No description available.

No description available.

This thesis presents a biologically inspired model of learning and development. This model decomposes the lifetime of a single learning system into a number of stages, analogous to the infant, juvenile, adolescent and adult stages of development in a biological system. This model is then applied to Kohonen's SOM algorithm.In order to better understand the operation of Kohonen's SOM algorithm, a theoretical analysis of self-organisation is performed. This analysis establishes the role played by lateral connections in organisation, and the significance of the Laplacian lateral connections common to many SOM architectures.This analysis of neighbourhood interactions is then used to develop three key variations on Kohonen's SOM algorithm. Firstly, a new scheme for parameter decay, known as Butterworth Step Decay, is presented. This decay scheme provides training times comparable to the best training times possible using traditional linear decay, but precludes the need for a priori knowledge of likely training times. In addition, this decay scheme allows Kohonen's SOM to learn in a continuous manner.Secondly, a method is presented for establishing core knowledge in the fundamental representation of a SOM. This technique is known as Syllabus Presentation. This technique involves using a selected training syllabus to reinforce knowledge known to be significant. A method for developing a training syllabus, known as Percept Masking, is also presented.Thirdly, a method is presented for preventing the loss of trained representations in a continuously learning SOM. This technique, known as Arbor Pruning, involves restricting the weight update process to prevent the loss of significant representations. This technique can be used if the data domain varies within a known set of dimensions. However, it cannot be used to control forgetfulness if dimensions are added to or removed from ++ / the data domain.

As digital crimes have risen, so has the need for digital forensics. Numerous state-of-the-art tools have been developed to assist digital investigators conduct proper investigations into digital crimes. However, digital investigations are becoming increasingly complex and time consuming due to the amount of data involved, and digital investigators can find themselves unable to conduct them in an appropriately efficient and effective manner. This situation has prompted the need for new tools capable of handling such large, complex investigations. Data mining is one such potential tool. It is still relatively unexplored from a digital forensics perspective, but the purpose of data mining is to discover new knowledge from data where the dimensionality, complexity or volume of data is prohibitively large for manual analysis. This study assesses the self-organising map (SOM), a neural network model and data mining technique that could potentially offer tremendous benefits to digital forensics. The focus of this study is to demonstrate how the SOM can help digital investigators to make better decisions and conduct the forensic analysis process more efficiently and effectively during a digital investigation. The SOM’s visualisation capabilities can not only be used to reveal interesting patterns, but can also serve as a platform for further, interactive analysis. / Dissertation (MSc (Computer Science))--University of Pretoria, 2007. / Computer Science / unrestricted

Wastewaters from intensive agricultural activities contain high concentrations of nitrogen and phosphorus that contributes to water management problems. During the past few years, there has been considerable interest in the use of constructed wetlands for treating surface water runoff from farmyards. If the contaminated runoff is not treated, this wastewater along with other non-point sources of pollution can seriously contaminate the surface water and groundwater. Integrated Constructed Wetlands (ICWs) are a type of free water surface wetlands. They are engineered systems that are designed, constructed and operated successfully for treating farmyard runoff in the British Isles. However, the long-term treatment performance of these systems, the processes involved in contaminant removal and the impact on associated water bodies are not well-known. The aims of this project were to assess the performance of full-scale integrated constructed wetlands and understand nutrient removal in them. Performance evaluation of these systems through physical, chemical and microbiological parameters collected for more than 7 years showed good removal efficiencies compared to international literature. The monitored nutrient concentrations in groundwater and surface waters indicate that ICW systems did not pollute the receiving waters. The role of plants (Typha latifolia) and sediment in removing nutrients was also assessed. More nitrogen and phosphorus were stored in wetland soils and sediments than in plants. The results demonstrate that the soil component of a mature wetland system is an important and sustainable nutrient storage compartment. A novel molecular toolbox was used to characterise and compare microbial diversity responsible for nitrogen removal in sediment and litter components of ICW systems. Diverse populations of nitrogen removing bacteria were detected. The litter component of the wetland systems supported more diverse nitrogen removing bacteria than the sediments. Nitrogen removing bacteria in the wetland systems appeared to be stochastically assembled from the same source community. The self-organising map model was applied as a prediction tool for the performance of ICW and to investigate an alternative method of analysing water quality performance indicators. The model performed very well in predicting nutrients and biochemical oxygen demand with easy to measure and cost-effective water quality parameters. The results indicate that the model was an appropriate approach to monitor wastewater treatment processes and can be used to support management of ICW in real-time.

Thesis (PhD)--Stellenbosch University, 2011. / ENGLISH ABSTRACT: The knee is the largest, most complicated and incongruent joint in the human body. It sustains very high forces and is susceptible to injury and disease. Osteoarthritis is a common disease prevalent among the elderly and causes softening or degradation of the cartilage and subcondral bone in the joint, which leads to a loss of function and pain. This problem can be alleviated through a surgical intervention commonly termed a “knee replacement”. The aim of a knee replacement procedure is to relieve pain and restore normal function. Ideally, the knee replacement prosthesis should have an articulating geometry similar to that of the patient’s healthy knee, and must allow for normal motion. Unfortunately, this is often problematic since knee prostheses are supplied in standard sizes from a variety of manufacturers and each one has a slightly different design. Furthermore, commercial prostheses are not always able to restore the complex geometry of an individual patient’s original articulating surfaces. This dissertation shows that there is a significant variation between knee geometries, regardless of gender and race. This research aims to resolve the problem in two parts: Firstly by presenting a method for preoperatively selecting the optimal knee prosthesis type and size for a specific patient, and secondly by presenting a design procedure for designing and manufacturing patient-specific unicompartmental knee replacements. The design procedure uses mathematical modelling and an artificial neural network to estimate the original and healthy articulating surfaces of a patient’s knee. The models are combined with medical images from the patient to create a knee prosthesis that is patient-specific. These patient-specific implants are then compared to conventional implants with respect to contact stresses and kinematics. The dissertation concludes that patient-specific implants can have characteristics that are comparable to or better than conventional prostheses. The unique design methodology presented in this dissertation introduces a significant advancement in knee replacement technology, with the potential to dramatically improve clinical outcomes of knee replacement surgery. / AFRIKAANSE OPSOMMING: Die knie is die grootste, mees komplekse en mees ongelyksoortige gewrig in die liggaam. Osteoarthritis is ’n siekte wat algemeen by bejaardes voorkom en die versagting of agteruitgang van die kraakbeen en subchondrale bene in die gewrig tot gevolg het, wat tot ’n verlies van funksionering en pyn lei. Hierdie probleem kan verlig word deur ’n chirurgiese ingryping wat algemeen as ’n “knievervanging” bekend staan. Die doel van ’n knievervangingsprosedure is om pyn te verlig en normale funksionering te herstel. Ideaal gesproke behoort die knievervangingsprostese ’n gewrigsgeometrie te hê wat soortgelyk aan die pasiënt se gesonde knie is, en normale beweging moontlik maak. Ongelukkig is dit dikwels problematies aangesien knieprosteses in standaardgroottes en deur ’n verskeidenheid vervaardigers verskaf word, wat elkeen se ontwerp effens anders maak. Verder kan kommersiële prosteses nie altyd die komplekse geometrie van ’n individuele pasiënt se oorspronklike gewrigsoppervlakke vervang nie. Hierdie proefskrif wys dat daar ’n betekenisvolle variasie tussen knieafmetings is, afgesien van geslag en ras. Hierdie navorsing is daarop gemik om die problem op tweërlei wyse te benader: Eerstens deur ’n metode aan te bied om die optimal knieprostesetipe en -grootte vir ’n spesifieke pasiënt voor die operasie uit te soek, en tweedens om ’n ontwerpprosedure aan te bied vir die ontwerp en vervaardiging van pasiëntspesifieke unikompartementele knievervangings. Die ontwerpprosedure gebruik wiskundige modellering en ’n kunsmatige neurale netwerk om die oorspronklike en gesonde gewrigsoppervlakke van ’n pasiënt se knie te bepaal. Die modelle word met mediese beelde van die pasiënt gekombineer om ’n knieprostese te skep wat pasiëntspesifiek is. Hierdie pasiëntspesifieke inplantings word dan met konvensionele inplantings vergelyk wat kontakstres en kinematika betref. Daar word tot die slotsom gekom dat die pasiëntspesifieke inplantings oor eienskappe kan beskik wat vergelykbaar is met of selfs beter is as dié van konvensionele prosteses. Die unieke ontwerpmetodologie wat in hierdie proefskrif aangebied word, stel beduidende vordering in knievervangingstegnologie bekend, met die potensiaal om die kliniese uitkomste van knievervangingsoperasies dramaties te verbeter.

Lithological core logging is a subjective and time consuming endeavour which could possibly be automated, the question is if and to what extent this automation would affect the resulting core logs. This study presents a case from the Zinkgruvan Zn-Pb-Ag mine, Bergslagen, Sweden; in which Classification and Regression Trees and K-means Clustering on the Self Organising Map were applied to X-Ray Flourescence lithogeochemistry data derived from automated core scan technology. These two methods are assessed through comparison to manual core logging. It is found that the X-Ray Fluorescence data are not sufficiently accurate or precise for the purpose of automated full lithological classification since not all elements are successfully quantified. Furthermore, not all lithologies are possible to distinquish with lithogeochemsitry alone furter hindering the success of automated lithological classification. This study concludes that; 1) K-means on the Self Organising Map is the most successful approach, however; this may be influenced by the method of domain validation, 2) the choice of ground truth for learning is important for both supervised learning and the assessment of machine learning accuracy and 3) geology, data resolution and choice of elements are important parameters for machine learning. Both the supervised method of Classification and Regression Trees and the unsupervised method of K-means clustering applied to Self Organising Maps show potential to assist core logging procedures.

This thesis proposes a novel approach to extracting pose information from image sequences. Current state of the art techniques focus exclusively on the image space occupied by the body for pose and action recognition. The method proposed here, however, focuses on the negative spaces: the areas surrounding the individual. This has resulted in the colour-coded negative space approach, an image preprocessing step that circumvents the need for complicated model fitting or template matching methods. The approach can be described as follows: negative spaces surrounding the human silhouette are extracted using horizontal and vertical scanning processes. These negative space areas are more numerous, and undergo more radical changes in shape than the single area occupied by the figure of the person performing an action. The colour-coded negative space representation is formed using the four binary images produced by the scanning processes. Features are then extracted from the colour-coded images. These are based on the percentage of area occupied by distinct coloured regions as well as the bounding box proportions. Pose clusters are identified using feedback from an independent action set. Subsequent images are classified using a simple Euclidean distance measure. An image sequence is thus temporally segmented into its corresponding pose representations. Action recognition simply becomes the detection of a temporally ordered sequence of poses that characterises the action. The method is purely vision-based, utilising monocular images with no need for body markers or special clothing. Two datasets were constructed using several actors performing different poses and actions. Some of these actions included actors waving their arms, sitting down or kicking a leg. These actions were recorded against a monochrome background to simplify the segmentation of the actors from the background. The actions were then recorded on DV cam and digitised into a data base. The silhouette images from these actions were isolated and placed in a frame or bounding box. The next step was to highlight the negative spaces using a directional scanning method. This scanning method colour-codes the negative spaces of each action. What became immediately apparent is that very distinctive colour patterns formed for different actions. To emphasise the action, different colours were allocated to negative spaces surrounding the image. For example, the space between the legs of an actor standing in a T - pose with legs apart would be allocated yellow, while the space below the arms were allocated different shades of green. The space surrounding the head would be different shades of purple. During an action when the actor moves one leg up in a kicking fashion, the yellow colour would increase. Inversely, when the actor closes his legs and puts them together, the yellow colour filling the negative space would decrease substantially. What also became apparent is that these coloured negative spaces are interdependent and that they influence each other during the course of an action. For example, when an actor lifts one of his legs, increasing the yellow-coded negative space, the green space between that leg and the arm decreases. This interrelationship between colours hold true for all poses and actions as presented in this thesis. In terms of pose recognition, it is significant that these colour coded negative spaces and the way the change during an action or a movement are substantial and instantly recognisable. Compare for example, looking at someone lifting an arm as opposed to seeing a vast negative space changing shape. In a controlled research environment, several actors were instructed to perform a number of different actions. After colour coding the negative spaces, it became apparent that every action can be recognised by a unique colour coded pattern. The challenge is to ascribe a numerical presentation, a mathematical quotation, to extract the essence of what is so visually apparent. The essence of pose recognition and it's measurability lies in the relationship between the colours in these negative spaces and how they impact on each other during a pose or an action. The simplest way of measuring this relationship is by calculating the percentage of each colour present during an action. These calculated percentages become the basis of pose and action recognition. By plotting these percentages on a graph confirms that the essence of these different actions and poses can in fact been captured and recognised. Despite variations in these traces caused by time differences, personal appearance and mannerisms, what emerged is a clear recognisable pattern that can be married to an action or different parts of an action. 7 Actors might lift their left leg, some slightly higher than others, some slower than others and these variations in terms of colour percentages would be recorded as a trace, but there would be very specific stages during the action where the traces would correspond, making the action recognisable.In conclusion, using negative space as a tool in human pose and tracking recognition presents an exiting research avenue because it is influenced less by variations such as difference in personal appearance and changes in the angle of observation. This approach is also simplistic and does not rely on complicated models and templates