Development of a three dimensional audio-visual next generation speech recognition system. To overcome the disadvantages of current Audio-Visual Speech Recognition Systems, we propose a set of robust algorithms in three dimensional computer vision and speech processing. The proposed system will have far-reaching implications in various areas, for example, human-machine interaction for speech recognition in automated dialog systems and voice-to-text conversions.
Audio Visual Speech Recognition. Even though significant advances have been made in automatic speech recognition using acoustic information, the recognition accuracies are still poor in noisy and hostile environments such as in crowds, traffic, factory floors etc. In many of these applications visual information is or can easily be made available in addition to the audio. The aim of this project is to achieve an order of magnitude improvement in speech recognition accuracies in adverse environme ....Audio Visual Speech Recognition. Even though significant advances have been made in automatic speech recognition using acoustic information, the recognition accuracies are still poor in noisy and hostile environments such as in crowds, traffic, factory floors etc. In many of these applications visual information is or can easily be made available in addition to the audio. The aim of this project is to achieve an order of magnitude improvement in speech recognition accuracies in adverse environments by joint processing and modelling of the acoustic modality with visual information in the form of lip shapes and movements. The outcomes will be useful in human computer interaction in adverse environments as well as in the transcription and mining of multimedia data.
Read moreRead less
Advancing System Identification using Modern Optimisation Methods. This project lies within an ARC Research Priority Area. Namely, "Frontier Technologies". It involves the development of new technologies and fundamental theory that take data records from physical or abstract systems and generate mathematical models for use in prediction, control and diagnosis of the underlying system. In light of this, the project also lies within the ARC Research Priority Area of "Smart Information Use",
New Approaches for the Estimation of Complex Dynamic System Models. This project lies within an ARC defined Research Priority Area.
Namely, "Frontier Technologies". It is directed towards taking
available data records from physical and other processes, ranging from
petrochemical plant outputs to share market values, and using them to
determine equations that allow for the prediction, diagnosis, and
control of the underlying determining systems. As such, it lies
within the ARC identified ....New Approaches for the Estimation of Complex Dynamic System Models. This project lies within an ARC defined Research Priority Area.
Namely, "Frontier Technologies". It is directed towards taking
available data records from physical and other processes, ranging from
petrochemical plant outputs to share market values, and using them to
determine equations that allow for the prediction, diagnosis, and
control of the underlying determining systems. As such, it lies
within the ARC identified area of "Smart Information Use", but has further
potential to foster new approaches and new technologies, by way of the
improved system modelling, prediction and diagnosis capabilities it will provide.Read moreRead less
New Methods for Dynamic System Estimation. The estimation of dynamic system models from records of normal operating data is a significant problem across many engineering applications ranging from (at least) chemical process control to the implementation of telecommunications networks. This project, based on a promising pilot study, will develop and analyse new methods for solving these problems. This will lead to significant advances in the simplicity, robustness and generality of methods for a ....New Methods for Dynamic System Estimation. The estimation of dynamic system models from records of normal operating data is a significant problem across many engineering applications ranging from (at least) chemical process control to the implementation of telecommunications networks. This project, based on a promising pilot study, will develop and analyse new methods for solving these problems. This will lead to significant advances in the simplicity, robustness and generality of methods for accurate model generation with attendant benefits of more efficient and higher performance operation of many engineering systems.Read moreRead less
System Identification of Complex System Models. This project lies within an ARC Research Priority Area. Namely, "Frontier Technologies". It involves the development of new technologies and fundamental theory that take data records from physical or abstract systems and generate mathematical models for use in prediction, control and diagnosis of the underlying system. In light of this, the project also lies within the ARC Research Priority Area of "Smart Information Use",
New Model Predictive Control Design Methods. Automatic computer control is fundamental to sustaining a wide range of manufacturing, mineral processing, chemical processing, and other industries vital to the Australian economy. Furthermore, the efficiency, profitability, and environmental impact of these operations is directly linked to the quality of this computer control. In many situations, even a few percent improvement in automatic control delivers dividends measured in many millions of doll ....New Model Predictive Control Design Methods. Automatic computer control is fundamental to sustaining a wide range of manufacturing, mineral processing, chemical processing, and other industries vital to the Australian economy. Furthermore, the efficiency, profitability, and environmental impact of these operations is directly linked to the quality of this computer control. In many situations, even a few percent improvement in automatic control delivers dividends measured in many millions of dollars. This project will develop design tools allowing for more sophisticated, high performance control to be more widely employed. This will deliver the potential for economic and environmental benefits and energy savings to be achieved across a range of industries.Read moreRead less
Approximate authentication systems for digital information. Assurance about the origin and integrity of digital content is crucial not only in high security applications but also in everyday life scenarios such as providing proof that an X-ray image presented as part of an insurance claim is authentic, or a news clip is not tampered with. The outcomes of this project will significantly enhance trustworthiness of multimedia information systems which are increasingly used in areas such as surveil ....Approximate authentication systems for digital information. Assurance about the origin and integrity of digital content is crucial not only in high security applications but also in everyday life scenarios such as providing proof that an X-ray image presented as part of an insurance claim is authentic, or a news clip is not tampered with. The outcomes of this project will significantly enhance trustworthiness of multimedia information systems which are increasingly used in areas such as surveillance (traffic control), health, digital content production and distribution, tourism and journalism. It will also result in the development of secure biometric authentication systems which are critical in securing cyber space.Read moreRead less
Novel time-frequency techniques for analysing and modeling non-stationary physical and engineering data. This project addresses an issue of fundamental importance in science and technology, where non-stationary data (which have time-varying statistics) are ubiquitous. Therefore, the development of time-frequency tools to model and analyse non-stationary data has great potential for impact in a wide range of areas reaching from seismic data analysis to biomedical signal processing to sonar and ra ....Novel time-frequency techniques for analysing and modeling non-stationary physical and engineering data. This project addresses an issue of fundamental importance in science and technology, where non-stationary data (which have time-varying statistics) are ubiquitous. Therefore, the development of time-frequency tools to model and analyse non-stationary data has great potential for impact in a wide range of areas reaching from seismic data analysis to biomedical signal processing to sonar and radar. Employing techniques to be developed in this proposal, we expect to be able to classify and detect features of non-stationary data that were unrecognisable using hitherto known methods.Read moreRead less
Classification of human movement patterns from a triaxial accelerometer for home telecare. The purpose of the proposed ambulatory monitoring system is to ensure that a community-dwelling elderly person receives automatic, timely aid in the event of an adverse situation while alone at home, such as a fall or stroke. This project is aimed at the priority area 'ageing well, ageing productively'. Providing rapid assistance after a fall will prevent much of the morbidity, distress and socio-economic ....Classification of human movement patterns from a triaxial accelerometer for home telecare. The purpose of the proposed ambulatory monitoring system is to ensure that a community-dwelling elderly person receives automatic, timely aid in the event of an adverse situation while alone at home, such as a fall or stroke. This project is aimed at the priority area 'ageing well, ageing productively'. Providing rapid assistance after a fall will prevent much of the morbidity, distress and socio-economic cost that are associated with falls. Specifically, this project brings new advanced digital signal processing algorithms to bear on the difficult signal enhancement and classification problems associated with the information stream received from the triaxial accelerometer worn by the elderly client.Read moreRead less