Towards improved speech recognition for resource poor languages

In recent times, the improved levels of accuracy obtained by Automatic Speech Recognition (ASR) technology has made it viable for use in a number of commercial products. Unfortunately, these types of applications are limited to only a few of the world’s languages, primarily because ASR development is reliant on the availability of large amounts of language specific resources. This motivates the need for techniques which reduce this language-specific, resource dependency. Ideally, these approaches should generalise across languages, thereby providing scope for rapid creation of ASR capabilities for resource poor languages. Cross Lingual ASR emerges as a means for addressing this need. Underpinning this approach is the observation that sound production is largely influenced by the physiological construction of the vocal tract, and accordingly, is human, and not language specific. As a result, a common inventory of sounds exists across languages; a property which is exploitable, as sounds from a resource poor, target language can be recognised using models trained on resource rich, source languages. One of the initial impediments to the commercial uptake of ASR technology was its fragility in more challenging environments, such as conversational telephone speech. Subsequent improvements in these environments has gained consumer confidence. Pragmatically, if cross lingual techniques are to considered a viable alternative when resources are limited, they need to perform under the same types of conditions. Accordingly, this thesis evaluates cross lingual techniques using two speech environments; clean read speech and conversational telephone speech. Languages used in evaluations are German, Mandarin, Japanese and Spanish. Results highlight that previously proposed approaches provide respectable results for simpler environments such as read speech, but degrade significantly when in the more taxing conversational environment. Two separate approaches for addressing this degradation are proposed. The first is based on deriving better target language lexical representation, in terms of the source language model set. The second, and ultimately more successful approach, focuses on improving the classification accuracy of context-dependent (CD) models, by catering for the adverse influence of languages specific phonotactic properties. Whilst the primary research goal in this thesis is directed towards improving cross lingual techniques, the catalyst for investigating its use was based on expressed interest from several organisations for an Indonesian ASR capability. In Indonesia alone, there are over 200 million speakers of some Malay variant, provides further impetus and commercial justification for speech related research on this language. Unfortunately, at the beginning of the candidature, limited research had been conducted on the Indonesian language in the field of speech science, and virtually no resources existed. This thesis details the investigative and development work dedicated towards obtaining an ASR system with a 10000 word recognition vocabulary for the Indonesian language.

Fabrication using a rapid prototyping system and in vitro characterization of PEG-PCL-PLA scaffolds for tissue engineering

n the field of tissue engineering new polymers are needed to fabricate scaffolds with specific properties depending on the targeted tissue. This work aimed at designing and developing a 3D scaffold with variable mechanical strength, fully interconnected porous network, controllable hydrophilicity and degradability. For this, a desktop-robot-based melt-extrusion rapid prototyping technique was applied to a novel tri-block co-polymer, namely poly(ethylene glycol)-block-poly(epsi-caprolactone)-block-poly(DL-lactide), PEG-PCL-P(DL)LA. This co-polymer was melted by electrical heating and directly extruded out using computer-controlled rapid prototyping by means of compressed purified air to build porous scaffolds. Various lay-down patterns (0/30/60/90/120/150°, 0/45/90/135°, 0/60/120° and 0/90°) were produced by using appropriate positioning of the robotic control system. Scanning electron microscopy and micro-computed tomography were used to show that 3D scaffold architectures were honeycomb-like with completely interconnected and controlled channel characteristics. Compression tests were performed and the data obtained agreed well with the typical behavior of a porous material undergoing deformation. Preliminary cell response to the as-fabricated scaffolds has been studied with primary human fibroblasts. The results demonstrated the suitability of the process and the cell biocompatibility of the polymer, two important properties among the many required for effective clinical use and efficient tissue-engineering scaffolding.

Direct writing of chitosan scaffolds using a robotic system

Purpose – This paper aims to present a novel rapid prototyping (RP) fabrication methods and preliminary characterization for chitosan scaffolds. Design – A desktop rapid prototyping robot dispensing (RPBOD) system has been developed to fabricate scaffolds for tissue engineering (TE) applications. The system is a computer-controlled four-axis machine with a multiple-dispenser head. Neutralization of the acetic acid by the sodium hydroxide results in a precipitate to form a gel-like chitosan strand. The scaffold properties were characterized by scanning electron microscopy, porosity calculation and compression test. An example of fabrication of a freeform hydrogel scaffold is demonstrated. The required geometric data for the freeform scaffold were obtained from CT-scan images and the dispensing path control data were converted form its volume model. The applications of the scaffolds are discussed based on its potential for TE. Findings – It is shown that the RPBOD system can be interfaced with imaging techniques and computational modeling to produce scaffolds which can be customized in overall size and shape allowing tissue-engineered grafts to be tailored to specific applications or even for individual patients. Research limitations/implications – Important challenges for further research are the incorporation of growth factors, as well as cell seeding into the 3D dispensing plotting materials. Improvements regarding the mechanical properties of the scaffolds are also necessary. Originality/value – One of the important aspects of TE is the design scaffolds. For customized TE, it is essential to be able to fabricate 3D scaffolds of various geometric shapes, in order to repair tissue defects. RP or solid free-form fabrication techniques hold great promise for designing 3D customized scaffolds; yet traditional cell-seeding techniques may not provide enough cell mass for larger constructs. This paper presents a novel attempt to fabricate 3D scaffolds, using hydrogels which in the future can be combined with cells.

Low-cost hardware speech enhancement for improved speech recognition in automotive environments

Voice recognition is one of the key enablers to reduce driver distraction as in-vehicle systems become more and more complex. With the integration of voice recognition in vehicles, safety and usability are improved as the driver’s eyes and hands are not required to operate system controls. Whilst speaker independent voice recognition is well developed, performance in high noise environments (e.g. vehicles) is still limited. La Trobe University and Queensland University of Technology have developed a low-cost hardware-based speech enhancement system for automotive environments based on spectral subtraction and delay–sum beamforming techniques. The enhancement algorithms have been optimised using authentic Australian English collected under typical driving conditions. Performance tests conducted using speech data collected under variety of vehicle noise conditions demonstrate a word recognition rate improvement in the order of 10% or more under the noisiest conditions. Currently developed to a proof of concept stage there is potential for even greater performance improvement.

3D visualisation of groundwater/surface water systems as a planning and communication tool

Groundwater is increasingly recognised as an important yet vulnerable natural resource, and a key consideration in water cycle management. However, communication of sub-surface water system behaviour, as an important part of encouraging better water management, is visually difficult. Modern 3D visualisation techniques can be used to effectively communicate these complex behaviours to engage and inform community stakeholders. Most software developed for this purpose is expensive and requires specialist skills. The Groundwater Visualisation System (GVS) developed by QUT integrates a wide range of surface and sub-surface data, to produce a 3D visualisation of the behaviour, structure and connectivity of groundwater/surface water systems. Surface data (elevation, surface water, land use, vegetation and geology) and data collected from boreholes (bore locations and subsurface geology) are combined to visualise the nature, structure and connectivity of groundwater/surface water systems. Time-series data (water levels, groundwater quality, rainfall, stream flow and groundwater abstraction) is displayed as an animation within the 3D framework, or graphically, to show water system condition changes over time. GVS delivers an interactive, stand-alone 3D Visualisation product that can be used in a standard PC environment. No specialised training or modelling skills are required. The software has been used extensively in the SEQ region to inform and engage both water managers and the community alike. Examples will be given of GVS visualisations developed in areas where there have been community concerns around groundwater over-use and contamination.

Design and evaluation of a prototype tracking system

In this paper, we present the design and construction of a prototype target tracking system. The experimental set up consists of three main modules for moving the object, detecting the motion of the object and its tracking. The mechanism for moving the object includes an object and two stepper motors and their driving and control circuitry. The detection of the object’s motion is realized by photo switch array. The tracking mechanism consists of a laser beam and two DC servomotors and their associated circuitry. The control algorithm is a standard fuzzy logic controller. The system is designed to operate in two modes in such a way that the role of target and tracker can be interchanged. Experimental results indicate that the fuzzy controller is capable of controlling the system in both modes.

GVS : A flexible, low-end, 3D visualisation framework for enhancing conceptual groundwater models for community, management and simulations

In this paper we discuss an advanced, 3D groundwater visualisation and animation system that allows scientists, government agencies and community groups to better understand the groundwater processes that effect community planning and decision-making. The system is unique in that it has been designed to optimise community engagement. Although it incorporates a powerful visualisation engine, this open-source system can be freely distributed and boasts a simple user interface allowing individuals to run and investigate the models on their own PCs and gain intimate knowledge of the groundwater systems. The initial version of the Groundwater Visualisation System (GVS v1.0), was developed from a coastal delta setting (Bundaberg, QLD), and then applied to a basalt catchment area (Obi Obi Creek, Maleny, QLD). Several major enhancements have been developed to produce higher quality visualisations, including display of more types of data, support for larger models and improved user interaction. The graphics and animation capabilities have also been enhanced, notably the display of boreholes, depth logs and time-series water level surfaces. The GVS software remains under continual development and improvement

Community engagement and groundwater investigation in a rural/peri urban basaltic aquifier system, Tamborine Mountain, South East Queensland

The Tamborine Mt area is a popular residential and tourist area in the Gold Coast hinterland, SE Qld. The 15km2 area occurs on elevated remnant Tertiary Basalts of the Beechmont Group, which comprise a number of mappable flow units originally derived from the Tweed volcanic centre to the south. The older Albert Basalt (Tertiary), which underlies the Beechmont Basalt at the southern end of the investigation area, is thought to be derived from the Focal Peak volcanic centre to the south west. The Basalts contain a locally significant ‘un-declared’ groundwater resource, which is utilised by the Tamborine Mt community for: • domestic purposes to supplement rainwater tank supplies, • commercial scale horticulture and • commercial export off-Mountain for bottled water. There is no reticulated water supply, and all waste water is treated on-site through domestic scale WTPs. Rainforest and other riparian ecosystems that attract residents and tourist dollars to the area, are also reliant on the groundwater that discharges to springs and surface streams on and around the plateau. Issues regarding a lack of compiled groundwater information, groundwater contamination, and groundwater sustainability are being investigated by QUT, utilising funding provided by the Federal Government’s ‘Caring for our Country’ programme through SEQ Catchments Ltd. The objectives of the two year project, which started in April 2009, are to: • Characterise the nature and condition of groundwater / surface water systems in the Tamborine Mountain area in terms of the issues being raised; • Engage and build capacity within the community to source local knowledge, encourage participation, raise awareness and improve understanding of the impacts of land and water use; • Develop a stand-alone 3D Visualisation model for dissemination into the community and use as a communication tool.

Security metrics for object-oriented designs

Several studies have developed metrics for software quality attributes of object-oriented designs such as reusability and functionality. However, metrics which measure the quality attribute of information security have received little attention. Moreover, existing security metrics measure either the system from a high level (i.e. the whole system’s level) or from a low level (i.e. the program code’s level). These approaches make it hard and expensive to discover and fix vulnerabilities caused by software design errors. In this work, we focus on the design of an object-oriented application and define a number of information security metrics derivable from a program’s design artifacts. These metrics allow software designers to discover and fix security vulnerabilities at an early stage, and help compare the potential security of various alternative designs. In particular, we present security metrics based on composition, coupling, extensibility, inheritance, and the design size of a given object-oriented, multi-class program from the point of view of potential information flow.

Assessing the impact of refactoring on security-critical object-oriented designs

Refactoring focuses on improving the reusability, maintainability and performance of programs. However, the impact of refactoring on the security of a given program has received little attention. In this work, we focus on the design of object-oriented applications and use metrics to assess the impact of a number of standard refactoring rules on their security by evaluating the metrics before and after refactoring. This assessment tells us which refactoring steps can increase the security level of a given program from the point of view of potential information flow, allowing application designers to improve their system’s security at an early stage.