Online self-supervised multi-instance segmentation of dynamic objects

This paper presents a method for the continuous segmentation of dynamic objects using only a vehicle mounted monocular camera without any prior knowledge of the object’s appearance. Prior work in online static/dynamic segmentation is extended to identify multiple instances of dynamic objects by introducing an unsupervised motion clustering step. These clusters are then used to update a multi-class classifier within a self-supervised framework. In contrast to many tracking-by-detection based methods, our system is able to detect dynamic objects without any prior knowledge of their visual appearance shape or location. Furthermore, the classifier is used to propagate labels of the same object in previous frames, which facilitates the continuous tracking of individual objects based on motion. The proposed system is evaluated using recall and false alarm metrics in addition to a new multi-instance labelled dataset to evaluate the performance of segmenting multiple instances of objects.

Insider comes out : an artist's inquiry and narrative about the relationship of art and mental health

Background Through an account of prevailing experiences of art and mental illness, this paper aims to raise awareness, open dialogue and create agency about art created by people with experience of mental illness. Methods This paper draws on personal narrative and inquiry by an artist with mental illness and data collected as part of a larger participatory action research project that investigated understandings of identity, art and mental illness. Result An inquiry through art raised awareness and attentiveness to the importance of choice in identity construction and exposed frequent dichotomies in art and mental illness that were negotiated to eschew prescribed social stratification. As an artist, the first author challenged values present in one idea and absent in the other, and the options and concessions available to authorise her own dialogue and agency of being an artist. Conclusion Constructing an identity is an important part of being human, the labels that we choose or are chosen for us attribute to our identity. Reflections and recommendations are offered to consider expanded ways of thinking about art and mental illness and the functions that art play in identity construction.

The past, present and future of carbon labelling for construction materials – a review

Global climate change is one of the most significant environmental issues that can harm human development. One central issue for the building and construction industry to address global climate change is the development of a credible and meaningful way to measure greenhouse gas (GHG) emissions. While Publicly Available Specification (PAS) 2050, the first international GHG standard, has been proven to be successful in standardizing the quantification process, its contribution to the management of carbon labels for construction materials is limited. With the recent publication of ISO 14067: Greenhouse gases – carbon footprint of products – requirements and guidelines for quantification and communication in May 2013, it is necessary for the building and construction industry to understand the past, present and future of the carbon labelling practices for construction materials. A systematic review shows that international GHG standards have been evolving in terms of providing additional guidance on communication and comparison, as well as less flexibility on the use of carbon labels. At the same time, carbon labelling schemes have been evolving on standardization and benchmarking. In addition, future actions are needed in the aspect of raising consumer awareness, providing benchmarking, ensuring standardization and developing simulation technologies in order for carbon labelling schemes for construction materials to provide credible, accurate and transparent information on GHG emissions.

Probability and story-telling : normative and descriptive models of juridical proof

The operation of the law rests on the selection of an account of the facts. Whether this involves prediction or postdiction, it is not possible to achieve certainty. Any attempt to model the operation of the law completely will therefore raise questions of how to model the process of proof. In the selection of a model a crucial question will be whether the model is to be used normatively or descriptively. Focussing on postdiction, this paper presents and contrasts the mathematical model with the story model. The former carries the normative stamp of scientific approval, whereas the latter has been developed by experimental psychologists to describe how humans reason. Neil Cohen's attempt to use a mathematical model descriptively provides an illustration of the dangers in not clearly setting this parameter of the modelling process. It should be kept in mind that the labels 'normative' and 'descriptive' are not eternal. The mathematical model has its normative limits, beyond which we may need to critically assess models with descriptive origins.

Flux density

Social Interiors (Julian Knowles, Rik Rue, Shane Fahey) are currently developing a major sound art project entitled Flux Density, in collaboration with a team of artists, focused on investigating the changing relationships between emerging digital technologies and traditional ‘obsolete’ analogue media. The project has two main components. – a curated compilation and a live performance. It is a large scale curatorial and performance project led by Social Interiors with assistant curators Joel Stern, Alessio Cavallaro and Shannon O’Neill. Presentation - International Symposium of Electronic Art. Social Interiors are one of Australia’s best known experimental sound ensembles. Project will consist of an online compilation of historic music emerging from the 80s cassette culture era, remix based works by Social Interiors, and work from new cassette labels established in a post internet era. Performance project will take place in Sydney and consist of Social Interiors in performance/collaboration with a range of well known artists. Partners include ABC Radio, ISEA, and Extreme Records.

The Hollywood effect : a comparative study of costume in Australia and the Proposition

This paper compares costuming practices in Baz Luhrmann’s Australia (2008) and John Hillcoat’s The Proposition (2005) and argues that high production values, such as in the blockbuster Australia, are not neutral mechanisms of production, but powerful prescriptive elements which do not result in a successful representation of cultural specificity. Australia is a typical blockbuster, it employs a large number of extras, it features compelling landscape shots, has been shot across four different locations and sets, and, importantly, is an international production with the 20th Century Fox. The film’s costumes were designed by Catherine Martin, who received an Oscar nomination in 2009. While global exposure of fashion in film and through celebrities’ endorsements has consolidated a historical synergy between the fashion industry and Hollywood, the Australian film and fashion industries have had a very limited exchange. Baz Luhrmann’s film is Australia’s first instance of promo-costuming and use of tie-in labels (Ferragamo, R.M.Williams, Prada, Paspaley). Catherine Martin thoroughly researched 1930s women’s wear, indigenous and stockmen’s clothing, and set up to make all costumes with a large team of costumiers and seamstresses, striving for authenticity. The Proposition won its costume designer Margot Wilson an AFI in 2005 for best costume, but compared to Australia the story, location and costumes are far harsher. Filmed around Winton in far west Queensland, the director John Hillcoat and Director of Photography Benoit Delhomme were insistent about realism, and emphasising the harshness of the Australian landscape. The realism of the costumes was derived from the fabrics and manufacturing, as well as the way they were shot, with the actors often wearing two or three layers of heavy wool during days of shooting in 50 degree heat, and the details of making and breaking down. The implication is that both films are culturally specific as they both deal with an Australian story. However, Australia is clearly produced according to a Hollywood blockbuster model, and closely matches Hollywood’s narrative and aesthetic characteristics, while The Proposition is a more modest film that eschews these conventions of beauty and glossed history. Despite its western genre-orientation, The Proposition is more successful than Australia when it comes to costuming, because its costumes are not only functional to the narrative, but, in Roland Barthes’ words, they also fulfil a prestation. This prestation highlights the social and cultural conflicts on which colonial Australia was founded, instead of gilding, and gliding, over them.

Understanding and analyzing a large collection of archived swimming videos

In elite sports, nearly all performances are captured on video. Despite the massive amounts of video that has been captured in this domain over the last 10-15 years, most of it remains in an 'unstructured' or 'raw' form, meaning it can only be viewed or manually annotated/tagged with higher-level event labels which is time consuming and subjective. As such, depending on the detail or depth of annotation, the value of the collected repositories of archived data is minimal as it does not lend itself to large-scale analysis and retrieval. One such example is swimming, where each race of a swimmer is captured on a camcorder and in-addition to the split-times (i.e., the time it takes for each lap), stroke rate and stroke-lengths are manually annotated. In this paper, we propose a vision-based system which effectively 'digitizes' a large collection of archived swimming races by estimating the location of the swimmer in each frame, as well as detecting the stroke rate. As the videos are captured from moving hand-held cameras which are located at different positions and angles, we show our hierarchical-based approach to tracking the swimmer and their different parts is robust to these issues and allows us to accurately estimate the swimmer location and stroke rates.

WikiLeaks : the Napster of secrets?

WikiLeaks has become a global phenomenon, and its founder and spokesman Julian Assange an international celebrity (or terrorist, depending on one’s perspective). But perhaps this focus on Assange and his website is as misplaced as the attacks against Napster and its founders were a decade ago: WikiLeaks itself only marks a new phase in a continuing shift in the balance of power between states and citizens, much as Napster helped to undermine the control of major music labels over the music industry. If the history of music filesharing is any guide, no level of punitive action against WikiLeaks and its supporters is going to re-contain the information WikiLeaks has set loose.

Across the divide - Criminology as other: Observations on the construction of the field

In 1978 Donald Cressey commented on an emerging division in the study of crime with some scholars concentrating on the development of a “crime fi ghting coalition” and others concerned with the processes associated with “making laws, breaking laws, and the reaction to the breaking of laws” (1978: 175). Since Cressey’s paper, many others have refl ected on the distinction between criminology and the sociology of crime and deviance (Akers, 1992; Garland, 1999; Garland & Sparks, 2000; Konty, 2007). But does such a distinction actually exist? Adopting a pragmatic position, the immediate answer is yes, if we assume that these categories have substance on the basis that they are grounded in everyday beliefs, institutional preferences and research practice (Konty, 2007). Moreover, these are viable categories in that some people studying crime label themselves criminologists (or are given this label by others) while others prefer or are given the label sociologist. Of course, there are further labels that may apply to persons studying crime, which include psychologist, penologist, biologist, chemist, and so on. One could argue that such labels are unimportant, however, it remains that these categories have a practical character. For criminology and the sociology of crime in particular, scholarly discourse frames these categories as oppositional (Bader et al., 1996.; Bendle, 1989; Laub & Sampson, 1991; Sibley, 2002) and to the extent that this has occurred, the categories have social relevance.

Emotional experiences of preservice science teachers in online learning : the formation, disruption and maintenance of social bonds

The enactment of learning to become a science teacher in online mode is an emotionally charged experience. We attend to the formation, maintenance and disruption of social bonds experienced by online preservice science teachers as they shared their emotional online learning experiences through blogs, or e-motion diaries, in reaction to videos of face-to-face lessons. A multi-theoretic framework drawing on microsociological perspectives of emotion informed our hermeneutic interpretations of students’ first-person accounts reported through an e-motion diary. These accounts were analyzed through our own database of emotion labels constructed from the synthesis of existing literature on emotion across a range of fields of inquiry. Preservice science teachers felt included in the face-to-face group as they watched videos of classroom transactions. The strength of these feelings of social solidarity were dependent on the quality of the video recording. E-motion diaries provided a resource for interactions focused on shared emotional experiences leading to formation of social bonds and the alleviation of feelings of fear, trepidation and anxiety about becoming science teachers. We offer implications to inform practitioners who wish to improve feelings of inclusion amongst their online learners in science education.

WikiLeaks : The Napster of Secrets?

WikiLeaks has become a global phenomenon, and its founder and spokesman Julian Assange an international celebrity (or terrorist, depending on one’s perspective). But perhaps this focus on Assange and his website is as misplaced as the attacks against Napster and its founders were a decade ago: WikiLeaks itself only marks a new phase in a continuing shift in the balance of power between states and citizens, much as Napster helped to undermine the control of major music labels over the music industry. If the history of music filesharing is any guide, no level of punitive action against WikiLeaks and its supporters is going to re-contain the information WikiLeaks has set loose.

Trashtopia by Maison Briz Vegas

Description of the Work Trashtopia was a fashion exhibition at Craft Queensland’s Artisan gallery showcasing outfits made entirely from rubbish materials. The exhibition was part of an on-going series by the Queensland Fashion Archives, called Remember or Revive. Maison Briz Vegas designers, Carla Binotto and Carla van Lunn created a dystopian beach holiday tableau referencing mid-century Californian and Gold Coast beach culture and style, and today’s plastic pollution of the world’s oceans. The display engaged a popular audience with ideas about environmental destruction and climate change while bringing twentieth and twenty-first century consumer and leisure culture into question. The medium of fashion was used as a means of amusement and provocation. The fashion objects and installation questioned current mores about the material value of rubbish and the installation was also a work of environmental activism. Statement of the Research Component The work was framed by critical reflections of contemporary consumer culture and research fields questioning value in waste materials and fashion objects. The work is situated in the context of conceptual and experimental fashion design practice and fashion presentation. The exhibited work transgressed the conventional production methods and material choice of designer fashion garments, for example, discarded plastic shopping bags were painstakingly shredded to mimic ostrich feathers. The viewer was prompted to reflect on the materiality of rubbish and its potential for transformation. The exhibition also sits in the context of culture jamming and contemporary activist practice. The work references and subverts twentieth century beach holiday culture, contrasting resort wear with a contemporary picture of plastic pollution of the oceans and climate change. Hawaiian style prints contained a playful and dark narrative of dying marine-life and the viewer was invited to take a “Greetings from Trashtopia” postcard depicting fashion models floating in oceans of plastic rubbish. This reflective creative practice sought to address the question of whether fashion made from recycled rubbish materials can critically and emotionally engage viewers with questions about contemporary consumer culture and material value. This work presents an innovative model of fashion design practice in which rubbish materials are transformed into designer garments and rubbish is placed centre stage in the public presentation of the designs. In overturning the traditional model of fashion presentation, the viewer is also given a deeper connection to the recycling process and complex ideas of waste and value. In 2015 two outfits from the exhibition were selected, along with works from three leading Australian fashion labels, and four leading New Zealand labels, for a commemorative ANZAC fashion collection shown at iD Dunedin Fashion Week. The show titled, “Together Alone, revisited” reprised an Australian and New Zealand fashion exhibition first held at the National Gallery of Victoria in 2009.

Clustering and labeling a web scale document collection using Wikipedia clusters

Clustering is an important technique in organising and categorising web scale documents. The main challenges faced in clustering the billions of documents available on the web are the processing power required and the sheer size of the datasets available. More importantly, it is nigh impossible to generate the labels for a general web document collection containing billions of documents and a vast taxonomy of topics. However, document clusters are most commonly evaluated by comparison to a ground truth set of labels for documents. This paper presents a clustering and labeling solution where the Wikipedia is clustered and hundreds of millions of web documents in ClueWeb12 are mapped on to those clusters. This solution is based on the assumption that the Wikipedia contains such a wide range of diverse topics that it represents a small scale web. We found that it was possible to perform the web scale document clustering and labeling process on one desktop computer under a couple of days for the Wikipedia clustering solution containing about 1000 clusters. It takes longer to execute a solution with finer granularity clusters such as 10,000 or 50,000. These results were evaluated using a set of external data.

The Fleet Store : devising a public retail exhibition to promote creative, innovative and sustainable business models for emerging fashion designers

1.Description of the Work The Fleet Store was devised as a creative output to establish an exhibition linked to a fashion business model where emerging designers were encouraged to research new and innovative strategies for creating design-driven and commercial collections for a public consumer. This was a project that was devised to break down the perceptions of emerging fashion designers that designing commercial collections linked to a sustainable business model is a boring and unnecessary process. The focus was to demystify the business of fashion and to link its importance to a design-driven and public outcome that is more familiar to fashion designers. The criterion for participation was that all designers had to be registered as a business with the Australian Taxation Office. Designers were chosen from the Creative Enterprise Australia Fashion Business Incubator, the QUT fashion graduate alumni and current QUT fashion design and double degree (fashion and business) students with existing businesses. The project evolved from a series of collaborative workshops where designers were introduced to new and innovative creative industries’ business models and the processes, costings and timings involved to create a niche, sustainable business for a public exhibition of design-driven commercial collections. All designers initiated their own business infra-structure but were then introduced to the concept of collaboration for successful and profitable exhibition and business outcomes. Collaborative strategies such as crowd funding, crowd sourcing, peer to peer mentoring and manufacturing were all researched, and strategies for the establishment of the retail exhibition were all devised in a collaborative environment. All participants also took on roles outside their ‘designer’ background to create a retail exhibition that was creative but also had critical mass and aesthetic for the consumer. The Fleet Store ‘popped up’ for 2 weeks (10 days), in a heritage-listed building in an inner city location. Passers-by were important, but the main consumer was enlisted by the use of interest and investment from crowd sourcing, crowd funding, ethical marketing, corporate social responsibility projects and collaborative public relations and social media strategies. The research has furthered discussion on innovative strategies for emerging fashion designers to initiate and maintain sustainable businesses and suggests that collaboration combined with a design-driven and business focus can create a sustainable and economically viable retail exhibition. 2. Research Statement Research Background The research field involved developing a new ethical, design-driven, collaborative and sustainable model for fashion design practice and management. The research asked can a public, design-driven, collaborative retail exhibition create a platform for promoting creative, innovative and sustainable business models for emerging fashion designers. The methodology was primarily practice-led as all participants were designers in their own right and the project manager acted as a mentor and curator to guide the process and analyse the potential of the research question. The Fleet Store offers new knowledge in design practice and management; with the creation of a model where design outcomes and business models are inextricably linked to the success of the creative output. Key innovations include extending the commercialisation of emerging fashion businesses by creating a curated retail gallery for collaborative and sustainable strategies to support niche fashion designer labels. This has contributed to a broader conversation on how to nurture and sustain competitive Australian fashion designers/labels. Research Contribution and Significance The Fleet Store has contributed to a growing body of research into innovative and sustainable business models for niche fashion and creative industries’ practitioners. All participants have maintained their business infra-structure and many are currently growing their businesses, using the strategies tested for the Fleet Store. The exhibition space was visited by over 1,000 people and sales of $27,000 were made in 10 days of opening. (Follow up sales of $3,000 has also been reported.) Three of the designers were ‘discovered’ from the exhibition and have received substantial orders from high profile national buyers and retailers for next season delivery. Several participants have since collaborated to create other pop up retail environments and are now mentoring other emerging designers on the significance of a collaborative retail exhibition to consolidate niche business models for emerging fashion designers.

Supervised latent dirichlet allocation models for efficient activity representation

Local spatio-temporal features with a Bag-of-visual words model is a popular approach used in human action recognition. Bag-of-features methods suffer from several challenges such as extracting appropriate appearance and motion features from videos, converting extracted features appropriate for classification and designing a suitable classification framework. In this paper we address the problem of efficiently representing the extracted features for classification to improve the overall performance. We introduce two generative supervised topic models, maximum entropy discrimination LDA (MedLDA) and class- specific simplex LDA (css-LDA), to encode the raw features suitable for discriminative SVM based classification. Unsupervised LDA models disconnect topic discovery from the classification task, hence yield poor results compared to the baseline Bag-of-words framework. On the other hand supervised LDA techniques learn the topic structure by considering the class labels and improve the recognition accuracy significantly. MedLDA maximizes likelihood and within class margins using max-margin techniques and yields a sparse highly discriminative topic structure; while in css-LDA separate class specific topics are learned instead of common set of topics across the entire dataset. In our representation first topics are learned and then each video is represented as a topic proportion vector, i.e. it can be comparable to a histogram of topics. Finally SVM classification is done on the learned topic proportion vector. We demonstrate the efficiency of the above two representation techniques through the experiments carried out in two popular datasets. Experimental results demonstrate significantly improved performance compared to the baseline Bag-of-features framework which uses kmeans to construct histogram of words from the feature vectors.