Human identification from video using advanced gait recognition techniques

The solutions proposed in this thesis contribute to improve gait recognition performance in practical scenarios that further enable the adoption of gait recognition into real world security and forensic applications that require identifying humans at a distance. Pioneering work has been conducted on frontal gait recognition using depth images to allow gait to be integrated with biometric walkthrough portals. The effects of gait challenging conditions including clothing, carrying goods, and viewpoint have been explored. Enhanced approaches are proposed on segmentation, feature extraction, feature optimisation and classification elements, and state-of-the-art recognition performance has been achieved. A frontal depth gait database has been developed and made available to the research community for further investigation. Solutions are explored in 2D and 3D domains using multiple images sources, and both domain-specific and independent modality gait features are proposed.

Towards pose-robust face recognition on video

This thesis investigates face recognition in video under the presence of large pose variations. It proposes a solution that performs simultaneous detection of facial landmarks and head poses across large pose variations, employs discriminative modelling of feature distributions of faces with varying poses, and applies fusion of multiple classifiers to pose-mismatch recognition. Experiments on several benchmark datasets have demonstrated that improved performance is achieved using the proposed solution.

Text recognition approaches for indoor robotics : a comparison

This paper evaluates the performance of different text recognition techniques for a mobile robot in an indoor (university campus) environment. We compared four different methods: our own approach using existing text detection methods (Minimally Stable Extremal Regions detector and Stroke Width Transform) combined with a convolutional neural network, two modes of the open source program Tesseract, and the experimental mobile app Google Goggles. The results show that a convolutional neural network combined with the Stroke Width Transform gives the best performance in correctly matched text on images with single characters whereas Google Goggles gives the best performance on images with multiple words. The dataset used for this work is released as well.

Machine learning for activity recognition : hip versus wrist data

Problem addressed Wrist-worn accelerometers are associated with greater compliance. However, validated algorithms for predicting activity type from wrist-worn accelerometer data are lacking. This study compared the activity recognition rates of an activity classifier trained on acceleration signal collected on the wrist and hip. Methodology 52 children and adolescents (mean age 13.7 +/- 3.1 year) completed 12 activity trials that were categorized into 7 activity classes: lying down, sitting, standing, walking, running, basketball, and dancing. During each trial, participants wore an ActiGraph GT3X+ tri-axial accelerometer on the right hip and the non-dominant wrist. Features were extracted from 10-s windows and inputted into a regularized logistic regression model using R (Glmnet + L1). Results Classification accuracy for the hip and wrist was 91.0% +/- 3.1% and 88.4% +/- 3.0%, respectively. The hip model exhibited excellent classification accuracy for sitting (91.3%), standing (95.8%), walking (95.8%), and running (96.8%); acceptable classification accuracy for lying down (88.3%) and basketball (81.9%); and modest accuracy for dance (64.1%). The wrist model exhibited excellent classification accuracy for sitting (93.0%), standing (91.7%), and walking (95.8%); acceptable classification accuracy for basketball (86.0%); and modest accuracy for running (78.8%), lying down (74.6%) and dance (69.4%). Potential Impact Both the hip and wrist algorithms achieved acceptable classification accuracy, allowing researchers to use either placement for activity recognition.

Towards Vision-Based Pose- and Condition-Invariant Place Recognition along Routes

Vision-based place recognition involves recognising familiar places despite changes in environmental conditions or camera viewpoint (pose). Existing training-free methods exhibit excellent invariance to either of these challenges, but not both simultaneously. In this paper, we present a technique for condition-invariant place recognition across large lateral platform pose variance for vehicles or robots travelling along routes. Our approach combines sideways facing cameras with a new multi-scale image comparison technique that generates synthetic views for input into the condition-invariant Sequence Matching Across Route Traversals (SMART) algorithm. We evaluate the system’s performance on multi-lane roads in two different environments across day-night cycles. In the extreme case of day-night place recognition across the entire width of a four-lane-plus-median-strip highway, we demonstrate performance of up to 44% recall at 100% precision, where current state-of-the-art fails.

Old mastheads and new media : newspapers striving to adapt in Australia and South-east Asia

As a key element in their response to new media forcing transformations in mass media and media use, newspapers have deployed various strategies to not only establish online and mobile products, and develop healthy business plans, but to set out to be dominant portals. Their response to change was the subject of an early investigation by one of the present authors (Keshvani 2000). That was part of a set of short studies inquiring into what impact new software applications and digital convergence might have on journalism practice (Tickle and Keshvani 2000), and also looking for demonstrations of the way that innovations, technologies and protocols then under development might produce a “wireless, streamlined electronic news production process (Tickle and Keshvani 2001).” The newspaper study compared the online products of The Age in Melbourne and the Straits Times in Singapore. It provided an audit of the Singapore and Australia Information and Communications Technology (ICT) climate concentrating on the state of development of carrier networks, as a determining factor in the potential strength of the two services with their respective markets. In the outcome, contrary to initial expectations, the early cable roll-out and extensive ‘wiring’ of the city in Singapore had not produced a level of uptake of Internet services as strong as that achieved in Melbourne by more ad hoc and varied strategies. By interpretation, while news websites and online content were at an early stage of development everywhere, and much the same as one another, no determining structural imbalance existed to separate these leading media participants in Australia and South-east Asia. The present research revisits that situation, by again studying the online editions of the two large newspapers in the original study, and one other, The Courier Mail, (recognising the diversification of types of product in this field, by including it as a representative of Newscorp, now a major participant). The inquiry works through the principle of comparison. It is an exercise in qualitative, empirical research that establishes a comparison between the situation in 2000 as described in the earlier work, and the situation in 2014, after a decade of intense development in digital technology affecting the media industries. It is in that sense a follow-up study on the earlier work, although this time giving emphasis to content and style of the actual products as experienced by their users. It compares the online and print editions of each of these three newspapers; then the three mastheads as print and online entities, among themselves; and finally it compares one against the other two, as representing a South-east Asian model and Australian models. This exercise is accompanied by a review of literature on the developments in ICT affecting media production and media organisations, to establish the changed context. The new study of the online editions is conducted as a systematic appraisal of the first level, or principal screens, of the three publications, over the course of six days (10-15.2.14 inclusive). For this, categories for analysis were made, through conducting a preliminary examination of the products over three days in the week before. That process identified significant elements of media production, such as: variegated sourcing of materials; randomness in the presentation of items; differential production values among media platforms considered, whether text, video or stills images; the occasional repurposing and repackaging of top news stories of the day and the presence of standard news values – once again drawn out of the trial ‘bundle’ of journalistic items. Reduced in this way the online artefacts become comparable with the companion print editions from the same days. The categories devised and then used in the appraisal of the online products have been adapted to print, to give the closest match of sets of variables. This device, to study the two sets of publications on like standards -- essentially production values and news values—has enabled the comparisons to be made. This comparing of the online and print editions of each of the three publications was set up as up the first step in the investigation. In recognition of the nature of the artefacts, as ones that carry very diverse information by subject and level of depth, and involve heavy creative investment in the formulation and presentation of the information; the assessment also includes an open section for interpreting and commenting on main points of comparison. This takes the form of a field for text, for the insertion of notes, in the table employed for summarising the features of each product, for each day. When the sets of comparisons as outlined above are noted, the process then becomes interpretative, guided by the notion of change. In the context of changing media technology and publication processes, what substantive alterations have taken place, in the overall effort of news organisations in the print and online fields since 2001; and in their print and online products separately? Have they diverged or continued along similar lines? The remaining task is to begin to make inferences from that. Will the examination of findings enforce the proposition that a review of the earlier study, and a forensic review of new models, does provide evidence of the character and content of change --especially change in journalistic products and practice? Will it permit an authoritative description on of the essentials of such change in products and practice? Will it permit generalisation, and provide a reliable base for discussion of the implications of change, and future prospects? Preliminary observations suggest a more dynamic and diversified product has been developed in Singapore, well themed, obviously sustained by public commitment and habituation to diversified online and mobile media services. The Australian products suggest a concentrated corporate and journalistic effort and deployment of resources, with a strong market focus, but less settled and ordered, and showing signs of limitations imposed by the delay in establishing a uniform, large broadband network. The scope of the study is limited. It is intended to test, and take advantage of the original study as evidentiary material from the early days of newspaper companies’ experimentation with online formats. Both are small studies. The key opportunity for discovery lies in the ‘time capsule’ factor; the availability of well-gathered and processed information on major newspaper company production, at the threshold of a transformational decade of change in their industry. The comparison stands to identify key changes. It should also be useful as a reference for further inquiries of the same kind that might be made, and for monitoring of the situation in regard to newspaper portals on line, into the future.

Opportunity cost of unavailable surgical instruments in Australian hospitals

Hospitals invest considerable resources organizing operating suites and having surgeons and theatre staff available on an agreed schedule. A common impediment to efficiency is perioperative delay,including delays getting to the operating room or during the operation. Perioperative delays entail significant costs for hospitals,wasting staff time and operating theatre resources. They may also affect patient outcomes; prolonged surgery is a predictor for unanticipated admission following elective ambulatory surgery...

Visual place recognition for persistent robot navigation in changing environments

This thesis demonstrates that robots can learn about how the world changes, and can use this information to recognise where they are, even when the appearance of the environment has changed a great deal. The ability to localise in highly dynamic environments using vision only is a key tool for achieving long-term, autonomous navigation in unstructured outdoor environments. The proposed learning algorithms are designed to be unsupervised, and can be generated by the robot online in response to its observations of the world, without requiring information from a human operator or other external source.

Unsupervised online learning of condition-invariant images for place recognition

This paper presents an online, unsupervised training algorithm enabling vision-based place recognition across a wide range of changing environmental conditions such as those caused by weather, seasons, and day-night cycles. The technique applies principal component analysis to distinguish between aspects of a location’s appearance that are condition-dependent and those that are condition-invariant. Removing the dimensions associated with environmental conditions produces condition-invariant images that can be used by appearance-based place recognition methods. This approach has a unique benefit – it requires training images from only one type of environmental condition, unlike existing data-driven methods that require training images with labelled frame correspondences from two or more environmental conditions. The method is applied to two benchmark variable condition datasets. Performance is equivalent or superior to the current state of the art despite the lesser training requirements, and is demonstrated to generalise to previously unseen locations.

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.

Androgen-targeted therapy-induced epithelial mesenchymal plasticity and neuroendocrine transdifferentiation in prostate cancer : an opportunity for intervention

Androgens regulate biological pathways to promote proliferation, differentiation, and survival of benign and malignant prostate tissue. Androgen receptor (AR) targeted therapies exploit this dependence and are used in advanced prostate cancer to control disease progression. Contemporary treatment regimens involve sequential use of inhibitors of androgen synthesis or AR function. Although targeting the androgen axis has clear therapeutic benefit, its effectiveness is temporary, as prostate tumor cells adapt to survive and grow. The removal of androgens (androgen deprivation) has been shown to activate both epithelial-to-mesenchymal transition (EMT) and neuroendocrine transdifferentiation (NEtD) programs. EMT has established roles in promoting biological phenotypes associated with tumor progression (migration/invasion, tumor cell survival, cancer stem cell-like properties, resistance to radiation and chemotherapy) in multiple human cancer types. NEtD in prostate cancer is associated with resistance to therapy, visceral metastasis, and aggressive disease. Thus, activation of these programs via inhibition of the androgen axis provides a mechanism by which tumor cells can adapt to promote disease recurrence and progression. Brachyury, Axl, MEK, and Aurora kinase A are molecular drivers of these programs, and inhibitors are currently in clinical trials to determine therapeutic applications. Understanding tumor cell plasticity will be important in further defining the rational use of androgen-targeted therapies clinically and provides an opportunity for intervention to prolong survival of men with metastatic prostate cancer.

A Likelihood-Maximizing Framework for Enhanced In-Car Speech Recognition Based on Speech Dialog System Interaction

Speech recognition in car environments has been identified as a valuable means for reducing driver distraction when operating noncritical in-car systems. Under such conditions, however, speech recognition accuracy degrades significantly, and techniques such as speech enhancement are required to improve these accuracies. Likelihood-maximizing (LIMA) frameworks optimize speech enhancement algorithms based on recognized state sequences rather than traditional signal-level criteria such as maximizing signal-to-noise ratio. LIMA frameworks typically require calibration utterances to generate optimized enhancement parameters that are used for all subsequent utterances. Under such a scheme, suboptimal recognition performance occurs in noise conditions that are significantly different from that present during the calibration session – a serious problem in rapidly changing noise environments out on the open road. In this chapter, we propose a dialog-based design that allows regular optimization iterations in order to track the ever-changing noise conditions. Experiments using Mel-filterbank noise subtraction (MFNS) are performed to determine the optimization requirements for vehicular environments and show that minimal optimization is required to improve speech recognition, avoid over-optimization, and ultimately assist with semireal-time operation. It is also shown that the proposed design is able to provide improved recognition performance over frameworks incorporating a calibration session only.

Functional MRI of facial emotion recognition deficits in schizophrenia and their electrophysiological correlates

Empirical evidence suggests impaired facial emotion recognition in schizophrenia. However, the nature of this deficit is the subject of ongoing research. The current study tested the hypothesis that a generalized deficit at an early stage of face-specific processing (i.e. putatively subserved by the fusiform gyrus) accounts for impaired facial emotion recognition in schizophrenia as opposed to the Negative Emotion-specific Deficit Model, which suggests impaired facial information processing at subsequent stages. Event-related potentials (ERPs) were recorded from 11 schizophrenia patients and 15 matched controls while performing a gender discrimination and a facial emotion recognition task. Significant reduction of the face-specific vertex positive potential (VPP) at a peak latency of 165 ms was confirmed in schizophrenia subjects whereas their early visual processing, as indexed by P1, was found to be intact. Attenuated VPP was found to correlate with subsequent P3 amplitude reduction and to predict accuracy when performing a facial emotion discrimination task. A subset of ten schizophrenia patients and ten matched healthy control subjects also performed similar tasks in the magnetic resonance imaging scanner. Patients showed reduced blood oxygenation level-dependent (BOLD) activation in the fusiform, inferior frontal, middle temporal and middle occipital gyrus as well as in the amygdala. Correlation analyses revealed that VPP and the subsequent P3a ERP components predict fusiform gyrus BOLD activation. These results suggest that problems in facial affect recognition in schizophrenia may represent flow-on effects of a generalized deficit in early visual processing.