A method to quantify a descriptor's illumination variance

This paper presents a new metric, which we call the lighting variance ratio, for quantifying descriptors in terms of their variance to illumination changes. In many applications it is desirable to have descriptors that are robust to changes in illumination, especially in outdoor environments. The lighting variance ratio is useful for comparing descriptors and determining if a descriptor is lighting invariant enough for a given environment. The metric is analysed across a number of datasets, cameras and descriptors. The results show that the upright SIFT descriptor is typically the most lighting invariant descriptor.

Mobile app rating scale: A new tool for assessing the quality of health mobile apps

Background The use of mobile apps for health and well being promotion has grown exponentially in recent years. Yet, there is currently no app-quality assessment tool beyond “star”-ratings. Objective The objective of this study was to develop a reliable, multidimensional measure for trialling, classifying, and rating the quality of mobile health apps. Methods A literature search was conducted to identify articles containing explicit Web or app quality rating criteria published between January 2000 and January 2013. Existing criteria for the assessment of app quality were categorized by an expert panel to develop the new Mobile App Rating Scale (MARS) subscales, items, descriptors, and anchors. There were sixty well being apps that were randomly selected using an iTunes search for MARS rating. There were ten that were used to pilot the rating procedure, and the remaining 50 provided data on interrater reliability. Results There were 372 explicit criteria for assessing Web or app quality that were extracted from 25 published papers, conference proceedings, and Internet resources. There were five broad categories of criteria that were identified including four objective quality scales: engagement, functionality, aesthetics, and information quality; and one subjective quality scale; which were refined into the 23-item MARS. The MARS demonstrated excellent internal consistency (alpha = .90) and interrater reliability intraclass correlation coefficient (ICC = .79). Conclusions The MARS is a simple, objective, and reliable tool for classifying and assessing the quality of mobile health apps. It can also be used to provide a checklist for the design and development of new high quality health apps.

Test-retest reliability of graph theory measures of structural brain connectivity

The human connectome has recently become a popular research topic in neuroscience, and many new algorithms have been applied to analyze brain networks. In particular, network topology measures from graph theory have been adapted to analyze network efficiency and 'small-world' properties. While there has been a surge in the number of papers examining connectivity through graph theory, questions remain about its test-retest reliability (TRT). In particular, the reproducibility of structural connectivity measures has not been assessed. We examined the TRT of global connectivity measures generated from graph theory analyses of 17 young adults who underwent two high-angular resolution diffusion (HARDI) scans approximately 3 months apart. Of the measures assessed, modularity had the highest TRT, and it was stable across a range of sparsities (a thresholding parameter used to define which network edges are retained). These reliability measures underline the need to develop network descriptors that are robust to acquisition parameters.

On robust face recognition via sparse coding: The good, the bad and the ugly

In the field of face recognition, sparse representation (SR) has received considerable attention during the past few years, with a focus on holistic descriptors in closed-set identification applications. The underlying assumption in such SR-based methods is that each class in the gallery has sufficient samples and the query lies on the subspace spanned by the gallery of the same class. Unfortunately, such an assumption is easily violated in the face verification scenario, where the task is to determine if two faces (where one or both have not been seen before) belong to the same person. In this study, the authors propose an alternative approach to SR-based face verification, where SR encoding is performed on local image patches rather than the entire face. The obtained sparse signals are pooled via averaging to form multiple region descriptors, which then form an overall face descriptor. Owing to the deliberate loss of spatial relations within each region (caused by averaging), the resulting descriptor is robust to misalignment and various image deformations. Within the proposed framework, they evaluate several SR encoding techniques: l1-minimisation, Sparse Autoencoder Neural Network (SANN) and an implicit probabilistic technique based on Gaussian mixture models. Thorough experiments on AR, FERET, exYaleB, BANCA and ChokePoint datasets show that the local SR approach obtains considerably better and more robust performance than several previous state-of-the-art holistic SR methods, on both the traditional closed-set identification task and the more applicable face verification task. The experiments also show that l1-minimisation-based encoding has a considerably higher computational cost when compared with SANN-based and probabilistic encoding, but leads to higher recognition rates.

Visual fingerprints of the acoustic environment: The use of acoustic indices to characterise natural habitats

Acoustic recordings play an increasingly important role in monitoring terrestrial environments. However, due to rapid advances in technology, ecologists are accumulating more audio than they can listen to. Our approach to this big-data challenge is to visualize the content of long-duration audio recordings by calculating acoustic indices. These are statistics which describe the temporal-spectral distribution of acoustic energy and reflect content of ecological interest. We combine spectral indices to produce false-color spectrogram images. These not only reveal acoustic content but also facilitate navigation. An additional analytic challenge is to find appropriate descriptors to summarize the content of 24-hour recordings, so that it becomes possible to monitor long-term changes in the acoustic environment at a single location and to compare the acoustic environments of different locations. We describe a 24-hour ‘acoustic-fingerprint’ which shows some preliminary promise.

Assessment and rationalization of genetic diversity of Papua New Guinea taro (Colocasia esculenta) using SSR DNA fingerprinting

Taro (Colocasia esculenta) accessions were collected from 15 provinces of Papua New Guinea (PNG). The collection, totalling 859 accessions was collated for characterization and a core collection of 81 accessions (10%) was established on the basis of characterization data generated on 30 agro-morphological descriptors, and DNA fingerprinting using seven SSR primers. The selection of accessions was based on cluster analysis of the morphological data enabling initial selection of 20% accessions. The 20% sample was then reduced and rationalized to 10% based on molecular data generated by SSR primers. This represents the first national core collection of any species established in PNG based on molecular markers. The core has been integrated with core from other Pacific Island countries, contributing to a Pacific regional core collection, which is conserved in vitro in the South Pacific Regional Germplasm Centre at Fiji. The core collection is a valuable resource for food security of the South Pacific region and is currently being utilized by the breeding programmes of small Pacific Island countries to broaden the genetic base of the crop.

Strawberry breeding in a subtropical environment

Strawberry breeding aims to provide cultivars that maximise consumer satisfaction and producer profitability in a changing environment. In this paper some concepts of profitability, consumer satisfaction and sustainability are explored for a subtropical climate using Queensland Australia, and Florida USA, as examples. The typical production environment is annual autumn planting of bare rooted runners into polythene covered raised beds at about 40000 plants/ha. Harvesting is late autumn to early spring, with fruit arriving at the major markets up to 2000km away from the production area within 1-4 days of harvest. The basic premise in the breed-big work is that consumers must enjoy the experience of eating strawberries, and that perceived flavour, sweetness, and juiciness are the major contributors to this experience. Using market chain information, we developed a basic value model comprised of costs, returns, and sustainability of market. To this basic outline are applied operational descriptors, such as 'speed of harvest', and associated plant characteristics, such as 'fruit display'. The expression of each plant characteristic is ascribed a value or level and together numerically describe the phenotype. This description is mathematically manipulated to provide a 'value index' for the cultivar. Nine cultivars including 'Strawberry Festival', 'Kabarla', 'DPI Rubygem' and 'Sweet Charlie' are described, and environmental issues that may impact on the subtropical strawberry breeding objectives are discussed. Product differentiation and the use of exotic germplasm as a new source of genes for flavour and resistance to disease and environmental stress will likely be the cornerstones of future progress in subtropical strawberry breeding. This approach should satisfy both consumers and producers.

The influence of a novel simulated learning environment upon student clinical subjective refraction performance: A pilot study

Background: Optometry students are taught the process of subjective refraction through lectures and laboratory based practicals before progressing to supervised clinical practice. Simulated learning environments (SLEs) are an emerging technology that are used in a range of health disciplines, however, there is limited evidence regarding the effectiveness of clinical simulators as an educational tool. Methods: Forty optometry students (20 fourth year and 20 fifth year) were assessed twice by a qualified optometrist (two examinations separated by 4-8 weeks) while completing a monocular non-cycloplegic subjective refraction on the same patient with an unknown refractive error simulated using contact lenses. Half of the students were granted access to an online SLE, The Brien Holden Vision Institute (BHVI®) Virtual Refractor, and the remaining students formed a control group. The primary outcome measures at each visit were; accuracy of the clinical refraction compared to a qualified optometrist and relative to the Optometry Council of Australia and New Zealand (OCANZ) subjective refraction examination criteria. Secondary measures of interest included descriptors of student SLE engagement, student self-reported confidence levels and correlations between performance in the simulated and real world clinical environment. Results: Eighty percent of students in the intervention group interacted with the SLE (for an average of 100 minutes); however, there was no correlation between measures of student engagement with the BHVI® Virtual Refractor and speed or accuracy of clinical subjective refractions. Fifth year students were typically more confident and refracted more accurately and quickly than fourth year students. A year group by experimental group interaction (p = 0.03) was observed for accuracy of the spherical component of refraction, and post hoc analysis revealed that less experienced students exhibited greater gains in clinical accuracy following exposure to the SLE intervention. Conclusions: Short-term exposure to a SLE can positively influence clinical subjective refraction outcomes for less experienced optometry students and may be of benefit in increasing the skills of novice refractionists to levels appropriate for commencing supervised clinical interactions.

Direct measurement of inner prosthesis loading during activities of daily living: Proposal for classification of functional outcome

Background The purpose of this presentation is to outline the relevance of the categorization of the load regime data to assess the functional output and usage of the prosthesis of lower limb amputees. The objectives are • To highlight the need for categorisation of activities of daily living • To present a categorization of load regime applied on residuum, • To present some descriptors of the four types of activity that could be detected, • To provide an example the results for a case. Methods The load applied on the osseointegrated fixation of one transfemoral amputee was recorded using a portable kinetic system for 5 hours. The load applied on the residuum was divided in four types of activities corresponding to inactivity, stationary loading, localized locomotion and directional locomotion as detailed in previously publications. Results The periods of directional locomotion, localized locomotion, and stationary loading occurred 44%, 34%, and 22% of recording time and each accounted for 51%, 38%, and 12% of the duration of the periods of activity, respectively. The absolute maximum force during directional locomotion, localized locomotion, and stationary loading was 19%, 15%, and 8% of the body weight on the anteroposterior axis, 20%, 19%, and 12% on the mediolateral axis, and 121%, 106%, and 99% on the long axis. A total of 2,783 gait cycles were recorded. Discussion Approximately 10% more gait cycles and 50% more of the total impulse than conventional analyses were identified. The proposed categorization and apparatus have the potential to complement conventional instruments, particularly for difficult cases.

Rhythm Analysis Using Vectorcardiograms

Fourier shape descriptors of vectorcardiograms have been proposed for cardiac rhythm analysis. The technique characterizes the differences in shape and size of the normal and abnormal vectorcardiograms. The specific abnormalities considered are premature ventricular contractions (PVC's) and supraventricular premature contractions (SVPC's).

Defining the unique flavours of Australian native foods

Australian native plant foods provide new and exciting eating experiences for consumers and have the potential to re-position ‘Australian cuisine’ as a contemporary food choice for consumers worldwide. The development of a common set of flavour and aroma descriptors and characteristics was identified as a key priority for the Australian native food industry. This research assists in the development and supply of product information to support market access and market growth for this emerging industry. This work was targeted so that a concise, consistent and accurate marketing message of the flavours of these ingredients could be delivered to customers. This report details the results of the development of the first ‘Australian native flavour wheel’ and sensory descriptions for sixteen of the key commercial native food species including fruits, berries, herbs, spices and seeds.

Eliciting and encoding expert knowledge on variable selection into classical or Bayesian species distribution models

The quality of species distribution models (SDMs) relies to a large degree on the quality of the input data, from bioclimatic indices to environmental and habitat descriptors (Austin, 2002). Recent reviews of SDM techniques, have sought to optimize predictive performance e.g. Elith et al., 2006. In general SDMs employ one of three approaches to variable selection. The simplest approach relies on the expert to select the variables, as in environmental niche models Nix, 1986 or a generalized linear model without variable selection (Miller and Franklin, 2002). A second approach explicitly incorporates variable selection into model fitting, which allows examination of particular combinations of variables. Examples include generalized linear or additive models with variable selection (Hastie et al. 2002); or classification trees with complexity or model based pruning (Breiman et al., 1984, Zeileis, 2008). A third approach uses model averaging, to summarize the overall contribution of a variable, without considering particular combinations. Examples include neural networks, boosted or bagged regression trees and Maximum Entropy as compared in Elith et al. 2006. Typically, users of SDMs will either consider a small number of variable sets, via the first approach, or else supply all of the candidate variables (often numbering more than a hundred) to the second or third approaches. Bayesian SDMs exist, with several methods for eliciting and encoding priors on model parameters (see review in Low Choy et al. 2010). However few methods have been published for informative variable selection; one example is Bayesian trees (O’Leary 2008). Here we report an elicitation protocol that helps makes explicit a priori expert judgements on the quality of candidate variables. This protocol can be flexibly applied to any of the three approaches to variable selection, described above, Bayesian or otherwise. We demonstrate how this information can be obtained then used to guide variable selection in classical or machine learning SDMs, or to define priors within Bayesian SDMs.

Occupational determinants of pancreatic cancer

Objective and background. Tobacco smoking, pancreatitis and diabetes mellitus are the only known causes of pancreatic cancer, leaving ample room for yet unidentified determinants. This is an empirical study on a Finnish data on occupational exposures and pancreatic cancer risk, and a non-Bayesian and a hierarchical Bayesian meta-analysis of data on occupational factors and pancreatic cancer. Methods. The case-control study analyzed 595 incident cases of pancreatic cancer and 1,622 controls of stomach, colon, and rectum cancer, diagnosed 1984-1987 and known to be dead by 1990 in Finland. The next-of-kin responded to a mail questionnaire on job and medical histories and lifestyles. Meta-analysis of occupational risk factors of pancreatic cancer started off with 1,903 identified studies. The analyses were based on different subsets of that database. Five epidemiologists examined the reports and extracted the pertinent data using a standardized extraction form that covered 20 study descriptors and the relevant relative risk estimates. Random effects meta-analyses were applied for 23 chemical agents. In addition, hierarchical Bayesian models for meta-analysis were applied to the occupational data of 27 job titles using job exposure matrix as a link matrix and estimating the relative risks of pancreatic cancer associated with nine occupational agents. Results. In the case-control study, logistic regressions revealed excess risks of pancreatic cancer associated with occupational exposures to ionizing radiation, nonchlorinated solvents, and pesticides. Chlorinated hydrocarbon solvents and related compounds, used mainly in metal degreasing and dry cleaning, are emerging as likely risk factors of pancreatic cancer in the non-Bayesian and the hierarchical Bayesian meta-analysis. Consistent excess risk was found for insecticides, and a high excess for nickel and nickel compounds in the random effects meta-analysis but not in the hierarchical Bayesian meta-analysis. Conclusions. In this study occupational exposure to chlorinated hydrocarbon solvents and related compounds and insecticides increase risk of pancreatic cancer. Hierarchical Bayesian meta-analysis is applicable when studies addressing the agent(s) under study are lacking or very few, but several studies address job titles with potential exposure to these agents. A job-exposure matrix or a formal expert assessment system is necessary in this situation.

Class-specific sparse codes for representing activities

In this paper we investigate the effectiveness of class specific sparse codes in the context of discriminative action classification. The bag-of-words representation is widely used in activity recognition to encode features, and although it yields state-of-the art performance with several feature descriptors it still suffers from large quantization errors and reduces the overall performance. Recently proposed sparse representation methods have been shown to effectively represent features as a linear combination of an over complete dictionary by minimizing the reconstruction error. In contrast to most of the sparse representation methods which focus on Sparse-Reconstruction based Classification (SRC), this paper focuses on a discriminative classification using a SVM by constructing class-specific sparse codes for motion and appearance separately. Experimental results demonstrates that separate motion and appearance specific sparse coefficients provide the most effective and discriminative representation for each class compared to a single class-specific sparse coefficients.

Spatial ecology of food webs : herbivore-parasitoid communities on the pedunculate oak

Herbivorous insects, their host plants and natural enemies form the largest and most species-rich communities on earth. But what forces structure such communities? Do they represent random collections of species, or are they assembled by given rules? To address these questions, food webs offer excellent tools. As a result of their versatile information content, such webs have become the focus of intensive research over the last few decades. In this thesis, I study herbivore-parasitoid food webs from a new perspective: I construct multiple, quantitative food webs in a spatially explicit setting, at two different scales. Focusing on food webs consisting of specialist herbivores and their natural enemies on the pedunculate oak, Quercus robur, I examine consistency in food web structure across space and time, and how landscape context affects this structure. As an important methodological development, I use DNA barcoding to resolve potential cryptic species in the food webs, and to examine their effect on food web structure. I find that DNA barcoding changes our perception of species identity for as many as a third of the individuals, by reducing misidentifications and by resolving several cryptic species. In terms of the variation detected in food web structure, I find surprising consistency in both space and time. From a spatial perspective, landscape context leaves no detectable imprint on food web structure, while species richness declines significantly with decreasing connectivity. From a temporal perspective, food web structure remains predictable from year to year, despite considerable species turnover in local communities. The rate of such turnover varies between guilds and species within guilds. The factors best explaining these observations are abundant and common species, which have a quantitatively dominant imprint on overall structure, and suffer the lowest turnover. By contrast, rare species with little impact on food web structure exhibit the highest turnover rates. These patterns reveal important limitations of modern metrics of quantitative food web structure. While they accurately describe the overall topology of the web and its most significant interactions, they are disproportionately affected by species with given traits, and insensitive to the specific identity of species. As rare species have been shown to be important for food web stability, metrics depicting quantitative food web structure should then not be used as the sole descriptors of communities in a changing world. To detect and resolve the versatile imprint of global environmental change, one should rather use these metrics as one tool among several.