Doctoral Degrees (Computer Science)

Permanent URI for this collectionhttps://hdl.handle.net/10413/7113

Browse

Now showing 1 - 20 of 38

Deep learning framework for speech emotion classification.
(2024) Akinpelu, Samson Adebisi.; Viriri, Serestina.
A robust deep learning-based approach for the recognition and classification of speech emotion is proposed in this research work. Emotion recognition and classification occupy a conspicuous position in human-computer interaction (HCI) and by extension, determine the reasons and justification for human action. Emotion plays a critical role in decision-making as well. Distinguishing among various emotions (angry, sad, happy, neutral, disgust, fear, and surprise) that exist from speech signals has however been a long-term challenge. There have been some limitations associated with existing deep learning techniques as a result of the complexity of features from human speech (sequential data) which consists of insufficient label datasets, Noise and Environmental Factors, Cross-cultural and Linguistic Differences, Speakers’ Variability and Temporal Dynamics. There is also a heavy reliance on huge parameter tunning, especially for millions of parameters before the model can learn the expected emotional features necessary for classification emotion, which often results in computational complexity, over-fitting, and poor generalization. This thesis presents an innovative deep learning framework-based approach for the recognition and classification of speech emotions. The deep learning techniques currently in use for speech-emotion classification are exhaustively and analytically reviewed in this thesis. This research models various approaches and architectures based on deep learning to build a framework that is dependable and efficient for classifying emotions from speech signals. This research proposes a deep transfer learning model that addresses the shortcomings of inadequate training datasets for the classification of speech emotions. The research also models advanced deep transfer learning in conjunction with a feature selection algorithm to obtain more accurate results regarding the classification of speech emotion. Speech emotion classification is further enhanced by combining the regularized feature selection (RFS) techniques and attention-based networks for the classification of speech emotion with a significant improvement in the emotion recognition results. The problem of misclassification of emotion is alleviated through the selection of salient features that are relevant to emotion classification from speech signals. By combining regularized feature selection with attention-based mechanisms, the model can better understand emotional complexities and outperform conventional ML model emotion detection algorithms. The proposed approach is very resilient to background noise and cultural differences, which makes it suitable for real-world applications. Having investigated the reasons behind the enormous computing resources required for many deep learning based methods, the research proposed a lightweight deep learning approach that can be deployed on low-memory devices for speech emotion classification. A redesigned VGGNet with an overall model size of 7.94MB is utilized, combined with the best-performing classifier (Random Forest). Extensive experiments and comparisons with other deep learning models (DenseNet, MobileNet, InceptionNet, and ResNet) over three publicly available speech emotion datasets show that the proposed lightweight model improves the performance of emotion classification with minimal parameter size. The research further devises a new method that minimizes computational complexity using a vision transformer (ViT) network for speech emotion classification. The ViT model’s capabilities allow the mel-spectrogram input to be fed into the model, allowing for the capturing of spatial dependencies and high-level features from speech signals that are suitable indicators of emotional states. Finally, the research proposes a novel transformer model that is based on shift-window for efficient classification of speech emotion on bi-lingual datasets. Because this method promotes feature reuse, it needs fewer parameters and works well with smaller datasets. The proposed model was evaluated using over 3000 speech emotion samples from the publicly available TESS, EMODB, EMOVO, and bilingual TESS-EMOVO datasets. The results showed 98.0%, 98.7%, and 97.0% accuracy, F1-Score, and precision, respectively, across the 7 classes of emotion.
Pancreatic cancer survival prediction using Deep learning techniques.
(2023) Bakasa, Wilson.; Viriri, Serestina.
Abstract available in PDF.
Forest image classification based on deep learning and ontologies.
(2024) Kwenda, Clopas.; Gwetu, Mandlenkosi Victor.; Fonou-Dombeu, Jean Vincent.
Forests contribute abundantly to nature’s natural resources and they significantly contribute to a wide range of environmental, socio-cultural, and economic benefits. Classifications of forest vegetation offer a practical method for categorising information about patterns of forest vegetation. This information is required to successfully plan for land use, map landscapes, and preserve natural habitats. Remote sensing technology has provided high spatio-temporal resolution images with many spectral bands that make conducting research in forestry easy. In that regard, artificial intelligence technologies assess forest damage. The field of remote sensing research is constantly adapting to leverage newly developed computational algorithms and increased computing power. Both the theory and the practice of remote sensing have significantly changed as a result of recent technological advancements, such as the creation of new sensors and improvements in data accessibility. Data-driven methods, including supervised classifiers (such as Random Forests) and deep learning classifiers, are gaining much importance in processing big earth observation data due to their accuracy in creating observable images. Though deep learning models produce satisfactory results, researchers find it difficult to understand how they make predictions because they are regarded as black-box in nature, owing to their complicated network structures. However, when inductive inference from data learning is taken into consideration, data-driven methods are less efficient in working with symbolic information. In data-driven techniques, the specialized knowledge that environmental scientists use to evaluate images obtained through remote sensing is typically disregarded. This limitation presents a significant obstacle for end users of Earth Observation applications who are accustomed to working with symbolic information, such as ecologists, agronomists, and other related professionals. This study advocates for the incorporation of ontologies in forest image classification owing to their ability in representing domain expert knowledge. The future of remote sensing science should be supported by knowledge representation techniques such as ontologies. The study presents a methodological framework that integrates deep learning techniques and ontologies with the aim of enhancing domain expert confidence as well as increasing the accuracy of forest image classification. In addressing this challenge, this study followed the following systematic steps (i) A critical review of existing methods for forest image classification (ii) A critical analysis of appropriate methods for forest image classification (iii) Development of the state-of-the-art model for forest image segmentation (iv) Design of a hybrid model of deep learning and machine learning model for forest image classification (v) A state-of-the-art ontological framework for forest image classification. The ontological framework was flexible to capture the expression of the domain expert knowledge. The ontological state-of-the-art model performed well as it achieved a classification accuracy of 96%, with a Root Mean Square Error of 0.532. The model can also be used in the fruit industry and supermarkets to classify fruits into their respective categories. It can also be potentially used to classify trees with respect to their species. As a way of enhancing confidence in deep learning models by domain experts, the study recommended the adoption of explainable artificial intelligence (XAI) methods because they unpack the process by which deep learning models reach their decision. The study also recommended the adoption of high-resolution networks (HRNets) as an alternative to traditional deep learning models, because they can convert low-resolution representation to high-resolution and have efficient block structures developed according to new standards and they are excellent at being used for feature extraction.
Deep learning for brain tumor segmentation and survival prediction.
(2024) Magadza, Tirivangani Batanai Hendrix Takura.; Viriri, Serestina.
A brain tumor is an abnormal growth of cells in the brain that multiplies uncontrolled. The death of people due to brain tumors has increased over the past few decades. Early diagnosis of brain tumors is essential in improving treatment possibilities and increasing the survival rate of patients. The life expectancy of patients with glioblastoma multiforme (GBM), the most malignant glioma, using the current standard of care is, on average, 14 months after diagnosis despite aggressive surgery, radiation, and chemotherapies. Despite considerable efforts in brain tumor segmentation research, patient diagnosis remains poor. Accurate segmentation of pathological regions may significantly impact treatment decisions, planning, and outcome monitoring. However, the large spatial and structural variability among brain tumors makes automatic segmentation a challenging problem, leaving brain tumor segmentation an open challenge that warrants further research endeavors. While several methods automatically segment brain tumors, deep learning methods are becoming widespread in medical imaging due to their resounding performance. However, the boost in performance comes at the cost of high computational complexity. Therefore, to improve the adoption rate of computer-assisted diagnosis in clinical setups, especially in developing countries, there is a need for more computational and memoryefficient models. In this research, using a few computational resources, we explore various techniques to develop deep learning models accurately for segmenting the different glioma sub-regions, namely the enhancing tumor, the tumor core, and the whole tumor. We quantitatively evaluate the performance of our proposed models against the state-of-the-art methods using magnetic resolution imaging (MRI) datasets provided by the Brain Tumor Segmentation (BraTS) Challenge. Lastly, we use segmentation labels produced by the segmentation task and MRI multimodal data to extract appropriate imaging/radiomic features to train a deep learning model for overall patient survival prediction.
Exploration of ear biometrics with deep learning.
(2024) Booysens, Aimee Anne.; Viriri, Serestina.
Biometrics is the recognition of a human using biometric characteristics for identification, which may be physiological or behavioural. Numerous models have been proposed to distinguish biometric traits used in multiple applications, such as forensic investigations and security systems. With the COVID-19 pandemic, facial recognition systems failed due to users wearing masks; however, human ear recognition proved more suitable as it is visible. This thesis explores efficient deep learning-based models for accurate ear biometrics recognition. The ears were extracted and identified from 2D profiles and facial images, focusing on both left and right ears. With the numerous datasets used, with particular mention of BEAR, EarVN1.0, IIT, ITWE and AWE databases. Many machine learning techniques were explored, such as Naïve Bayes, Decision Tree, K-Nearest Neighbor, and innovative deep learning techniques: Transformer Network Architecture, Lightweight Deep Learning with Model Compression and EfficientNet. The experimental results showed that the Transformer Network achieved a high accuracy of 92.60% and 92.56% with epochs of 50 and 90, respectively. The proposed ReducedFireNet Model reduces the input size and increases computation time, but it detects more robust ear features. The EfficientNet variant B8 achieved a classification accuracy of 98.45%. The results achieved are more significant than those of other works, with the highest achieved being 98.00%. The overall results showed that deep learning models can improve ear biometrics recognition when both ears are computed.
Post-quantum cloud security and data exchange using artificial intelligence.
(2023) Mosola, Napo Nathnael.; Blackledge, Jonathan Michael.; Dombeu, Jean Vincent Fonou.
This thesis investigates the application of plausible modern-day cryptographic solutions for securing the cloud and exchanging confidential data. The context followed is such that the strength of an encryption algorithm is based on the difficulty to cryptanalyse it. This means the more difficult the crypto-system is to cryptanalyse, the stronger and more trusted it is. The success of cryptanalysis on a cryptographic algorithm has been a function of the computational power available at the time of performing the cryptanalysis, without consideration of future innovations, specifically, without careful consideration of Moore’s law. A significant number of public-key crypto-systems can and will be compromised by a quantum computer coupled with the implementation of Shor’s algorithm. This has brought a lot of focus regarding research on cryptographic solutions post quantum computing (PQC) due to the following: ˆ cryptographic algorithms are based on the intractability of prime number factorisation using a conventional computing power; ˆ a quantum computer can factorize prime numbers with relative ease. In the past, the quantum computing paradigm was a hypothetical concept. Thus, cryptanalysis using quantum resources was a theoretical idea. This is no longer the case with the loom of quantum computers eminent. Consequently, prime number based encryption is becoming increasingly irrelevant. Low Qubit quantum computers now exist. Research and development in this area is growing. Hence the existence of the post-quantum cryptography paradigm. This paradigm is based on encryption algorithms developed and considered secure enough to withstand quantum attacks. Thus, the National Institute of Standards and Technology made a call for projects clustered under the Open Quantum Safe project (OQSP), which began in 2016. The ultimate goal of this project is development of future quantum resistant cryptographic algorithms for secure communication and data exchange. The OQSP aims to gather open source libraries which can be standalone or integrated into the public key encryption schemes to improve their security against ii quantum attacks in the quest to achieve quantum resistance. The major focus is placed on quantum key exchange (QKE). It is against this background that the material presented in this thesis reports on a spectrum of algorithms that are thought to be quantum resistant based on a coherence of ideas, methods, models and software implementation, trying to meet the NIST requirements and contributing to new knowledge in the field of cryptography. The aim is to provide confidentiality guarantees on cloud-hosted data as well as secure data exchange between communicating entities, while also tackling the cumbersome key exchange and management problem. The results show that the algorithms presented in this thesis introduce new ideas in cryptography and can be tested to withstand cryptanalytic quantum attacks, while a plausible encryption key distribution and management solution is proposed. In this context, the material presented in this thesis report on a spectrum of algorithms that are proposed to be quantum resistant based on a coherence of ideas, methods and software implementation, aimed at providing security of cloud-hosted data as well as data exchange between communicating entities. The cloud has a flexible, scalable and low cost properties. This is due to two concepts which are fundamental to cloud computing: ˆ virtualization; ˆ multi-occupancy. These above concepts have brought infinitely many benefits which make the cloud an attractive paradigm. Among the benefits are reduced capital and maintenance costs, high processing power, enormous storage facilities etc. However, security concerns affecting confidentiality of cloud-hosted data still plague bring concerns when it comes to cloud adoption. Data confidentiality can be achieved through encryption, which is in turn implemented by cryptographic algorithms. Hence, this thesis proposes and puts into practice cryptographic algorithms to solve issues of confidentiality, specifically in the cloud.
Search and selection methods for hyper-heuristics.
(2018) Akandwanaho, Stephen Mugisha.; Petruccione, Francesco.; Sinayskiy, Ilya.
Hyper-heuristics (HHs) search from a search space of heuristics for an optimal heuristic that can be mapped onto a problem to generate good solutions. One of the heuristic selection methods used by hyper-heuristics is a choice function (CF) which assigns scores to heuristics according to their performance. An investigation is conducted on a choice function for single-point selective hyper-heuristics. The drawbacks of the existing choice function are: discarding heuristics due to poor performance on a problem when they could exhibit good performance on another problem and premature convergence in the heuristic search space. In order to address the drawbacks of a choice function, a new selection method called an efficient choice function (ECF) is introduced, based on a three-pronged improvement approach. Firstly, a new element is introduced, which collects previously poor-performing heuris- tics. The best heuristic of the poorly performing heuristics is obtained and compared with the best heuristic from the general pool of heuristics. Secondly, the pairwise comparison of the best heuristics from both the poorly performing heuristics and the general pool of heuristics is applied at every point in the iteration to maintain competition through- out the search process and generate high-quality outcomes. Thirdly, another element is introduced that randomly divides heuristics into different groups, ranks the collective performance of each group and makes performance comparisons between disparate groups of heuristics. The proposed heuristic selection method is tested on several well-known combinatorial optimization problems which include, the vehicle routing problem, the bin packing problem, the permutation ow shop problem, the personnel scheduling problem and the patient transportation problem. Results show a better performance of an efficient choice function than the existing methods. The second contribution of this thesis is to enhance searching for optima in dynamic search spaces. An investigation of dynamic selection hyper-heuristics is performed and a new non-stationary covariance function is introduced. The Gaussian process regression is used as a predictive measure for moving optima in dynamic search environments and the proposed method is applied to the dynamic traveling salesman problem, which yields better performance than the existing approach. The third contribution is a spy search method (SSM) for a memetic algorithm (MA) in dynamic environments. Given that the proposed efficient choice function for hyper- heuristics is based on a memetic to perform the search for an optimal heuristic, improving a MA search enhances the capacity of the efficient choice function to nd good solutions. The proposed SSM shows a better performance than the nine existing methods on a set of different dynamic problems.
Retinal blood vessel segmentation using random forest Gabor feature selection and automatic thresholding.
(2019) Gwetu, Mandlenkosi Victor.; Tapamo, Jules-Raymond.; Viriri, Serestina.
Successful computer aided diagnosis of ocular diseases is normally dependent on the accurate detection of components such as blood vessels, optic disk, fovea and microaneurysms. The properties of these components can be indicative of the presence and/or severity of pathology. Since most prevalent forms of ocular diseases emanate from vascular disorders, it is expected that accurate detection of blood vessels is essential for ocular diagnosis. In this research work, we investigate several opportunities for improvement of retinal blood vessel segmentation with the hope that they will ultimately lead to improvement in the diagnosis of vascular related ocular diseases. We complement existing work in this domain by introducing new Gabor lter features and selecting the most e ective of these using Random Forests feature selection. The actual segmentation of blood vessels is then done using an improved automatic thresholding scheme based on the preferred Gabor feature. We propose Random Forest (RF) feature ranking algorithms that demonstrate reliable feature set partitions over several University of California, Irvine (UCI) datasets. To circumvent instances of unreliable rankings, we also propose feature rank and RF strength correlation as an alternative indicator. Of the four proposed Gabor features, the maximum magnitude response is con rmed as the most e ective, as is the general trend in previous literature. The proposed Selective Valley Emphasis thresholding technique achieves identical segmentation results to the legacy approach while improving on computational e ciency. Sensitivity and speci city outcomes of up to 76.8% and 97.9% as well as 78.8% and 97.8% are achieved on the DRIVE and STARE datasets, respectively.
A machine learning approach to facial-based ethnicity classification.
(2017) Momin, Hajra Mehbub.; Tapamo, Jules-Raymond.
The determination of ethnicity of an individual can be very useful in a face recognition and person identification system in general. The face displays a complex range of information about identity, age, sex, race as well as emotional and intentional state. It is commonly assumed that the biological unit of human classification is the ethnic group, with hereditary physical features making up the group classification, based on the qualities such as the skin colour, the build, the head shape, the hair, the face shape, and the blood type. In this thesis, the aim is to investigate methods and techniques to perform ethnicity classification of face images. Automatic face-based ethnicity classification has various applications in human computer interaction, surveillance, video and image retrieval, database indexing, and can give helpful insight for face recognition and identification. Since biometric systems have to deal with very large databases, it can be a good idea to partition the face database according to the ethnicity of a person. In addition, this has the potential to significantly improve the search speed, efficiency and accuracy of biometric systems. Automatic face and landmark detection on images is very important for face recognition, face identification and for ethnicity classification. This study presents an approach for detecting face and facial features such as the eyes, the nose and the mouth in gray-scale images. In addition, the study makes use of thresholding and connected component labelling algorithms in order to detect a face and extract features that characterize this face. This study investigates three different feature methods for the ethnicity classification of face images. A new ethnicity classification based on skin colour is proposed. Skin colour is one of the most important features in the human face. The skin colour differs from individual to individual belonging to different ethnic groups and from people across different regions. For instance, theskin colour of people belonging to White, Asian and Black groups is different from one another and extended from white to yellow to dark brown. Based on this different colour spaces are used to create a feature vector representing a given face image. A second feature model based on textures is proposed. Gabor filters are used to extract texture features. Thirdly, a combination of colour and texture features are used to further improve the ethnicity classification accuracy. Four different classifiers, namely K-Means clustering, Naive Bayesian (NB), Multilayer Perceptron (MLP) and Support Vector Machine (SVM), were used to test the effectiveness of the automatic characterization of ethnicity by using the proposed features models. The ethnic groups considered were Asian, Indian, White and Black. Extensive experiments demonstrate that our models achieve very good results, confirming the consistently overwhelming performance of Asian classification. The proposed models also achieve very good classification results for different ethnic groups when compared with existing models.
Mathematical and numerical analysis of the discrete fragmentation coagulation equation with growth, decay and sedimentation.
(2018) Joel, Luke Oluwaseye; Banasiak, Jacek.; Shindin, Sergey Konstantinovich.
Fragmentation-coagulation equations arise naturally in many branches of engineering and science, the applications stretching from astrophysics, blood clotting, colloidal chemistry and polymer science to molecular beam epitaxy. In realistic application, the fragmentation and coagulation are often coupled with growth, decay and/or sedimentation processes. The resulting models are used to describe the evolution of a population in which individuals can grow, coalesce, split or divide, and die. For example, in the phytoplankton dynamics, in addition to forming or breaking of clusters, individuals within them are born or die and so the latter processes must be adequately represented in the models. In the continuous case, the birth or death processes are incorporated into the model by adding an appropriate first order transport term, analogously to the age and size structured McKendrick models. In the discrete case, these vital processes are modelled by adding weighted differences operators. In this study, we focus on the discrete fragmentation-coagulation models with growth, decay or/and sedimentation. The problem is treated as an infinite-dimensional differential equation, which consists of a linear part (fragmentation, growth, decay and sedimentation term) and a nonlinear part (coagulation term), posed in a suitable Banach space, X. We use the theory of semigroups of linear operators, perturbation of positive semigroups and semilinear operators for the mathematical analysis of these models. The linear part of the models is shown to generate a semigroup which is analytic, compact and irreducible and thus has the asynchronous exponential growth property. These results are used to demonstrate the existence of global classical solutions to the semilinear fragmentation-coagulation equations with growth, decay and sedimentation for a class of unbounded coagulation kernels. Theoretical conclusions are supported by numerical simulations.
Modelling of artificial intelligence based demand side management techniques for mitigating energy poverty in smart grids.
(2018) Monyei, Chukwuka Gideon.; Viriri, Serestina.
This research work proposes an artificial intelligence (AI) based model for smart grid initiatives (for South Africa and by extension sub-Saharan Africa, (SSA)) and further incorporates energy justice principles. Spanning the social, technical, economic, environmental, policy and overall impact of smart and just electricity grids, this research begins by investigating declining electricity consumption and demand side management (DSM) potential across South Africa. In addition, technical frameworks such as the combined energy management system (CEMS), co-ordinated centralized energy management system (ConCEMS) and biased load manager home energy management system (BLM-HEMS) are modelled. These systems provide for the integration of all aspects of the electricity grid and their optimization in achieving cost reduction for both the utility and consumers as well as improvement in the consumers quality of life (QoL) and reduction of emissions. Policy and economy-wise, this research work further proposes and models an integrated electrification and expansion model (IEEM) for South Africa, and also addresses the issue of rural marginalization due to poor electricity access for off-grid communities. This is done by proposing a hybrid generation scheme (HGS) which is shown to satisfy sufficiently the requirements of the energy justice framework while significantly reducing the energy burden of households and reducing carbon emissions by over 70%.
Structure based partial solution search for the examination timetabling problem.
(2021) Rajah, Christopher Bradley.; Pillay, Nelishia.
The aim of this work is to present a new approach, namely, Structure Based Partial Solution Search (SBPSS) to solve the Examination Timetabling Problem. The success of the Developmental Approach in this problem domain suggested that the strategy of searching the spaces of partial timetables whilst constructing them is promising and worth pursuing. This work adopts a similar strategy. Multiple timetables are incrementally constructed at the same time. The quality of the partial timetables is improved upon by searching their partial solution spaces at every iteration during construction. Another key finding from the literature survey revealed that although timetables may exhibit the same behaviour in terms of their objective values, their structures or exam schedules may be different. The challenge with this finding is to decide on which regions to pursue because some regions may not be worth investigating due to the difficulty in searching them. These problematic areas may have solutions that are not amenable to change which makes it difficult to improve them. Another reason is that the neighbourhoods of solutions in these areas may be less connected than others which may restrict the ability of the search to move to a better solution in that neighbourhood. By moving to these problematic areas of the search space the search may stagnate and waste expensive computational resources. One way to overcome this challenge is to use both structure and behaviour in the search and not only behaviour alone to guide the search. A search that is guided by structure is able to find new regions by considering the structural components of the candidate solutions which indicate which part of the search space the same candidates occupy. Another benefit to making use of a structure-based search is that it has no objective value bias because it is not guided by only the objective value. This statement is consistent with the literature survey where it is suggested that in order to achieve good performance the search should not be guided by only the objective value. The proposed method has been tested on three popular benchmark sets for examination timetabling, namely, the Carter benchmark set; the benchmark set from the International Timetabling competition in 2007 and the Yeditepe benchmark set. The SBPSS found the best solutions for two of the Carter problem instances. The SBPSS found the best solutions for four of the competition problem instances. Lastly, the SBPSS improved on the best results for all the Yeditepe problem instances.
Tuberculosis diagnosis from pulmonary chest x-ray using deep learning.
(2022) Oloko-Oba, Mustapha Olayemi.; Viriri, Serestina.
Tuberculosis (TB) remains a life-threatening disease, and it is one of the leading causes of mortality in developing countries. This is due to poverty and inadequate medical resources. While treatment for TB is possible, it requires an accurate diagnosis first. Several screening tools are available, and the most reliable is Chest X-Ray (CXR), but the radiological expertise for accurately interpreting the CXR images is often lacking. Over the years, CXR has been manually examined; this process results in delayed diagnosis, is time-consuming, expensive, and is prone to misdiagnosis, which could further spread the disease among individuals. Consequently, an algorithm could increase diagnosis efficiency, improve performance, reduce the cost of manual screening and ultimately result in early/timely diagnosis. Several algorithms have been implemented to diagnose TB automatically. However, these algorithms are characterized by low accuracy and sensitivity leading to misdiagnosis. In recent years, Convolutional Neural Networks (CNN), a class of Deep Learning, has demonstrated tremendous success in object detection and image classification task. Hence, this thesis proposed an efficient Computer-Aided Diagnosis (CAD) system with high accuracy and sensitivity for TB detection and classification. The proposed model is based firstly on novel end-to-end CNN architecture, then a pre-trained Deep CNN model that is fine-tuned and employed as a features extractor from CXR. Finally, Ensemble Learning was explored to develop an Ensemble model for TB classification. The Ensemble model achieved a new stateof- the-art diagnosis accuracy of 97.44% with a 99.18% sensitivity, 96.21% specificity and 0.96% AUC. These results are comparable with state-of-the-art techniques and outperform existing TB classification models.
Automatic dental caries detection in bitewing radiographs.
(2022) Majanga, Vincent Idah.; Viriri, Serestina.
Dental Caries is one of the most prevalent chronic disease around the globe. Distinguishing carious lesions has been a challenging task. Conventional computer aided diagnosis and detection methods in the past have heavily relied on visual inspection of teeth. These are only effective on large and clearly visible caries on affected teeth. Conventional methods have been limited in performance due to the complex visual characteristics of dental caries images, which consists of hidden or inaccessible lesions. Early detection of dental caries is an important determinant for treatment and benefits much from the introduction of new tools such as dental radiography. A method for the segmentation of teeth in bitewing X-rays is presented in this thesis, as well as a method for the detection of dental caries on X-ray images using a supervised model. The diagnostic method proposed uses an assessment protocol that is evaluated according to a set of identifiers obtained from a learning model. The proposed technique automatically detects hidden and inaccessible dental caries lesions in bitewing radio graphs. The approach employed data augmentation to increase the number of images in the data set in order to have a total of 11,114 dental images. Image pre-processing on the data set was through the use of Gaussian blur filters. Image segmentation was handled through thresholding, erosion and dilation morphology, while image boundary detection was achieved through active contours method. Furthermore, the deep learning based network through the sequential model in Keras extracts features from the images through blob detection. Finally, a convexity threshold value of 0.9 is introduced to aid in the classification of caries as either present or not present. The relative efficacy of the supervised model in diagnosing dental caries when compared to current systems is indicated by the results detailed in this thesis. The proposed model achieved a 97% correct diagnostic which proved quite competitive with existing models.
Facial expression recognition and intensity estimation.
(2022) Ekundayo, Olufisayo Sunday.; Viriri, Serestina.
Facial Expression is one of the profound non-verbal channels through which human emotion state is inferred from the deformation or movement of face components when facial muscles are activated. Facial Expression Recognition (FER) is one of the relevant research fields in Computer Vision (CV) and Human-Computer Interraction (HCI). Its application is not limited to: robotics, game, medical, education, security and marketing. FER consists of a wealth of information. Categorising the information into primary emotion states only limit its performance. This thesis considers investigating an approach that simultaneously predicts the emotional state of facial expression images and the corresponding degree of intensity. The task also extends to resolving FER ambiguous nature and annotation inconsistencies with a label distribution learning method that considers correlation among data. We first proposed a multi-label approach for FER and its intensity estimation using advanced machine learning techniques. According to our findings, this approach has not been considered for emotion and intensity estimation in the field before. The approach used problem transformation to present FER as a multilabel task, such that every facial expression image has unique emotion information alongside the corresponding degree of intensity at which the emotion is displayed. A Convolutional Neural Network (CNN) with a sigmoid function at the final layer is the classifier for the model. The model termed ML-CNN (Multilabel Convolutional Neural Network) successfully achieve concurrent prediction of emotion and intensity estimation. ML-CNN prediction is challenged with overfitting and intraclass and interclass variations. We employ Visual Geometric Graphics-16 (VGG-16) pretrained network to resolve the overfitting challenge and the aggregation of island loss and binary cross-entropy loss to minimise the effect of intraclass and interclass variations. The enhanced ML-CNN model shows promising results and outstanding performance than other standard multilabel algorithms. Finally, we approach data annotation inconsistency and ambiguity in FER data using isomap manifold learning with Graph Convolutional Networks (GCN). The GCN uses the distance along the isomap manifold as the edge weight, which appropriately models the similarity between adjacent nodes for emotion predictions. The proposed method produces a promising result in comparison with the state-of-the-art methods.
The adoption of Web 2.0 tools in teaching and learning by in-service secondary school teachers: the Mauritian context.
(2018) Pyneandee, Marday.; Govender, Desmond Wesley.; Oogarah-Pratap, Brinda.
With the current rapid increase in use of Web 2.0 tools by students, it is becoming necessary for teachers to understand what is happening in this social networking phenomenon, so that they can better understand the new spaces that students inhabit and the implications for students’ learning and investigate the wealth of available Web 2.0 tools, and work to incorporate some into their pedagogical and learning practices. Teachers are using the Internet and social networking tools in their personal lives. However, there is little empirical evidence on teachers’ viewpoints and usage of social media and other online technologies to support their classroom practice. This study stemmed from the urgent need to address this gap by exploring teachers’ perceptions, and experience of the integration of online technologies, social media, in their personal lives and for professional practice to find the best predictors of the possibility of teachers’ using Web 2.0 tools in their professional practice. Underpinning the study is a conceptual framework consisting of core ideas found in the unified theory of acceptance and use of technology (UTAUT) and technology pedagogy and content knowledge (TPACK) models. The conceptual framework, together with a review of relevant literature, enabled the formulation of a theoretical model for understanding teachers’ intention to exploit the potential of Web 2.0 tools. The model was then further developed using a mixed-method, two-phase methodology. In the first phase, a survey instrument was designed and distributed to in-service teachers following a Postgraduate Certificate in Education course at the institution where the researcher works. Using the data collected from the survey, exploratory factor analysis, correlational analysis and multiple regression analysis were used to refine the theoretical model. Other statistical methods were also used to gain further insights into teachers’ perceptions of use of Web 2.0 tools in their practices. In the second phase of the study, survey respondents were purposefully selected, based on quantitative results, to participate in interviews. The qualitative data yielded from the interviews was used to support and enrich understanding of the quantitative findings. The constructs teacher knowledge and technology pedagogy knowledge from the TPACK model and the constructs effort expectancy, facilitating conditions and performance expectancy are the best predictors of teachers’ intentions to use Web 2.0 tools in their professional practice. There was an interesting finding on the relationship between UTAUT and TPACK constructs. The constructs performance expectancy and effort expectancy had a significant relationship with all the TPACK constructs – technology knowledge, technology pedagogy knowledge, pedagogical content knowledge (PCK), technology and content knowledge and TPACK – except for content knowledge and pedagogical knowledge. The association between the TPACK construct PCK with the UTAUT constructs performance expectancy and effort expectancy was an unexpected finding because PCK is only about PCK and has no technology component. The theoretical contribution of this study is the model, which is teachers’ intention of future use of Web 2.0 tools in their professional practice. The predictive model, together with other findings, enhances understanding of the nature of teachers’ intention to utilise Web 2.0 tools in their professional practice. Findings from this study have implications for school infrastructure, professional development of teachers and an ICT learning environment to support the adoption of Web 2.0 tools in teaching practices and are presented as guiding principles at the end of the study.
Automated design of genetic programming of classification algorithms.
(2018) Nyathi, Thambo.; Pillay, Nelishia.
Over the past decades, there has been an increase in the use of evolutionary algorithms (EAs) for data mining and knowledge discovery in a wide range of application domains. Data classification, a real-world application problem is one of the areas EAs have been widely applied. Data classification has been extensively researched resulting in the development of a number of EA based classification algorithms. Genetic programming (GP) in particular has been shown to be one of the most effective EAs at inducing classifiers. It is widely accepted that the effectiveness of a parameterised algorithm like GP depends on its configuration. Currently, the design of GP classification algorithms is predominantly performed manually. Manual design follows an iterative trial and error approach which has been shown to be a menial, non-trivial time-consuming task that has a number of vulnerabilities. The research presented in this thesis is part of a large-scale initiative by the machine learning community to automate the design of machine learning techniques. The study investigates the hypothesis that automating the design of GP classification algorithms for data classification can still lead to the induction of effective classifiers. This research proposes using two evolutionary algorithms,namely,ageneticalgorithm(GA)andgrammaticalevolution(GE)toautomatethe design of GP classification algorithms. The proof-by-demonstration research methodology is used in the study to achieve the set out objectives. To that end two systems namely, a genetic algorithm system and a grammatical evolution system were implemented for automating the design of GP classification algorithms. The classification performance of the automated designed GP classifiers, i.e., GA designed GP classifiers and GE designed GP classifiers were compared to manually designed GP classifiers on real-world binary class and multiclass classification problems. The evaluation was performed on multiple domain problems obtained from the UCI machine learning repository and on two specific domains, cybersecurity and financial forecasting. The automated designed classifiers were found to outperform the manually designed GP classifiers on all the problems considered in this study. GP classifiers evolved by GE were found to be suitable for classifying binary classification problems while those evolved by a GA were found to be suitable for multiclass classification problems. Furthermore, the automated design time was found to be less than manual design time. Fitness landscape analysis of the design spaces searched by a GA and GE were carried out on all the class of problems considered in this study. Grammatical evolution found the search to be smoother on binary classification problems while the GA found multiclass problems to be less rugged than binary class problems.
The enhanced best performance algorithm for global optimization with applications.
(2016) Chetty, Mervin.; Adewumi, Aderemi Oluyinka.
Abstract available in PDF file.
Leaf recognition for accurate plant classification.
(2017) Kala, Jules Raymond.; Viriri, Serestina.; Moodley, Deshendran.
Plants are the most important living organisms on our planet because they are sources of energy and protect our planet against global warming. Botanists were the first scientist to design techniques for plant species recognition using leaves. Although many techniques for plant recognition using leaf images have been proposed in the literature, the precision and the quality of feature descriptors for shape, texture, and color remain the major challenges. This thesis investigates the precision of geometric shape features extraction and improved the determination of the Minimum Bounding Rectangle (MBR). The comparison of the proposed improved MBR determination method to Chaudhuri's method is performed using Mean Absolute Error (MAE) generated by each method on each edge point of the MBR. On the top left point of the determined MBR, Chaudhuri's method has the MAE value of 26.37 and the proposed method has the MAE value of 8.14. This thesis also investigates the use of the Convexity Measure of Polygons for the characterization of the degree of convexity of a given leaf shape. Promising results are obtained when using the Convexity Measure of Polygons combined with other geometric features to characterize leave images, and a classification rate of 92% was obtained with a Multilayer Perceptron Neural Network classifier. After observing the limitations of the Convexity Measure of Polygons, a new shape feature called Convexity Moments of Polygons is presented in this thesis. This new feature has the invariant properties of the Convexity Measure of Polygons, but is more precise because it uses more than one value to characterize the degree of convexity of a given shape. Promising results are obtained when using the Convexity Moments of Polygons combined with other geometric features to characterize the leaf images and a classification rate of 95% was obtained with the Multilayer Perceptron Neural Network classifier. Leaf boundaries carry valuable information that can be used to distinguish between plant species. In this thesis, a new boundary-based shape characterization method called Sinuosity Coefficients is proposed. This method has been used in many fields of science like Geography to describe rivers meandering. The Sinuosity Coefficients is scale and translation invariant. Promising results are obtained when using Sinuosity Coefficients combined with other geometric features to characterize the leaf images, a classification rate of 80% was obtained with the Multilayer Perceptron Neural Network classifier. Finally, this thesis implements a model for plant classification using leaf images, where an input leaf image is described using the Convexity Moments, the Sinuosity Coefficients and the geometric features to generate a feature vector for the recognition of plant species using a Radial Basis Neural Network. With the model designed and implemented the overall classification rate of 97% was obtained.
A semantic sensor web framework for proactive environmental monitoring and control.
(2017) Adeleke, Jude Adekunle.; Moodley, Deshendran.; Rens, Gavin Brian.; Adewumi, Aderemi Oluyinka.
Observing and monitoring of the natural and built environments is crucial for main- taining and preserving human life. Environmental monitoring applications typically incorporate some sensor technology to continually observe specific features of inter- est in the physical environment and transmitting data emanating from these sensors to a computing system for analysis. Semantic Sensor Web technology supports se- mantic enrichment of sensor data and provides expressive analytic techniques for data fusion, situation detection and situation analysis. Despite the promising successes of the Semantic Sensor Web technology, current Semantic Sensor Web frameworks are typically focused at developing applications for detecting and reacting to situations detected from current or past observations. While these reactive applications provide a quick response to detected situations to minimize adverse effects, they are limited when it comes to anticipating future adverse situations and determining proactive control actions to prevent or mitigate these situations. Most current Semantic Sensor Web frameworks lack two essential mechanisms required to achieve proactive control, namely, mechanisms for antici- pating the future and coherent mechanisms for consistent decision processing and planning. Designing and developing proactive monitoring and control Semantic Sensor Web applications is challenging. It requires incorporating and integrating different tech- niques for supporting situation detection, situation prediction, decision making and planning in a coherent framework. This research proposes a coherent Semantic Sen- sor Web framework for proactive monitoring and control. It incorporates ontology to facilitate situation detection from streaming sensor observations, statistical ma- chine learning for situation prediction and Markov Decision Processes for decision making and planning. The efficacy and use of the framework is evaluated through the development of two different prototype applications. The first application is for proactive monitoring and control of indoor air quality to avoid poor air quality situations. The second is for proactive monitoring and control of electricity usage in blocks of residential houses to prevent strain on the national grid. These appli- cations show the effectiveness of the proposed framework for developing Semantic Sensor Web applications that proactively avert unwanted environmental situations before they occur.

Browse

Recent Submissions