in the solid state. (C) 2009 Wiley Periodicals, Inc. Biopolymers 91: 851-860, 2009.

A Lidl bit of grocery competition could cause a Coles rethink

Australian shoppers have inadvertently invited global discount grocers to our shores by demonstrating their readiness to adopt private labels. In 2001, German discounter Aldi opened its first store in Sydney. The impact this business format would have on the Australian grocery sector was underestimated.

The plasticity of NBS resistance genes in sorghum is driven by multiple evolutionary processes

Background Increased disease resistance is a key target of cereal breeding programs, with disease outbreaks continuing to threaten global food production, particularly in Africa. Of the disease resistance gene families, the nucleotide-binding site plus leucine-rich repeat (NBS-LRR) family is the most prevalent and ancient and is also one of the largest gene families known in plants. The sequence diversity in NBS-encoding genes was explored in sorghum, a critical food staple in Africa, with comparisons to rice and maize and with comparisons to fungal pathogen resistance QTL. Results In sorghum, NBS-encoding genes had significantly higher diversity in comparison to non NBS-encoding genes and were significantly enriched in regions of the genome under purifying and balancing selection, both through domestication and improvement. Ancestral genes, pre-dating species divergence, were more abundant in regions with signatures of selection than in regions not under selection. Sorghum NBS-encoding genes were also significantly enriched in the regions of the genome containing fungal pathogen disease resistance QTL; with the diversity of the NBS-encoding genes influenced by the type of co-locating biotic stress resistance QTL. Conclusions NBS-encoding genes are under strong selection pressure in sorghum, through the contrasting evolutionary processes of purifying and balancing selection. Such contrasting evolutionary processes have impacted ancestral genes more than species-specific genes. Fungal disease resistance hot-spots in the genome, with resistance against multiple pathogens, provides further insight into the mechanisms that cereals use in the “arms race” with rapidly evolving pathogens in addition to providing plant breeders with selection targets for fast-tracking the development of high performing varieties with more durable pathogen resistance.

QTL analysis in multiple sorghum populations facilitates the dissection of the genetic and physiological control of tillering

Tillering in sorghum can be associated with either the carbon supply–demand (S/D) balance of the plant or an intrinsic propensity to tiller (PTT). Knowledge of the genetic control of tillering could assist breeders in selecting germplasm with tillering characteristics appropriate for their target environments. The aims of this study were to identify QTL for tillering and component traits associated with the S/D balance or PTT, to develop a framework model for the genetic control of tillering in sorghum. Four mapping populations were grown in a number of experiments in south east Queensland, Australia. The QTL analysis suggested that the contribution of traits associated with either the S/D balance or PTT to the genotypic differences in tillering differed among populations. Thirty-four tillering QTL were identified across the populations, of which 15 were novel to this study. Additionally, half of the tillering QTL co-located with QTL for component traits. A comparison of tillering QTL and candidate gene locations identified numerous coincident QTL and gene locations across populations, including the identification of common non-synonymous SNPs in the parental genotypes of two mapping populations in a sorghum homologue of MAX1, a gene involved in the control of tiller bud outgrowth through the production of strigolactones. Combined with a framework for crop physiological processes that underpin genotypic differences in tillering, the co-location of QTL for tillering and component traits and candidate genes allowed the development of a framework QTL model for the genetic control of tillering in sorghum.

Isolation of multiple novel paramyxoviruses from pteropid bat urine

Bats have been found to harbor a number of new emerging viruses with zoonotic potential and there has been a great deal of interest in identifying novel bat pathogens to determine risk to human and animal health. Many groups have identified novel viruses in bats by detection of viral nucleic acid, however virus isolation is still a challenge and there are few reports of viral isolates from bats. In recent years, our group has developed optimized procedures for virus isolation from bat urine, including the use of primary bat cells. In previous reports we have described the isolation of Hendra virus, Menangle virus and Cedar virus, in Queensland, Australia. Here, we report the isolation of four additional novel bat paramyxoviruses from urine collected from beneath pteropid bat (flying fox) colonies in Queensland and New South Wales during 2009-2011.

Forgery attacks on ++AE authenticated encryption mode

In this paper, we analyse a block cipher mode of operation submitted in 2014 to the cryptographic competition for authenticated encryption (CAESAR). This mode is designed by Recacha and called ++AE (plus-plus-ae). We propose a chosen plaintext forgery attack on ++AE that requires only a single chosen message query to allow an attacker to construct multiple forged messages. Our attack is deterministic and guaranteed to pass ++AE integrity check. We demonstrate the forgery attack using 128-bit AES as the underlying block cipher. Hence, ++AE is insecure as an authenticated encryption mode of operation.

Multiple instance dictionary learning for activity representation

This paper presents an effective feature representation method in the context of activity recognition. Efficient and effective feature representation plays a crucial role not only in activity recognition, but also in a wide range of applications such as motion analysis, tracking, 3D scene understanding etc. In the context of activity recognition, local features are increasingly popular for representing videos because of their simplicity and efficiency. While they achieve state-of-the-art performance with low computational requirements, their performance is still limited for real world applications due to a lack of contextual information and models not being tailored to specific activities. We propose a new activity representation framework to address the shortcomings of the popular, but simple bag-of-words approach. In our framework, first multiple instance SVM (mi-SVM) is used to identify positive features for each action category and the k-means algorithm is used to generate a codebook. Then locality-constrained linear coding is used to encode the features into the generated codebook, followed by spatio-temporal pyramid pooling to convey the spatio-temporal statistics. Finally, an SVM is used to classify the videos. Experiments carried out on two popular datasets with varying complexity demonstrate significant performance improvement over the base-line bag-of-feature method.

A data fusion approach of multiple maintenance data sources for real-world reliability modelling

A central tenet in the theory of reliability modelling is the quantification of the probability of asset failure. In general, reliability depends on asset age and the maintenance policy applied. Usually, failure and maintenance times are the primary inputs to reliability models. However, for many organisations, different aspects of these data are often recorded in different databases (e.g. work order notifications, event logs, condition monitoring data, and process control data). These recorded data cannot be interpreted individually, since they typically do not have all the information necessary to ascertain failure and preventive maintenance times. This paper presents a methodology for the extraction of failure and preventive maintenance times using commonly-available, real-world data sources. A text-mining approach is employed to extract keywords indicative of the source of the maintenance event. Using these keywords, a Naïve Bayes classifier is then applied to attribute each machine stoppage to one of two classes: failure or preventive. The accuracy of the algorithm is assessed and the classified failure time data are then presented. The applicability of the methodology is demonstrated on a maintenance data set from an Australian electricity company.

Initialisation flaws in the A5-GMR-1 satphone encryption algorithm

A5-GMR-1 is a synchronous stream cipher used to provide confidentiality for communications between satellite phones and satellites. The keystream generator may be considered as a finite state machine, with an internal state of 81 bits. The design is based on four linear feedback shift registers, three of which are irregularly clocked. The keystream generator takes a 64-bit secret key and 19-bit frame number as inputs, and produces an output keystream of length between $2^8$ and $2^{10}$ bits. Analysis of the initialisation process for the keystream generator reveals serious flaws which significantly reduce the number of distinct keystreams that the generator can produce. Multiple (key, frame number) pairs produce the same keystream, and the relationship between the various pairs is easy to determine. Additionally, many of the keystream sequences produced are phase shifted versions of each other, for very small phase shifts. These features increase the effectiveness of generic time-memory tradeoff attacks on the cipher, making such attacks feasible.