Enhancing human action recognition with region proposals

Deep convolutional network models have dominated recent work in human action recognition as well as image classification. However, these methods are often unduly influenced by the image background, learning and exploiting the presence of cues in typical computer vision datasets. For unbiased robotics applications, the degree of variation and novelty in action backgrounds is far greater than in computer vision datasets. To address this challenge, we propose an “action region proposal” method that, informed by optical flow, extracts image regions likely to contain actions for input into the network both during training and testing. In a range of experiments, we demonstrate that manually segmenting the background is not enough; but through active action region proposals during training and testing, state-of-the-art or better performance can be achieved on individual spatial and temporal video components. Finally, we show by focusing attention through action region proposals, we can further improve upon the existing state-of-the-art in spatio-temporally fused action recognition performance.

The con-stitutional re-cognition (s)cam-pain: The campaign for the hidden recognition of first nations peoples' racial inferiority

The constitutional recognition campaign has received party-wide support and its efforts have been promoted by Prime Minister Tony Abbott as being something that would ‘complete our Constitution.’ The broader rhetoric surrounding this campaign suggests that it will result in a just, albeit delayed, recognition of indigenous peoples in the Australian legal system. However, beneath the surface of this seemingly benevolent gesture, is a reaffirmation of the colonial subordination and erasure of the several hundred original nations’ peoples and ways of being.

Recognition and enforcement in cross-border insolvency law: A proposal for judicial gap-filling

This Article analyzes the recognition and enforcement of cross-border insolvency judgments from the United States, United Kingdom, and Australia to determine whether the UNCITRAL Model Law’s goal of modified universalism is currently being practiced, and subjects the Model Law to analysis through the lens of international relations theories to elaborate a way forward. We posit that courts could use the express language of the Model Law text to confer recognition and enforcement of foreign insolvency judgments. The adoption of our proposal will reduce costs, maximize recovery for creditors, and ensure predictability for all parties.

Government entity contribution to an effective partnership privatisation

The partnership form of privatisation is increasingly being used, in particular to carry out complex and evolving bundles of services. These have not previously been privatised because of incomplete contracts and contract management difficulties. Improved performance of the government entity as contract administrator and member of the partnership is crucial to modern service delivery expectations yet the privatisation literature has focused on other aspects of partnerships leaving the understanding of factors impacting the effectiveness of the government entity underdeveloped. This paper proposes the development of knowledge as to the range of factors which impact the effectiveness of the government entity. There is limited data available as to the operation of trust in the partnership relationship, and as to the capability of a range of privatisation forms to achieve stewardship of infrastructure. This research will utilise the findings from that research to build a tentative framework which will be utilised in staged research interrogating first the privatization literature and then the literature of other disciplines and sectors. The combined data will be analysed to provide government and practitioners such as government entity CEO’s with a complete listing of the operation of the factors which impact the effectiveness of the government entity in contributing to improved service delivery.

Adaptive frequency scaled wavelet packet decomposition for frog call classification

Environmental changes have put great pressure on biological systems leading to the rapid decline of biodiversity. To monitor this change and protect biodiversity, animal vocalizations have been widely explored by the aid of deploying acoustic sensors in the field. Consequently, large volumes of acoustic data are collected. However, traditional manual methods that require ecologists to physically visit sites to collect biodiversity data are both costly and time consuming. Therefore it is essential to develop new semi-automated and automated methods to identify species in automated audio recordings. In this study, a novel feature extraction method based on wavelet packet decomposition is proposed for frog call classification. After syllable segmentation, the advertisement call of each frog syllable is represented by a spectral peak track, from which track duration, dominant frequency and oscillation rate are calculated. Then, a k-means clustering algorithm is applied to the dominant frequency, and the centroids of clustering results are used to generate the frequency scale for wavelet packet decomposition (WPD). Next, a new feature set named adaptive frequency scaled wavelet packet decomposition sub-band cepstral coefficients is extracted by performing WPD on the windowed frog calls. Furthermore, the statistics of all feature vectors over each windowed signal are calculated for producing the final feature set. Finally, two well-known classifiers, a k-nearest neighbour classifier and a support vector machine classifier, are used for classification. In our experiments, we use two different datasets from Queensland, Australia (18 frog species from commercial recordings and field recordings of 8 frog species from James Cook University recordings). The weighted classification accuracy with our proposed method is 99.5% and 97.4% for 18 frog species and 8 frog species respectively, which outperforms all other comparable methods.

Activity recognition using binary tree SVM

This paper presents an effective classification method based on Support Vector Machines (SVM) in the context of activity recognition. Local features that capture both spatial and temporal information in activity videos have made significant progress recently. Efficient and effective features, feature representation and classification plays a crucial role in activity recognition. For classification, SVMs are popularly used because of their simplicity and efficiency; however the common multi-class SVM approaches applied suffer from limitations including having easily confused classes and been computationally inefficient. We propose using a binary tree SVM to address the shortcomings of multi-class SVMs in activity recognition. We proposed constructing a binary tree using Gaussian Mixture Models (GMM), where activities are repeatedly allocated to subnodes until every new created node contains only one activity. Then, for each internal node a separate SVM is learned to classify activities, which significantly reduces the training time and increases the speed of testing compared to popular the `one-against-the-rest' multi-class SVM classifier. Experiments carried out on the challenging and complex Hollywood dataset demonstrates comparable performance over the baseline bag-of-features method.

Human action recognition from video sequences

This PhD research has proposed new machine learning techniques to improve human action recognition based on local features. Several novel video representation and classification techniques have been proposed to increase the performance with lower computational complexity. The major contributions are the construction of new feature representation techniques, based on advanced machine learning techniques such as multiple instance dictionary learning, Latent Dirichlet Allocation (LDA) and Sparse coding. A Binary-tree based classification technique was also proposed to deal with large amounts of action categories. These techniques are not only improving the classification accuracy with constrained computational resources but are also robust to challenging environmental conditions. These developed techniques can be easily extended to a wide range of video applications to provide near real-time performance.

Elliptical summary randomisation for sensor-based human activity recognition

Many conventional statistical machine learning al- gorithms generalise poorly if distribution bias ex- ists in the datasets. For example, distribution bias arises in the context of domain generalisation, where knowledge acquired from multiple source domains need to be used in a previously unseen target domains. We propose Elliptical Summary Randomisation (ESRand), an efficient domain generalisation approach that comprises of a randomised kernel and elliptical data summarisation. ESRand learns a domain interdependent projection to a la- tent subspace that minimises the existing biases to the data while maintaining the functional relationship between domains. In the latent subspace, ellipsoidal summaries replace the samples to enhance the generalisation by further removing bias and noise in the data. Moreover, the summarisation enables large-scale data processing by significantly reducing the size of the data. Through comprehensive analysis, we show that our subspace-based approach outperforms state-of-the-art results on several activity recognition benchmark datasets, while keeping the computational complexity significantly low.

A cloud-based intelligent computing system for contextual exploration on personal sleep-tracking data using association rule mining

With the development of wearable and mobile computing technology, more and more people start using sleep-tracking tools to collect personal sleep data on a daily basis aiming at understanding and improving their sleep. While sleep quality is influenced by many factors in a person’s lifestyle context, such as exercise, diet and steps walked, existing tools simply visualize sleep data per se on a dashboard rather than analyse those data in combination with contextual factors. Hence many people find it difficult to make sense of their sleep data. In this paper, we present a cloud-based intelligent computing system named SleepExplorer that incorporates sleep domain knowledge and association rule mining for automated analysis on personal sleep data in light of contextual factors. Experiments show that the same contextual factors can play a distinct role in sleep of different people, and SleepExplorer could help users discover factors that are most relevant to their personal sleep.