Convolutional Neural Network-based place recognition

Recently Convolutional Neural Networks (CNNs) have been shown to achieve state-of-the-art performance on various classification tasks. In this paper, we present for the first time a place recognition technique based on CNN models, by combining the powerful features learnt by CNNs with a spatial and sequential filter. Applying the system to a 70 km benchmark place recognition dataset we achieve a 75% increase in recall at 100% precision, significantly outperforming all previous state of the art techniques. We also conduct a comprehensive performance comparison of the utility of features from all 21 layers for place recognition, both for the benchmark dataset and for a second dataset with more significant viewpoint changes.

Pollen image classification using the classifynder system: Algorithm comparison and a case study on New Zealand honey

We describe an investigation into how Massey University’s Pollen Classifynder can accelerate the understanding of pollen and its role in nature. The Classifynder is an imaging microscopy system that can locate, image and classify slide based pollen samples. Given the laboriousness of purely manual image acquisition and identification it is vital to exploit assistive technologies like the Classifynder to enable acquisition and analysis of pollen samples. It is also vital that we understand the strengths and limitations of automated systems so that they can be used (and improved) to compliment the strengths and weaknesses of human analysts to the greatest extent possible. This article reviews some of our experiences with the Classifynder system and our exploration of alternative classifier models to enhance both accuracy and interpretability. Our experiments in the pollen analysis problem domain have been based on samples from the Australian National University’s pollen reference collection (2,890 grains, 15 species) and images bundled with the Classifynder system (400 grains, 4 species). These samples have been represented using the Classifynder image feature set.We additionally work through a real world case study where we assess the ability of the system to determine the pollen make-up of samples of New Zealand honey. In addition to the Classifynder’s native neural network classifier, we have evaluated linear discriminant, support vector machine, decision tree and random forest classifiers on these data with encouraging results. Our hope is that our findings will help enhance the performance of future releases of the Classifynder and other systems for accelerating the acquisition and analysis of pollen samples.

A comparison of classification algorithms within the Classifynder pollen imaging system

We describe an investigation into how Massey University's Pollen Classifynder can accelerate the understanding of pollen and its role in nature. The Classifynder is an imaging microscopy system that can locate, image and classify slide based pollen samples. Given the laboriousness of purely manual image acquisition and identification it is vital to exploit assistive technologies like the Classifynder to enable acquisition and analysis of pollen samples. It is also vital that we understand the strengths and limitations of automated systems so that they can be used (and improved) to compliment the strengths and weaknesses of human analysts to the greatest extent possible. This article reviews some of our experiences with the Classifynder system and our exploration of alternative classifier models to enhance both accuracy and interpretability. Our experiments in the pollen analysis problem domain have been based on samples from the Australian National University's pollen reference collection (2890 grains, 15 species) and images bundled with the Classifynder system (400 grains, 4 species). These samples have been represented using the Classifynder image feature set. In addition to the Classifynder's native neural network classifier, we have evaluated linear discriminant, support vector machine, decision tree and random forest classifiers on these data with encouraging results. Our hope is that our findings will help enhance the performance of future releases of the Classifynder and other systems for accelerating the acquisition and analysis of pollen samples. © 2013 AIP Publishing LLC.

Classification of inventory data suit for managing Australian railway bridges

With a focus to optimising the life cycle performance of Australian Railway bridges, new bridge classification and environmental classification systems are proposed. The new bridge classification system is mainly to facilitate the implementation of novel Bridge Management System (BMS) which optimise the life cycle cost both at project level and network level while environment classification is mainly to improve accuracy of Remaining Service Potential (RSP) module of the proposed BMS. In fact, limited capacity of the existing BMS to trigger the maintenance intervention point is an indirect result of inadequacies of the existing bridge and environmental classification systems. The proposed bridge classification system permits to identify the intervention points based on percentage deterioration of individual elements and maintenance cost, while allowing performance based rating technique to implement for maintenance optimisation and prioritisation. Simultaneously, the proposed environment classification system will enhance the accuracy of prediction of deterioration of steel components.

Some comparison on whole-proteome phylogeny of large dsDNA viruses based on dynamical language approach and feature frequency profiles method

There has been a growing interest in alignment-free methods for phylogenetic analysis using complete genome data. Among them, CVTree method, feature frequency profiles method and dynamical language approach were used to investigate the whole-proteome phylogeny of large dsDNA viruses. Using the data set of large dsDNA viruses from Gao and Qi (BMC Evol. Biol. 2007), the phylogenetic results based on the CVTree method and the dynamical language approach were compared in Yu et al. (BMC Evol. Biol. 2010). In this paper, we first apply dynamical language approach to the data set of large dsDNA viruses from Wu et al. (Proc. Natl. Acad. Sci. USA 2009) and compare our phylogenetic results with those based on the feature frequency profiles method. Then we construct the whole-proteome phylogeny of the larger dataset combining the above two data sets. According to the report of The International Committee on the Taxonomy of Viruses (ICTV), the trees from our analyses are in good agreement to the latest classification of large dsDNA viruses.

Family-based programmes for preventing smoking by children and adolescents (Review)

Background There is evidence that family and friends influence children's decisions to smoke. Objectives To assess the effectiveness of interventions to help families stop children starting smoking. Search methods We searched 14 electronic bibliographic databases, including the Cochrane Tobacco Addiction Group specialized register, MEDLINE, EMBASE, PsycINFO, CINAHL unpublished material, and key articles' reference lists. We performed free-text internet searches and targeted searches of appropriate websites, and hand-searched key journals not available electronically. We consulted authors and experts in the field. The most recent search was 3 April 2014. There were no date or language limitations. Selection criteria Randomised controlled trials (RCTs) of interventions with children (aged 5-12) or adolescents (aged 13-18) and families to deter tobacco use. The primary outcome was the effect of the intervention on the smoking status of children who reported no use of tobacco at baseline. Included trials had to report outcomes measured at least six months from the start of the intervention. Data collection and analysis We reviewed all potentially relevant citations and retrieved the full text to determine whether the study was an RCT and matched our inclusion criteria. Two authors independently extracted study data for each RCT and assessed them for risk of bias. We pooled risk ratios using a Mantel-Haenszel fixed effect model. Main results Twenty-seven RCTs were included. The interventions were very heterogeneous in the components of the family intervention, the other risk behaviours targeted alongside tobacco, the age of children at baseline and the length of follow-up. Two interventions were tested by two RCTs, one was tested by three RCTs and the remaining 20 distinct interventions were tested only by one RCT. Twenty-three interventions were tested in the USA, two in Europe, one in Australia and one in India. The control conditions fell into two main groups: no intervention or usual care; or school-based interventions provided to all participants. These two groups of studies were considered separately. Most studies had a judgement of 'unclear' for at least one risk of bias criteria, so the quality of evidence was downgraded to moderate. Although there was heterogeneity between studies there was little evidence of statistical heterogeneity in the results. We were unable to extract data from all studies in a format that allowed inclusion in a meta-analysis. There was moderate quality evidence family-based interventions had a positive impact on preventing smoking when compared to a no intervention control. Nine studies (4810 participants) reporting smoking uptake amongst baseline non-smokers could be pooled, but eight studies with about 5000 participants could not be pooled because of insufficient data. The pooled estimate detected a significant reduction in smoking behaviour in the intervention arms (risk ratio [RR] 0.76, 95% confidence interval [CI] 0.68 to 0.84). Most of these studies used intensive interventions. Estimates for the medium and low intensity subgroups were similar but confidence intervals were wide. Two studies in which some of the 4487 participants already had smoking experience at baseline did not detect evidence of effect (RR 1.04, 95% CI 0.93 to 1.17). Eight RCTs compared a combined family plus school intervention to a school intervention only. Of the three studies with data, two RCTS with outcomes for 2301 baseline never smokers detected evidence of an effect (RR 0.85, 95% CI 0.75 to 0.96) and one study with data for 1096 participants not restricted to never users at baseline also detected a benefit (RR 0.60, 95% CI 0.38 to 0.94). The other five studies with about 18,500 participants did not report data in a format allowing meta-analysis. One RCT also compared a family intervention to a school 'good behaviour' intervention and did not detect a difference between the two types of programme (RR 1.05, 95% CI 0.80 to 1.38, n = 388). No studies identified any adverse effects of intervention. Authors' conclusions There is moderate quality evidence to suggest that family-based interventions can have a positive effect on preventing children and adolescents from starting to smoke. There were more studies of high intensity programmes compared to a control group receiving no intervention, than there were for other compairsons. The evidence is therefore strongest for high intensity programmes used independently of school interventions. Programmes typically addressed family functioning, and were introduced when children were between 11 and 14 years old. Based on this moderate quality evidence a family intervention might reduce uptake or experimentation with smoking by between 16 and 32%. However, these findings should be interpreted cautiously because effect estimates could not include data from all studies. Our interpretation is that the common feature of the effective high intensity interventions was encouraging authoritative parenting (which is usually defined as showing strong interest in and care for the adolescent, often with rule setting). This is different from authoritarian parenting (do as I say) or neglectful or unsupervised parenting.

Injury narrative text classification using factorization model

Narrative text is a useful way of identifying injury circumstances from the routine emergency department data collections. Automatically classifying narratives based on machine learning techniques is a promising technique, which can consequently reduce the tedious manual classification process. Existing works focus on using Naive Bayes which does not always offer the best performance. This paper proposes the Matrix Factorization approaches along with a learning enhancement process for this task. The results are compared with the performance of various other classification approaches. The impact on the classification results from the parameters setting during the classification of a medical text dataset is discussed. With the selection of right dimension k, Non Negative Matrix Factorization-model method achieves 10 CV accuracy of 0.93.

Classification of activities of daily living of individual with limb loss

The monitoring of the actual activities of daily living of individuals with lower limb amputation is essential for an evidence-based fitting of the prosthesis, more particularly the choice of components (e.g., knees, ankles, feet)[1-4]. The purpose of this presentation was to give an overview of the categorization of the load regime data to assess the functional output and usage of the prosthesis of lower limb amputees has presented in several publications[5, 6]. The objectives were to present a categorization of load regime and to report the results for a case.

Centroid training to achieve effective text classification

Traditional text classification technology based on machine learning and data mining techniques has made a big progress. However, it is still a big problem on how to draw an exact decision boundary between relevant and irrelevant objects in binary classification due to much uncertainty produced in the process of the traditional algorithms. The proposed model CTTC (Centroid Training for Text Classification) aims to build an uncertainty boundary to absorb as many indeterminate objects as possible so as to elevate the certainty of the relevant and irrelevant groups through the centroid clustering and training process. The clustering starts from the two training subsets labelled as relevant or irrelevant respectively to create two principal centroid vectors by which all the training samples are further separated into three groups: POS, NEG and BND, with all the indeterminate objects absorbed into the uncertain decision boundary BND. Two pairs of centroid vectors are proposed to be trained and optimized through the subsequent iterative multi-learning process, all of which are proposed to collaboratively help predict the polarities of the incoming objects thereafter. For the assessment of the proposed model, F1 and Accuracy have been chosen as the key evaluation measures. We stress the F1 measure because it can display the overall performance improvement of the final classifier better than Accuracy. A large number of experiments have been completed using the proposed model on the Reuters Corpus Volume 1 (RCV1) which is important standard dataset in the field. The experiment results show that the proposed model has significantly improved the binary text classification performance in both F1 and Accuracy compared with three other influential baseline models.

Pattern-based topic modelling and its application for information filtering and information retrieval

This thesis targets on a challenging issue that is to enhance users' experience over massive and overloaded web information. The novel pattern-based topic model proposed in this thesis can generate high-quality multi-topic user interest models technically by incorporating statistical topic modelling and pattern mining. We have successfully applied the pattern-based topic model to both fields of information filtering and information retrieval. The success of the proposed model in finding the most relevant information to users mainly comes from its precisely semantic representations to represent documents and also accurate classification of the topics at both document level and collection level.

Towards robust automatic affective classification of images using facial expressions for practical applications

Affect is an important feature of multimedia content and conveys valuable information for multimedia indexing and retrieval. Most existing studies for affective content analysis are limited to low-level features or mid-level representations, and are generally criticized for their incapacity to address the gap between low-level features and high-level human affective perception. The facial expressions of subjects in images carry important semantic information that can substantially influence human affective perception, but have been seldom investigated for affective classification of facial images towards practical applications. This paper presents an automatic image emotion detector (IED) for affective classification of practical (or non-laboratory) data using facial expressions, where a lot of “real-world” challenges are present, including pose, illumination, and size variations etc. The proposed method is novel, with its framework designed specifically to overcome these challenges using multi-view versions of face and fiducial point detectors, and a combination of point-based texture and geometry. Performance comparisons of several key parameters of relevant algorithms are conducted to explore the optimum parameters for high accuracy and fast computation speed. A comprehensive set of experiments with existing and new datasets, shows that the method is effective despite pose variations, fast, and appropriate for large-scale data, and as accurate as the method with state-of-the-art performance on laboratory-based data. The proposed method was also applied to affective classification of images from the British Broadcast Corporation (BBC) in a task typical for a practical application providing some valuable insights.

Reliable fault diagnosis for incipient low-speed bearings using fault feature analysis based on a binary bat algorithm

In this paper, we propose a highly reliable fault diagnosis scheme for incipient low-speed rolling element bearing failures. The scheme consists of fault feature calculation, discriminative fault feature analysis, and fault classification. The proposed approach first computes wavelet-based fault features, including the respective relative wavelet packet node energy and entropy, by applying a wavelet packet transform to an incoming acoustic emission signal. The most discriminative fault features are then filtered from the originally produced feature vector by using discriminative fault feature analysis based on a binary bat algorithm (BBA). Finally, the proposed approach employs one-against-all multiclass support vector machines to identify multiple low-speed rolling element bearing defects. This study compares the proposed BBA-based dimensionality reduction scheme with four other dimensionality reduction methodologies in terms of classification performance. Experimental results show that the proposed methodology is superior to other dimensionality reduction approaches, yielding an average classification accuracy of 94.9%, 95.8%, and 98.4% under bearing rotational speeds at 20 revolutions-per-minute (RPM), 80 RPM, and 140 RPM, respectively.

Systems-based approach to investigate unsafe pedestrian behaviour at level crossings

Crashes at level crossings are a major issue worldwide. In Australia, as well as in other countries, the number of crashes with vehicles has declined in the past years, while the number of crashes involving pedestrians seems to have remained unchanged. A systematic review of research related to pedestrian behaviour highlighted a number of important scientific gaps in current knowledge. The complexity of such intersections imposes particular constraints to the understanding of pedestrians’ crossing behaviour. A new systems-based framework, called Pedestrian Unsafe Level Crossing framework (PULC) was developed. The PULC organises contributing factors to crossing behaviour on different system levels as per the hierarchical classification of Jens Rasmussen’s Framework for Risk Management. In addition, the framework adapts James Reason’s classification to distinguish between different types of unsafe behaviour. The framework was developed as a tool for collection of generalizable data that could be used to predict current or future system failures or to identify aspects of the system that require further safety improvement. To give it an initial support, the PULC was applied to the analysis of qualitative data from focus groups discussions. A total number of 12 pedestrians who regularly crossed the same level crossing were asked about their daily experience and their observations of others’ behaviour which allowed the extraction and classification of factors associated with errors and violations. Two case studies using Rasmussen’s AcciMap technique are presented as an example of potential application of the framework. A discussion on the identified multiple risk contributing factors and their interactions is provided, in light of the benefits of applying a systems approach to the understanding of the origins of individual’s behaviour. Potential actions towards safety improvement are discussed.

A survey: Particle Swarm Optimization based algorithms to solve premature convergence problem

Particle Swarm Optimization (PSO) is a biologically inspired computational search and optimization method based on the social behaviors of birds flocking or fish schooling. Although, PSO is represented in solving many well-known numerical test problems, but it suffers from the premature convergence. A number of basic variations have been developed due to solve the premature convergence problem and improve quality of solution founded by the PSO. This study presents a comprehensive survey of the various PSO-based algorithms. As part of this survey, the authors have included a classification of the approaches and they have identify the main features of each proposal. In the last part of the study, some of the topics within this field that are considered as promising areas of future research are listed.

Similarity-based birdcall retrieval from environmental audio

Automated digital recordings are useful for large-scale temporal and spatial environmental monitoring. An important research effort has been the automated classification of calling bird species. In this paper we examine a related task, retrieval of birdcalls from a database of audio recordings, similar to a user supplied query call. Such a retrieval task can sometimes be more useful than an automated classifier. We compare three approaches to similarity-based birdcall retrieval using spectral ridge features and two kinds of gradient features, structure tensor and the histogram of oriented gradients. The retrieval accuracy of our spectral ridge method is 94% compared to 82% for the structure tensor method and 90% for the histogram of gradients method. Additionally, this approach potentially offers a more compact representation and is more computationally efficient.