Random projections on manifolds of symmetric positive definite matrices for image classification

Recent advances suggest that encoding images through Symmetric Positive Definite (SPD) matrices and then interpreting such matrices as points on Riemannian manifolds can lead to increased classification performance. Taking into account manifold geometry is typically done via (1) embedding the manifolds in tangent spaces, or (2) embedding into Reproducing Kernel Hilbert Spaces (RKHS). While embedding into tangent spaces allows the use of existing Euclidean-based learning algorithms, manifold shape is only approximated which can cause loss of discriminatory information. The RKHS approach retains more of the manifold structure, but may require non-trivial effort to kernelise Euclidean-based learning algorithms. In contrast to the above approaches, in this paper we offer a novel solution that allows SPD matrices to be used with unmodified Euclidean-based learning algorithms, with the true manifold shape well-preserved. Specifically, we propose to project SPD matrices using a set of random projection hyperplanes over RKHS into a random projection space, which leads to representing each matrix as a vector of projection coefficients. Experiments on face recognition, person re-identification and texture classification show that the proposed approach outperforms several recent methods, such as Tensor Sparse Coding, Histogram Plus Epitome, Riemannian Locality Preserving Projection and Relational Divergence Classification.

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.

Improved image set classification via joint sparse approximated nearest subspaces

Existing multi-model approaches for image set classification extract local models by clustering each image set individually only once, with fixed clusters used for matching with other image sets. However, this may result in the two closest clusters to represent different characteristics of an object, due to different undesirable environmental conditions (such as variations in illumination and pose). To address this problem, we propose to constrain the clustering of each query image set by forcing the clusters to have resemblance to the clusters in the gallery image sets. We first define a Frobenius norm distance between subspaces over Grassmann manifolds based on reconstruction error. We then extract local linear subspaces from a gallery image set via sparse representation. For each local linear subspace, we adaptively construct the corresponding closest subspace from the samples of a probe image set by joint sparse representation. We show that by minimising the sparse representation reconstruction error, we approach the nearest point on a Grassmann manifold. Experiments on Honda, ETH-80 and Cambridge-Gesture datasets show that the proposed method consistently outperforms several other recent techniques, such as Affine Hull based Image Set Distance (AHISD), Sparse Approximated Nearest Points (SANP) and Manifold Discriminant Analysis (MDA).

Preventing industrial scale using bauxite refinery residues

An Australian green power (AGP) company produces energy from burning biomass from the sugar industry and recycled wood waste, however alkali in biomass is released into a recirculating stream that forms a scale as it becomes more concentrated. This investigation has shown that the addition of Bayer liquor (alumina waste residue) successfully removes scale-forming species from the recirculating stream and thus has the potential to reduce the rate of scaling. Characterisation of the scale and Bayer precipitates has been performed using X-ray diffraction (XRD), infrared spectroscopy (IR) and inductively coupled plasma optical emission spectroscopy (ICP-OES).

Polarization of forecast densities : a new approach to time series classification

Time series classification has been extensively explored in many fields of study. Most methods are based on the historical or current information extracted from data. However, if interest is in a specific future time period, methods that directly relate to forecasts of time series are much more appropriate. An approach to time series classification is proposed based on a polarization measure of forecast densities of time series. By fitting autoregressive models, forecast replicates of each time series are obtained via the bias-corrected bootstrap, and a stationarity correction is considered when necessary. Kernel estimators are then employed to approximate forecast densities, and discrepancies of forecast densities of pairs of time series are estimated by a polarization measure, which evaluates the extent to which two densities overlap. Following the distributional properties of the polarization measure, a discriminant rule and a clustering method are proposed to conduct the supervised and unsupervised classification, respectively. The proposed methodology is applied to both simulated and real data sets, and the results show desirable properties.

A pilot study on affective classification of facial images for emerging news topics

The proliferation of news reports published in online websites and news information sharing among social media users necessitates effective techniques for analysing the image, text and video data related to news topics. This paper presents the first study to classify affective facial images on emerging news topics. The proposed system dynamically monitors and selects the current hot (of great interest) news topics with strong affective interestingness using textual keywords in news articles and social media discussions. Images from the selected hot topics are extracted and classified into three categorized emotions, positive, neutral and negative, based on facial expressions of subjects in the images. Performance evaluations on two facial image datasets collected from real-world resources demonstrate the applicability and effectiveness of the proposed system in affective classification of facial images in news reports. Facial expression shows high consistency with the affective textual content in news reports for positive emotion, while only low correlation has been observed for neutral and negative. The system can be directly used for applications, such as assisting editors in choosing photos with a proper affective semantic for a certain topic during news report preparation.

People’s perceptions about the importance of forests on Borneo

We ascertained villagers’ perceptions about the importance of forests for their livelihoods and health through 1,837 reliably answered interviews of mostly male respondents from 185 villages in Indonesian and Malaysian Borneo. Variation in these perceptions related to several environmental and social variables, as shown in classification and regression analyses. Overall patterns indicated that forest use and cultural values are highest among people on Borneo who live close to remaining forest, and especially among older Christian residents. Support for forest clearing depended strongly on the scale at which deforestation occurs. Deforestation for small-scale agriculture was generally considered to be positive because it directly benefits people’s welfare. Large-scale deforestation (e.g., for industrial oil palm or acacia plantations), on the other hand, appeared to be more context-dependent, with most respondents considering it to have overall negative impacts on them, but with people in some areas considering the benefits to outweigh the costs. The interviews indicated high awareness of negative environmental impacts of deforestation, with high levels of concern over higher temperatures, air pollution and loss of clean water sources. Our study is unique in its geographic and trans-national scale. Our findings enable the development of maps of forest use and perceptions that could inform land use planning at a range of scales. Incorporating perspectives such as these could significantly reduce conflict over forest resources and ultimately result in more equitable development processes.

Large scale read classification for next generation sequencing

Next Generation Sequencing (NGS) has revolutionised molecular biology, resulting in an explosion of data sets and an increasing role in clinical practice. Such applications necessarily require rapid identification of the organism as a prelude to annotation and further analysis. NGS data consist of a substantial number of short sequence reads, given context through downstream assembly and annotation, a process requiring reads consistent with the assumed species or species group. Highly accurate results have been obtained for restricted sets using SVM classifiers, but such methods are difficult to parallelise and success depends on careful attention to feature selection. This work examines the problem at very large scale, using a mix of synthetic and real data with a view to determining the overall structure of the problem and the effectiveness of parallel ensembles of simpler classifiers (principally random forests) in addressing the challenges of large scale genomics.

Effective routing with utilisation of cross-layer information for industrial wireless sensor networks

This thesis focuses on providing reliable data transmissions in large-scale industrial wireless sensor networks through improving network layer protocols. It addresses three major problems: scalability, dynamic industrial environments and coexistence of multiple types of data traffic in a network. Theoretical developments are conducted, followed by simulation studies for verification of theoretic results. The approach proposed in this thesis has been shown to be effective for large-scale network implementation and to provide improved data transmission reliability for both periodic and sporadic traffic.

Decomposition analysis of air pollution abatement in China: empirical study for ten industrial sectors from 1998 to 2009

This study analyzes the management of air pollutant substance in Chinese industrial sectors from 1998 to 2009. Decomposition analysis applying the logarithmic mean divisia index is used to analyze changes in emissions of air pollutants with a focus on the following five factors: coal pollution intensity (CPI), end-of-pipe treatment (EOP), the energy mix (EM), productive efficiency change (EFF), and production scale changes (PSC). Three pollutants are the main focus of this study: sulfur dioxide (SO2), dust, and soot. The novelty of this paper is focusing on the impact of the elimination policy on air pollution management in China by type of industry using the scale merit effect for pollution abatement technology change. First, the increase in SO2 emissions from Chinese industrial sectors because of the increase in the production scale is demonstrated. However, the EOP equipment that induced this change and improvements in energy efficiency has prevented an increase in SO2 emissions that is commensurate with the increase in production. Second, soot emissions were successfully reduced and controlled in all industries except the steel industry between 1998 and 2009, even though the production scale expanded for these industries. This reduction was achieved through improvements in EOP technology and in energy efficiency. Dust emissions decreased by nearly 65% between 1998 and 2009 in the Chinese industrial sectors. This successful reduction in emissions was achieved by implementing EOP technology and pollution prevention activities during the production processes, especially in the cement industry. Finally, pollution prevention in the cement industry is shown to result from production technology development rather than scale merit. © 2013 Elsevier Ltd. All rights reserved.

Determinants of eco-efficiency in the Chinese industrial sector

This study measures productive inefficiency within the context of multi-environmental pollution (eco-efficiency) in the Chinese industrial sector. The weighted Russell directional distance model is applied to measure eco-efficiency using production technology. The objective is to clarify how external factors affect eco-efficiency. The major findings are that both foreign direct investment and investment for pollution abatement improve eco-efficiency as measured by air pollutant substances. A levy system for wastewater discharge improves eco-efficiency as measured by wastewater pollutant substances. However, an air pollutant levy does not significantly affect eco-efficiency as measured by air pollutants.

Social networking and digital gaming media convergence : Classification and its consequences for appropriation

Abstract Within the field of Information Systems, a good proportion of research is concerned with the work organisation and this has, to some extent, restricted the kind of application areas given consideration. Yet, it is clear that information and communication technology deployments beyond the work organisation are acquiring increased importance in our lives. With this in mind, we offer a field study of the appropriation of an online play space known as Habbo Hotel. Habbo Hotel, as a site of media convergence, incorporates social networking and digital gaming functionality. Our research highlights the ethical problems such a dual classification of technology may bring. We focus upon a particular set of activities undertaken within and facilitated by the space – scamming. Scammers dupe members with respect to their ‘Furni’, virtual objects that have online and offline economic value. Through our analysis we show that sometimes, online activities are bracketed off from those defined as offline and that this can be related to how the technology is classified by members – as a social networking site and/or a digital game. In turn, this may affect members’ beliefs about rights and wrongs. We conclude that given increasing media convergence, the way forward is to continue the project of educating people regarding the difficulties of determining rights and wrongs, and how rights and wrongs may be acted out with respect to new technologies of play online and offline.