http://nova.newcastle.edu.au/vital/access/services/Feed ${session.getAttribute("locale")} 5 http://nova.newcastle.edu.au/vital/access/manager/Repository/uon:8416 This paper introduces an automatic liver parenchyma segmentation algorithm that can delineate liver in abdominal CT images. The proposed approach consists of three main steps. Firstly, a texture analysis is applied onto input abdominal CT images to extract pixel level features. Here, two main categories of features, namely wavelet coefficients and Haralick texture descriptors are investigated. Secondly, support vector machines (SVM) are implemented to classify the data into pixel-wised liver or non-liver. Finally, specially combined morphological operations are designed as a post processor to remove noise and to delineate the liver. Our unique contributions to liver segmentation are twofold: one is that it has been proved through experiments that wavelet features present better classification than Haralick texture descriptors when SVMs are used; the other is that the combination of morphological operations with a pixel-wised SVM classifier can delineate volumetric liver accurately. The algorithm can be used in an advanced computer-aided liver disease diagnosis and surgical planning systems. Examples of applying the algorithm on real CT data are presented with performance validation based on the automatically segmented results and that of manually segmented ones. 2013-02-07T01:53:31.066Z ]]> http://nova.newcastle.edu.au/vital/access/manager/Repository/uon:11813 Research Doctorate - Doctor of Philosophy (PhD) 2012-10-25T04:45:30.599Z ]]> http://nova.newcastle.edu.au/vital/access/manager/Repository/uon:2938 The aim of this work is to contribute some insights and a partial overview of how machine learning methods are used in robotics. We first discuss typical general issues in the relationship between robotics and machine learning. Then we focus on projects associated with the RoboCup competition and symposium, and review the extent to which machine learning approaches have been used in the 4-legged league at RoboCup during the years 1998–2003. Further, we summarise the machine learning methods that were used by our own RoboCup team—the NUbots—in 2002/2003. 2012-01-30T04:38:40.843Z ]]> http://nova.newcastle.edu.au/vital/access/manager/Repository/uon:8371 Reinforcement learning (RL) consists of methods that automatically adjust behaviour based on numerical rewards and penalties. While use of the attribute-value framework is widespread in RL, it has limited expressive power. Logic languages, such as first-order logic, provide a more expressive framework, and their use in RL has led to the field of relational RL. This thesis develops a system for relational RL based on learning classifier systems (LCS). In brief, the system generates, evolves, and evaluates a population of condition-action rules, which take the form of definite clauses over first-order logic. Adopting the LCS approach allows the resulting system to integrate several desirable qualities: model-free and "tabula rasa" learning; a Markov Decision Process problem model; and importantly, support for variables as a principal mechanism for generalisation. The utility of variables is demonstrated by the system's ability to learn genuinely scalable behaviour - behaviour learnt in small environments that translates to arbitrary large versions of the environment without the need for retraining. 2011-07-20T05:20:13.035Z ]]> http://nova.newcastle.edu.au/vital/access/manager/Repository/uon:7864 The detection of long term trends in woody vegetation in Queensland, Australia, from the Landsat-5 TM and Landsat-7 ETM+ sensors requires the automated prediction of overstorey foliage projective cover (FPC) from a large volume of Landsat imagery. This paper presents a comparison of parametric (Multiple Linear Regression, Generalized Linear Models) and machine learning (Random Forests, Support Vector Machines) regression models for predicting overstorey FPC from Landsat-5 TM and Landsat-7 ETM+ imagery. Estimates of overstorey FPC were derived from field measured stand basal area (RMSE 7.26%) for calibration of the regression models. Independent estimates of overstorey FPC were derived from field and airborne LiDAR (RMSE 5.34%) surveys for validation of model predictions. The airborne LiDAR-derived estimates of overstorey FPC enabled the bias and variance of model predictions to be quantified in regional areas. The results showed all the parametric and machine learning models had similar prediction errors (RMSE < 10%), but the machine learning models had less bias than the parametric models at greater than ~60% overstorey FPC. All models showed greater than 10% bias in plant communities with high herbaceous or understorey FPC. The results of this work indicate that use of overstorey FPC products derived from Landsat-5 TM or Landsat-7 ETM+ data in Queensland using any of the regression models requires the assumption of senescent or absent herbaceous foliage at the time of image acquisition. 2011-06-08T06:10:09.656Z ]]> http://nova.newcastle.edu.au/vital/access/manager/Repository/uon:7779 Robot learning is a growing area of research at the intersection of robotics and machine learning. The main contributions of this paper include a review of how machine learning has been used on Sony AIBO robots and at RoboCup, with a focus on the four-legged league during the years 1998-2004. The review shows that the application-oriented use of machine learning in the four-legged league was still conservative and restricted to a few well-known and easy-to-use methods such as standard decision trees, evolutionary hill climbing, and support vector machines. Method-oriented spin-off studies emerged more frequently and increasingly addressed new and advanced machine learning techniques. Further, the paper presents some details about the growing impact of machine learning in the software system developed by the authors' robot soccer team - the NUbots. 2011-05-26T03:10:04.769Z ]]> http://nova.newcastle.edu.au/vital/access/manager/Repository/uon:6653 The first part of the present chapter provides some theoretical background and explains in the next section the general idea of kernel machines and kernelisation. Then the three fundamental machine learning paradigms dimensionality reduction, regression, and classification as well as associated questions of kernel and parameter selection are addressed. The chapter's second part gives a survey of typical questions and tasks arising in finance applications and how kernel methods have been applied to solve them. Finally follows a brief overview of relevant software toolboxes. 2010-09-10T02:10:03.016Z ]]> http://nova.newcastle.edu.au/vital/access/manager/Repository/uon:5878 A novel automatic system for extracting the semantic descriptions of medical image content and concept in text form is presented. We first extract and analyse image features and the features are mapped to semantic descriptions by fuzzy functions. Based on these semantic descriptions, our system facilitates knowledge base construction using a machine learning scheme. The result will be useful for other researchers in medical image retrieval area, who can take advantage of both text-based queries and image queries. 2010-07-20T02:40:08.507Z ]]>