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.

The Library beyond (just) the book: Collaboration and co-working spaces for new technology and new technological practices

Libraries have often been first adopters of many new technological innovations, such as, punch cards, computers, barcodes, and e-book readers. It is thus not surprising that many libraries have embraced the advent of the internet as an opportunity to move away from just being repositories of books, towards becoming ideas stores and local network hubs for entrepreneurial thinking and new creative practices. This presentation will look at the case of “The Edge” – an initiative of the State Library of Queensland in Brisbane, Australia, to establish a digital culture centre and learning environment deliberately designed for the co-creation and co-construction of knowledge. This initiative illustrates the potential role of libraries as testing grounds for new technologies and technological practices, which is particularly relevant in the context of the NBN rollout across Australia. It also provides an example of new engagement strategies for innovative co-working spaces that are a vital element in a trend that sees professionals, creatives and designers leave their traditional places of work and embrace the city as their office.

Leveraging the repository infrastructure to curate institutional heritage materials, learning objects and data assets

Developing and maintaining a successful institutional repository for research publications requires a considerable investment by the institution. Most of the money is spent on developing the skill-sets of existing staff or hiring new staff with the necessary skills. The return on this investment can be magnified by using this valuable infrastructure to curate collections of other materials such as learning objects, student work, conference proceedings and institutional or local community heritage materials. When Queensland University of Technology (QUT) implemented its repository for research publications (QUT ePrints) over 11 years ago, it was one of the first institutional repositories to be established in Australia. Currently, the repository holds over 29,000 open access research publications and the cumulative total number of full-text downloads for these document now exceeds 16 million. The full-text deposit rate for recently-published peer reviewed papers (currently over 74%) shows how well the repository has been embraced by QUT researchers. The success of QUT ePrints has resulted in requests to accommodate a plethora of materials which are ‘out of scope’ for this repository. QUT Library saw this as an opportunity to use its repository infrastructure (software, technical know-how and policies) to develop and implement a metadata repository for its research datasets (QUT Research Data Finder), a repository for research-related software (QUT Software Finder) and to curate a number of digital collections of institutional and local community heritage materials (QUT Digital Collections). This poster describes the repositories and digital collections curated by QUT Library and outlines the value delivered to the institution, and the wider community, by these initiatives.

A normal-distribution based rating aggregation method for generating product reputations

With the extensive use of rating systems in the web, and their significance in decision making process by users, the need for more accurate aggregation methods has emerged. The Naïve aggregation method, using the simple mean, is not adequate anymore in providing accurate reputation scores for items [6 ], hence, several researches where conducted in order to provide more accurate alternative aggregation methods. Most of the current reputation models do not consider the distribution of ratings across the different possible ratings values. In this paper, we propose a novel reputation model, which generates more accurate reputation scores for items by deploying the normal distribution over ratings. Experiments show promising results for our proposed model over state-of-the-art ones on sparse and dense datasets.

Bayesian computational methods for spatial analysis of images

This thesis introduces a new way of using prior information in a spatial model and develops scalable algorithms for fitting this model to large imaging datasets. These methods are employed for image-guided radiation therapy and satellite based classification of land use and water quality. This study has utilized a pre-computation step to achieve a hundredfold improvement in the elapsed runtime for model fitting. This makes it much more feasible to apply these models to real-world problems, and enables full Bayesian inference for images with a million or more pixels.

Open access spatial data and collaboration

On 19 June 2015, representatives from over 40 Australian research institutions gathered in Canberra to launch their Open Data Collections. The one day event, hosted by the Australian National Data Service (ANDS), showcased to government and a range of national stakeholders the rich variety of data collections that have been generated through the Major Open Data Collections (MODC) project. Colin Eustace attended the showcase for QUT Library and presented a poster that reflected the work that he and Jodie Vaughan generated through the project. QUT’s Blueprint 4, the University’s five-year institutional strategic plan, outlines the key priorities of developing a commitment to working in partnership with industry, as well as combining disciplinary strengths with interdisciplinary application. The Division of Technology, Information and Learning Support (TILS) has undertaken a number of Australian National Data Service (ANDS) funded projects since 2009 with the aim of developing improved research data management services within the University to support these strategic aims. By leveraging existing tools and systems developed during these projects, the Major Open Data Collection (MODC) project delivered support to multi-disciplinary collaborative research activities through partnership building between QUT researchers and Queensland government agencies, in order to add to and promote the discovery and reuse of a collection of spatially referenced datasets. The MODC project built upon existing Research Data Finder infrastructure (which uses VIVO open source software, developed by Cornell University) to develop a separate collection, Spatial Data Finder (https://researchdatafinder.qut.edu.au/spatial) as the interface to display the spatial data collection. During the course of the project, 62 dataset descriptions were added to Spatial Data Finder, 7 added to Research Data Finder and two added to Software Finder, another separate collection. The project team met with 116 individual researchers and attended 13 school and faculty meetings to promote the MODC project and raise awareness of the Library’s services and resources for research data management.

Investigating in-domain data requirements for PLDA training

This paper analyzes the limitations upon the amount of in- domain (NIST SREs) data required for training a probabilistic linear discriminant analysis (PLDA) speaker verification system based on out-domain (Switchboard) total variability subspaces. By limiting the number of speakers, the number of sessions per speaker and the length of active speech per session available in the target domain for PLDA training, we investigated the relative effect of these three parameters on PLDA speaker verification performance in the NIST 2008 and NIST 2010 speaker recognition evaluation datasets. Experimental results indicate that while these parameters depend highly on each other, to beat out-domain PLDA training, more than 10 seconds of active speech should be available for at least 4 sessions/speaker for a minimum of 800 speakers. If further data is available, considerable improvement can be made over solely out-domain PLDA training.

The QUT-NOISE-SRE protocol for the evaluation of noisy speaker recognition

The QUT-NOISE-SRE protocol is designed to mix the large QUT-NOISE database, consisting of over 10 hours of back- ground noise, collected across 10 unique locations covering 5 common noise scenarios, with commonly used speaker recognition datasets such as Switchboard, Mixer and the speaker recognition evaluation (SRE) datasets provided by NIST. By allowing common, clean, speech corpora to be mixed with a wide variety of noise conditions, environmental reverberant responses, and signal-to-noise ratios, this protocol provides a solid basis for the development, evaluation and benchmarking of robust speaker recognition algorithms, and is freely available to download alongside the QUT-NOISE database. In this work, we use the QUT-NOISE-SRE protocol to evaluate a state-of-the-art PLDA i-vector speaker recognition system, demonstrating the importance of designing voice-activity-detection front-ends specifically for speaker recognition, rather than aiming for perfect coherence with the true speech/non-speech boundaries.

EVpedia: A community web portal for extracellular vesicles research

Motivation Extracellular vesicles (EVs) are spherical bilayered proteolipids, harboring various bioactive molecules. Due to the complexity of the vesicular nomenclatures and components, online searches for EV-related publications and vesicular components are currently challenging. Results We present an improved version of EVpedia, a public database for EVs research. This community web portal contains a database of publications and vesicular components, identification of orthologous vesicular components, bioinformatic tools and a personalized function. EVpedia includes 6879 publications, 172 080 vesicular components from 263 high-throughput datasets, and has been accessed more than 65 000 times from more than 750 cities. In addition, about 350 members from 73 international research groups have participated in developing EVpedia. This free web-based database might serve as a useful resource to stimulate the emerging field of EV research. Availability and implementation The web site was implemented in PHP, Java, MySQL and Apache, and is freely available at http://evpedia.info.

Tesla motors’ open source revolution: Intellectual property and the carbon crisis

On the 12th June 2014, Elon Musk, the chief executive officer of the electric car manufacturer, Tesla Motors, announced in a blog that ‘all our patents belong to you.’ He explained that the company would adopt an open source philosophy in respect of its intellectual property in order to encourage the development of the industry of electric cars, and address the carbon crisis. Elon Musk made the dramatic, landmark announcement: Yesterday, there was a wall of Tesla patents in the lobby of our Palo Alto headquarters. That is no longer the case. They have been removed, in the spirit of the open source movement, for the advancement of electric vehicle technology.

Active learning: A step towards automating medical concept extraction

Objective This paper presents an automatic active learning-based system for the extraction of medical concepts from clinical free-text reports. Specifically, (1) the contribution of active learning in reducing the annotation effort, and (2) the robustness of incremental active learning framework across different selection criteria and datasets is determined. Materials and methods The comparative performance of an active learning framework and a fully supervised approach were investigated to study how active learning reduces the annotation effort while achieving the same effectiveness as a supervised approach. Conditional Random Fields as the supervised method, and least confidence and information density as two selection criteria for active learning framework were used. The effect of incremental learning vs. standard learning on the robustness of the models within the active learning framework with different selection criteria was also investigated. Two clinical datasets were used for evaluation: the i2b2/VA 2010 NLP challenge and the ShARe/CLEF 2013 eHealth Evaluation Lab. Results The annotation effort saved by active learning to achieve the same effectiveness as supervised learning is up to 77%, 57%, and 46% of the total number of sequences, tokens, and concepts, respectively. Compared to the Random sampling baseline, the saving is at least doubled. Discussion Incremental active learning guarantees robustness across all selection criteria and datasets. The reduction of annotation effort is always above random sampling and longest sequence baselines. Conclusion Incremental active learning is a promising approach for building effective and robust medical concept extraction models, while significantly reducing the burden of manual annotation.

Exhaustive search of the SNP-sNP interactome identifies epistatic effects on brain volume in two cohorts

The SNP-SNP interactome has rarely been explored in the context of neuroimaging genetics mainly due to the complexity of conducting approximately 10(11) pairwise statistical tests. However, recent advances in machine learning, specifically the iterative sure independence screening (SIS) method, have enabled the analysis of datasets where the number of predictors is much larger than the number of observations. Using an implementation of the SIS algorithm (called EPISIS), we used exhaustive search of the genome-wide, SNP-SNP interactome to identify and prioritize SNPs for interaction analysis. We identified a significant SNP pair, rs1345203 and rs1213205, associated with temporal lobe volume. We further examined the full-brain, voxelwise effects of the interaction in the ADNI dataset and separately in an independent dataset of healthy twins (QTIM). We found that each additional loading in the epistatic effect was associated with approximately 5% greater brain regional brain volume (a protective effect) in both the ADNI and QTIM samples.

Multi-site genetic analysis of diffusion images and voxelwise heritability analysis: A pilot project of the ENIGMA-DTI working group

The ENIGMA (Enhancing NeuroImaging Genetics through Meta-Analysis) Consortium was set up to analyze brain measures and genotypes from multiple sites across the world to improve the power to detect genetic variants that influence the brain. Diffusion tensor imaging (DTI) yields quantitative measures sensitive to brain development and degeneration, and some common genetic variants may be associated with white matter integrity or connectivity. DTI measures, such as the fractional anisotropy (FA) of water diffusion, may be useful for identifying genetic variants that influence brain microstructure. However, genome-wide association studies (GWAS) require large populations to obtain sufficient power to detect and replicate significant effects, motivating a multi-site consortium effort. As part of an ENIGMA-DTI working group, we analyzed high-resolution FA images from multiple imaging sites across North America, Australia, and Europe, to address the challenge of harmonizing imaging data collected at multiple sites. Four hundred images of healthy adults aged 18-85 from four sites were used to create a template and corresponding skeletonized FA image as a common reference space. Using twin and pedigree samples of different ethnicities, we used our common template to evaluate the heritability of tract-derived FA measures. We show that our template is reliable for integrating multiple datasets by combining results through meta-analysis and unifying the data through exploratory mega-analyses. Our results may help prioritize regions of the FA map that are consistently influenced by additive genetic factors for future genetic discovery studies. Protocols and templates are publicly available at (http://enigma.loni.ucla.edu/ongoing/dti-working-group/).

Multi-site study of additive genetic effects on fractional anisotropy of cerebral white matter: Comparing meta and megaanalytical approaches for data pooling

Combining datasets across independent studies can boost statistical power by increasing the numbers of observations and can achieve more accurate estimates of effect sizes. This is especially important for genetic studies where a large number of observations are required to obtain sufficient power to detect and replicate genetic effects. There is a need to develop and evaluate methods for joint-analytical analyses of rich datasets collected in imaging genetics studies. The ENIGMA-DTI consortium is developing and evaluating approaches for obtaining pooled estimates of heritability through meta-and mega-genetic analytical approaches, to estimate the general additive genetic contributions to the intersubject variance in fractional anisotropy (FA) measured from diffusion tensor imaging (DTI). We used the ENIGMA-DTI data harmonization protocol for uniform processing of DTI data from multiple sites. We evaluated this protocol in five family-based cohorts providing data from a total of 2248 children and adults (ages: 9-85) collected with various imaging protocols. We used the imaging genetics analysis tool, SOLAR-Eclipse, to combine twin and family data from Dutch, Australian and Mexican-American cohorts into one large "mega-family". We showed that heritability estimates may vary from one cohort to another. We used two meta-analytical (the sample-size and standard-error weighted) approaches and a mega-genetic analysis to calculate heritability estimates across-population. We performed leave-one-out analysis of the joint estimates of heritability, removing a different cohort each time to understand the estimate variability. Overall, meta- and mega-genetic analyses of heritability produced robust estimates of heritability.

Multiple stages classification of Alzheimer’s disease based on structural brain networks using generalized low rank approximations (GLRAM)

To classify each stage for a progressing disease such as Alzheimer’s disease is a key issue for the disease prevention and treatment. In this study, we derived structural brain networks from diffusion-weighted MRI using whole-brain tractography since there is growing interest in relating connectivity measures to clinical, cognitive, and genetic data. Relatively little work has usedmachine learning to make inferences about variations in brain networks in the progression of the Alzheimer’s disease. Here we developed a framework to utilize generalized low rank approximations of matrices (GLRAM) and modified linear discrimination analysis for unsupervised feature learning and classification of connectivity matrices. We apply the methods to brain networks derived from DWI scans of 41 people with Alzheimer’s disease, 73 people with EMCI, 38 people with LMCI, 47 elderly healthy controls and 221 young healthy controls. Our results show that this new framework can significantly improve classification accuracy when combining multiple datasets; this suggests the value of using data beyond the classification task at hand to model variations in brain connectivity.