QUT speaker identity verification system for EVALITA 2009

This document outlines the system submitted by the Speech and Audio Research Laboratory at the Queensland University of Technology (QUT) for the Speaker Identity Verication: Application task of EVALITA 2009. This submission consisted of a score-level fusion of three component systems, a joint-factor GMM system and two SVM systems using GLDS and GMM supervector kernels. Development and evaluation results are presented, demonstrating the effectiveness of this fused system approach.

Collaborative business process modeling using 3D virtual environments

Process modeling is a complex organizational task that requires many iterations and communication between the business analysts and the domain specialists involved in the process modeling. The challenge of process modeling is exacerbated, when the process of modeling has to be performed in a cross-organizational, distributed environment. Some systems have been developed to support collaborative process modeling, all of which use traditional 2D interfaces. We present an environment for collaborative process modeling, using 3D virtual environment technology. We make use of avatar instantiations of user ego centres, to allow for the spatial embodiment of the user with reference to the process model. We describe an innovative prototype collaborative process modeling approach, implemented as a modeling environment in Second Life. This approach leverages the use of virtual environments to provide user context for editing and collaborative exercises. We present a positive preliminary report on a case study, in which a test group modelled a business process using the system in Second Life.

On the cognitive effectiveness of routing symbols in process modeling languages

Process models provide visual support for analyzing and improving complex organizational processes. In this paper, we discuss differences of process modeling languages using cognitive effectiveness considerations, to make statements about the ease of use and quality of user experience. Aspects of cognitive effectiveness are of importance for learning a modeling language, creating models, and understanding models. We identify the criteria representational clarity, perceptual discriminability, perceptual immediacy, visual expressiveness, and graphic parsimony to compare and assess the cognitive effectiveness of different modeling languages. We apply these criteria in an analysis of the routing elements of UML Activity Diagrams, YAWL, BPMN, and EPCs, to uncover their relative strengths and weaknesses from a quality of user experience perspective. We draw conclusions that are relevant to the usability of these languages in business process modeling projects.

On the usage of labels and icons in business process modeling

The value of business process models is dependent not only on the choice of graphical elements in the model, but also on their annotation with additional textual and graphical information. This research discusses the use of text and icons for labeling the graphical constructs in a process model. We use two established verb classification schemes to examine the choice of activity labels in process modeling practice. Based on our findings, we synthesize a set of twenty-five activity label categories. We propose a systematic approach for graphically representing these label categories through the use of graphical icons, such that the resulting process models are easier and more readily understandable by end users. Our findings contribute to an ongoing stream of research investigating the practice of process modeling and thereby contribute to the body of knowledge about conceptual modeling quality overall.

Data-driven background dataset selection for SVM-based speaker verification

The recently proposed data-driven background dataset refinement technique provides a means of selecting an informative background for support vector machine (SVM)-based speaker verification systems. This paper investigates the characteristics of the impostor examples in such highly-informative background datasets. Data-driven dataset refinement individually evaluates the suitability of candidate impostor examples for the SVM background prior to selecting the highest-ranking examples as a refined background dataset. Further, the characteristics of the refined dataset were analysed to investigate the desired traits of an informative SVM background. The most informative examples of the refined dataset were found to consist of large amounts of active speech and distinctive language characteristics. The data-driven refinement technique was shown to filter the set of candidate impostor examples to produce a more disperse representation of the impostor population in the SVM kernel space, thereby reducing the number of redundant and less-informative examples in the background dataset. Furthermore, data-driven refinement was shown to provide performance gains when applied to the difficult task of refining a small candidate dataset that was mis-matched to the evaluation conditions.

Exploiting multiple feature sets in data-driven impostor dataset selection for speaker verification

This study assesses the recently proposed data-driven background dataset refinement technique for speaker verification using alternate SVM feature sets to the GMM supervector features for which it was originally designed. The performance improvements brought about in each trialled SVM configuration demonstrate the versatility of background dataset refinement. This work also extends on the originally proposed technique to exploit support vector coefficients as an impostor suitability metric in the data-driven selection process. Using support vector coefficients improved the performance of the refined datasets in the evaluation of unseen data. Further, attempts are made to exploit the differences in impostor example suitability measures from varying features spaces to provide added robustness.

Modeling multivariate autoregressive conditional heteroskedasticity with the double smooth transition conditional correlation GARCH Model

In this paper, we propose a multivariate GARCH model with a time-varying conditional correlation structure. The new double smooth transition conditional correlation (DSTCC) GARCH model extends the smooth transition conditional correlation (STCC) GARCH model of Silvennoinen and Teräsvirta (2005) by including another variable according to which the correlations change smoothly between states of constant correlations. A Lagrange multiplier test is derived to test the constancy of correlations against the DSTCC-GARCH model, and another one to test for another transition in the STCC-GARCH framework. In addition, other specification tests, with the aim of aiding the model building procedure, are considered. Analytical expressions for the test statistics and the required derivatives are provided. Applying the model to the stock and bond futures data, we discover that the correlation pattern between them has dramatically changed around the turn of the century. The model is also applied to a selection of world stock indices, and we find evidence for an increasing degree of integration in the capital markets.

Transaction mining for fraud detection in ERP Systems

Despite all attempts to prevent fraud, it continues to be a major threat to industry and government. Traditionally, organizations have focused on fraud prevention rather than detection, to combat fraud. In this paper we present a role mining inspired approach to represent user behaviour in Enterprise Resource Planning (ERP) systems, primarily aimed at detecting opportunities to commit fraud or potentially suspicious activities. We have adapted an approach which uses set theory to create transaction profiles based on analysis of user activity records. Based on these transaction profiles, we propose a set of (1) anomaly types to detect potentially suspicious user behaviour, and (2) scenarios to identify inadequate segregation of duties in an ERP environment. In addition, we present two algorithms to construct a directed acyclic graph to represent relationships between transaction profiles. Experiments were conducted using a real dataset obtained from a teaching environment and a demonstration dataset, both using SAP R/3, presently the predominant ERP system. The results of this empirical research demonstrate the effectiveness of the proposed approach.

Vision-based detection and tracking of aerial targets for UAV collision avoidance

Machine vision represents a particularly attractive solution for sensing and detecting potential collision-course targets due to the relatively low cost, size, weight, and power requirements of the sensors involved (as opposed to radar). This paper describes the development and evaluation of a vision-based collision detection algorithm suitable for fixed-wing aerial robotics. The system was evaluated using highly realistic vision data of the moments leading up to a collision. Based on the collected data, our detection approaches were able to detect targets at distances ranging from 400m to about 900m. These distances (with some assumptions about closing speeds and aircraft trajectories) translate to an advanced warning of between 8-10 seconds ahead of impact, which approaches the 12.5 second response time recommended for human pilots. We make use of the enormous potential of graphic processing units to achieve processing rates of 30Hz (for images of size 1024-by- 768). Currently, integration in the final platform is under way.

Airborne vision-based collision-detection system

Machine vision represents a particularly attractive solution for sensing and detecting potential collision-course targets due to the relatively low cost, size, weight, and power requirements of vision sensors (as opposed to radar and TCAS). This paper describes the development and evaluation of a real-time vision-based collision detection system suitable for fixed-wing aerial robotics. Using two fixed-wing UAVs to recreate various collision-course scenarios, we were able to capture highly realistic vision (from an onboard camera perspective) of the moments leading up to a collision. This type of image data is extremely scarce and was invaluable in evaluating the detection performance of two candidate target detection approaches. Based on the collected data, our detection approaches were able to detect targets at distances ranging from 400m to about 900m. These distances (with some assumptions about closing speeds and aircraft trajectories) translate to an advanced warning of between 8-10 seconds ahead of impact, which approaches the 12.5 second response time recommended for human pilots. We overcame the challenge of achieving real-time computational speeds by exploiting the parallel processing architectures of graphics processing units found on commercially-off-the-shelf graphics devices. Our chosen GPU device suitable for integration onto UAV platforms can be expected to handle real-time processing of 1024 by 768 pixel image frames at a rate of approximately 30Hz. Flight trials using manned Cessna aircraft where all processing is performed onboard will be conducted in the near future, followed by further experiments with fully autonomous UAV platforms.

Modeling the effects of strategic flexibility, organisational learning on e-business adoption and competitive advantage : an empirical investigation

The role of intangible firm capabilities as a source of competitive advantage has come into prominence in marketing strategy literature, due to the Resource Based View. This paper applies the Resource Based View and hypothesizes that strategic flexibility and organisation learning, conceptualised as capabilities, positively effect e-business adoption and competitive advantage. Partial Lease Squares analysis suggest that theoretical constructs function as hypothesised and explain a significant variation on e-business adoption and competitive advantage. Firms adopting e-business should develop capabilities such as strategic flexibility and organisation learning and that vendor firms may segment their potential clients based on these capabilities.

Toward robust image detection of crown-of-thorns starfish for autonomous population monitoring

Robust texture recognition in underwater image sequences for marine pest population control such as Crown-Of-Thorns Starfish (COTS) is a relatively unexplored area of research. Typically, humans count COTS by laboriously processing individual images taken during surveys. Being able to autonomously collect and process images of reef habitat and segment out the various marine biota holds the promise of allowing researchers to gain a greater understanding of the marine ecosystem and evaluate the impact of different environmental variables. This research applies and extends the use of Local Binary Patterns (LBP) as a method for texture-based identification of COTS from survey images. The performance and accuracy of the algorithms are evaluated on a image data set taken on the Great Barrier Reef.

Obstacle detection using optical flow

Motion has been examined in biology to be a critical component for obstacle avoidance and navigation. In particular, optical flow is a powerful motion cue that has been exploited in many biological systems for survival. In this paper, we investigate an obstacle detection system that uses optical flow to obtain range information to objects. Our experimental results demonstrate that optical flow is capable of providing good obstacle information but has obvious failure modes. We acknowledge that our optical flow system has certain disadvantages and cannot be solely used for navigation. Instead, we believe that optical flow is a critical visual subsystem used when moving at reason- able speeds. When combined with other visual subsystems, considerable synergy can result.

Fuzzy associative memory for humanoid robot joint control

Traditional approaches to joint control required accurate modelling of the system dynamic of the plant in question. Fuzzy Associative Memory (FAM) control schemes allow adequate control without a model of the system to be controlled. This paper presents a FAM based joint controller implemented on a humanoid robot. An empirically tuned PI velocity control loop is augmented with this feed forward FAM, with considerable reduction in joint position error achieved online and with minimal additional computational overhead.

Lip detection for audio-visual speech recognition in-car environment

Acoustically, car cabins are extremely noisy and as a consequence audio-only, in-car voice recognition systems perform poorly. As the visual modality is immune to acoustic noise, using the visual lip information from the driver is seen as a viable strategy in circumventing this problem by using audio visual automatic speech recognition (AVASR). However, implementing AVASR requires a system being able to accurately locate and track the drivers face and lip area in real-time. In this paper we present such an approach using the Viola-Jones algorithm. Using the AVICAR [1] in-car database, we show that the Viola- Jones approach is a suitable method of locating and tracking the driver’s lips despite the visual variability of illumination and head pose for audio-visual speech recognition system.