2023
Ahmad, Amar S; Al-Hassan, Munther; Hussain, Hamid Y; Juber, Nirmin F; Kiwanuka, Fred N; Hag-Ali, Mohammed; Ali, Raghib
A method of correction for heaping error in the variables using validation data Journal Article
In: Statistical Papers, pp. 1–18, 2023.
Abstract | Links | BibTeX | Tags:
@article{ahmad2023method,
title = {A method of correction for heaping error in the variables using validation data},
author = {Amar S Ahmad and Munther Al-Hassan and Hamid Y Hussain and Nirmin F Juber and Fred N Kiwanuka and Mohammed Hag-Ali and Raghib Ali},
url = {https://link.springer.com/article/10.1007/s00362-023-01405-4},
year = {2023},
date = {2023-01-01},
urldate = {2023-01-01},
journal = {Statistical Papers},
pages = {1–18},
publisher = {Springer},
abstract = {When self-reported data are used in statistical analysis to estimate the mean and variance, as well as the regression parameters, the estimates tend, in many cases, to be biased. This is because interviewees have a tendency to heap their answers to certain values. The aim of the paper is to examine the bias-inducing effect of the heaping error in self-reported data, and study the effect on the heaping error on the mean and variance of a distribution as well as the regression parameters. As a result a new method is introduced to correct the effects of bias due to the heaping error using validation data. Using publicly available data and simulation studies, it can be shown that the newly developed method is practical and can easily be applied to correct the bias in the estimated mean and variance, as well as in the estimated regression parameters computed from self-reported data. Hence, using the method of correction presented in this paper allows researchers to draw accurate conclusions leading to the right decisions, e.g. regarding health care planning and delivery.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
When self-reported data are used in statistical analysis to estimate the mean and variance, as well as the regression parameters, the estimates tend, in many cases, to be biased. This is because interviewees have a tendency to heap their answers to certain values. The aim of the paper is to examine the bias-inducing effect of the heaping error in self-reported data, and study the effect on the heaping error on the mean and variance of a distribution as well as the regression parameters. As a result a new method is introduced to correct the effects of bias due to the heaping error using validation data. Using publicly available data and simulation studies, it can be shown that the newly developed method is practical and can easily be applied to correct the bias in the estimated mean and variance, as well as in the estimated regression parameters computed from self-reported data. Hence, using the method of correction presented in this paper allows researchers to draw accurate conclusions leading to the right decisions, e.g. regarding health care planning and delivery.
Amin, Anang Hudaya Muhamad; Kiwanuka, Fred N; Abdelmajid, Nabih TJ; AlKaabi, Saif Hamad; Ahli, Sultan Khalid Abdulqader Rashed
Composite Identity of Things (CIDoT) on Permissioned Blockchain Network for Identity Management of IoT Devices Book Section
In: Research Anthology on Convergence of Blockchain, Internet of Things, and Security, pp. 382–401, IGI Global, 2023.
Abstract | Links | BibTeX | Tags:
@incollection{amin2023composite,
title = {Composite Identity of Things (CIDoT) on Permissioned Blockchain Network for Identity Management of IoT Devices},
author = {Anang Hudaya Muhamad Amin and Fred N Kiwanuka and Nabih TJ Abdelmajid and Saif Hamad AlKaabi and Sultan Khalid Abdulqader Rashed Ahli},
url = {https://www.igi-global.com/chapter/composite-identity-of-things-cidot-on-permissioned-blockchain-network-for-identity-management-of-iot-devices/310459},
year = {2023},
date = {2023-01-01},
urldate = {2023-01-01},
booktitle = {Research Anthology on Convergence of Blockchain, Internet of Things, and Security},
pages = {382–401},
publisher = {IGI Global},
abstract = {Internet of things (IoT) is in the forefront of many existing smart applications, including autonomous systems and green technology. IoT devices have been commonly used in the monitoring of energy efficiency and process automation. As the application spreads across different kinds of applications and technology, a large number of IoT devices need to be managed and configured, as they are capable of generating massive amount of sensory data. Looking from this perspective, there is a need for a proper mechanism to identify each IoT devices within the system and their respective applications. Participation of these IoT devices in complex systems requires a tamper-proof identity to be generated and stored for the purpose of device identification and verification. This chapter presents a comprehensive approach on identity management of IoT devices using a composite identity of things (CIDoT) with permissioned},
keywords = {},
pubstate = {published},
tppubtype = {incollection}
}
Internet of things (IoT) is in the forefront of many existing smart applications, including autonomous systems and green technology. IoT devices have been commonly used in the monitoring of energy efficiency and process automation. As the application spreads across different kinds of applications and technology, a large number of IoT devices need to be managed and configured, as they are capable of generating massive amount of sensory data. Looking from this perspective, there is a need for a proper mechanism to identify each IoT devices within the system and their respective applications. Participation of these IoT devices in complex systems requires a tamper-proof identity to be generated and stored for the purpose of device identification and verification. This chapter presents a comprehensive approach on identity management of IoT devices using a composite identity of things (CIDoT) with permissioned
2021
Kiwanuka, Fred N; Abuelmaatti, Omar Eltaher; Amin, Anang Hudaya Muhamad; Mukwaya, Brian J
Tropical Skin Disease Classification using Connected Attribute Filters. Proceedings Article
In: VISIGRAPP (5: VISAPP), pp. 338–345, 2021.
Abstract | Links | BibTeX | Tags:
@inproceedings{kiwanuka2021tropical,
title = {Tropical Skin Disease Classification using Connected Attribute Filters.},
author = {Fred N Kiwanuka and Omar Eltaher Abuelmaatti and Anang Hudaya Muhamad Amin and Brian J Mukwaya},
url = {https://pdfs.semanticscholar.org/0a62/87049ac53fcdbd5cbd077f84e63099e81784.pdf},
year = {2021},
date = {2021-01-01},
urldate = {2021-01-01},
booktitle = {VISIGRAPP (5: VISAPP)},
pages = {338–345},
abstract = {Morphological connected filters operate on an image through flat zones which comprise the largest connected components with a constant signal. These filters identify and ultimately extract the whole connected components in an image without alteration of their boundaries and thus shape preserving. This is a desirable property in many image processing and analysis applications. However, due to the variability of the number of connected components, even in the case of images of the same resolution and size, their application in classification tasks has been limited. In this study, we propose an approach that computes the shape and size features of connected components and use these features for the classification of bacterial and viral tropical skin infections. We demonstrate the performance of the approach using gradient boosting machines and compare the results to deep learning approaches. Results show that the performance of our approach is comparable to that of Convolutional Neural Networks (CNN) based approach when trained on 1460 images. Moreover, CNN was pre-trained and required augmentation to achieve that perfomance. However, our approach is at least 56% faster than CNN.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Morphological connected filters operate on an image through flat zones which comprise the largest connected components with a constant signal. These filters identify and ultimately extract the whole connected components in an image without alteration of their boundaries and thus shape preserving. This is a desirable property in many image processing and analysis applications. However, due to the variability of the number of connected components, even in the case of images of the same resolution and size, their application in classification tasks has been limited. In this study, we propose an approach that computes the shape and size features of connected components and use these features for the classification of bacterial and viral tropical skin infections. We demonstrate the performance of the approach using gradient boosting machines and compare the results to deep learning approaches. Results show that the performance of our approach is comparable to that of Convolutional Neural Networks (CNN) based approach when trained on 1460 images. Moreover, CNN was pre-trained and required augmentation to achieve that perfomance. However, our approach is at least 56% faster than CNN.
Kiwanuka, Fred N; Karadsheh, Louay; Amin, Anang Hudaya Muhamad; others,
Modeling Employee Flexible Work Scheduling As A Classification Problem Journal Article
In: Procedia Computer Science, vol. 192, pp. 3281–3290, 2021.
Abstract | Links | BibTeX | Tags:
@article{kiwanuka2021modeling,
title = {Modeling Employee Flexible Work Scheduling As A Classification Problem},
author = {Fred N Kiwanuka and Louay Karadsheh and Anang Hudaya Muhamad Amin and others},
url = {https://www.sciencedirect.com/science/article/pii/S1877050921018408},
year = {2021},
date = {2021-01-01},
urldate = {2021-01-01},
journal = {Procedia Computer Science},
volume = {192},
pages = {3281–3290},
publisher = {Elsevier},
abstract = {Many organizations have adapted flexible working arrangements during COVID19 pandemic because of restrictions on the number of employees required on site at any time. Unfortunately, current employee scheduling methods are more suited for compressed working arrangements. The problem of automating compressed employee scheduling has been studied by many researchers and is widely adopted by many organizations in an attempt to achieve high quality scheduling. During process of employee scheduling many constraints may have to be considered and may require negotiating a large dimension of constraints like in flexible working. These constraints make scheduling a challenging task in these working arrangements. Most scheduling algorithms are modeled as constraint optimization problems and suited for compressed work but for flexible working with large constraint dimensions, achieving accurate scheduling is even more challenging. In this research, we propose a machine learning approach that takes advantage of mining user-defined constraints or soft constraints and transform employee scheduling into a classification problem. We propose automatically extracting employee personal schedules like calendars in order to extract their availability. We then show how to use the extracted knowledge in a multi-label classification approach in order to generate a schedule for faculty staff in a University that supports flexible working. We show that the results of this approach are comparable to that of a constraint satisfaction and optimization method that is commonly used in literature. Results show that our approach achieved accuracy of 93.1% of satisfying constraints as compared to 92.7% of a common constraint programming approach.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Many organizations have adapted flexible working arrangements during COVID19 pandemic because of restrictions on the number of employees required on site at any time. Unfortunately, current employee scheduling methods are more suited for compressed working arrangements. The problem of automating compressed employee scheduling has been studied by many researchers and is widely adopted by many organizations in an attempt to achieve high quality scheduling. During process of employee scheduling many constraints may have to be considered and may require negotiating a large dimension of constraints like in flexible working. These constraints make scheduling a challenging task in these working arrangements. Most scheduling algorithms are modeled as constraint optimization problems and suited for compressed work but for flexible working with large constraint dimensions, achieving accurate scheduling is even more challenging. In this research, we propose a machine learning approach that takes advantage of mining user-defined constraints or soft constraints and transform employee scheduling into a classification problem. We propose automatically extracting employee personal schedules like calendars in order to extract their availability. We then show how to use the extracted knowledge in a multi-label classification approach in order to generate a schedule for faculty staff in a University that supports flexible working. We show that the results of this approach are comparable to that of a constraint satisfaction and optimization method that is commonly used in literature. Results show that our approach achieved accuracy of 93.1% of satisfying constraints as compared to 92.7% of a common constraint programming approach.
2020
Amin, Anang Hudaya Muhamad; Abdelmajid, Nabih; Kiwanuka, Fred N
Identity-of-Things Model using Composite Identity on Permissioned Blockchain Network Proceedings Article
In: 2020 Seventh International Conference on Software Defined Systems (SDS), pp. 171–176, IEEE 2020.
Abstract | Links | BibTeX | Tags:
@inproceedings{amin2020identity,
title = {Identity-of-Things Model using Composite Identity on Permissioned Blockchain Network},
author = {Anang Hudaya Muhamad Amin and Nabih Abdelmajid and Fred N Kiwanuka},
url = {https://ieeexplore.ieee.org/abstract/document/9143887},
year = {2020},
date = {2020-01-01},
urldate = {2020-01-01},
booktitle = {2020 Seventh International Conference on Software Defined Systems (SDS)},
pages = {171–176},
organization = {IEEE},
abstract = {The growing prevalence of Internet-of-Things (IoT) technology has led to an increase in the development of heterogeneous smart applications. Smart applications may involve a collaborative participation between IoT devices. Participation of IoT devices for specific application requires a tamper-proof identity to be generated and stored, in order to completely represent the device, as well as to eliminate the possibility of identity spoofing and presence of rogue devices in a network. In this paper, we present a composite Identity-of-Things (IDoT) approach on IoT devices with permissioned blockchain implementation for distributed identity management model. Our proposed approach considers both application and device domains in generating the composite identity. In addition, the use of permissioned blockchain for identity storage and verification allows the identity to be immutable. A simulation has been carried out to demonstrate the application of the proposed identity management model.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
The growing prevalence of Internet-of-Things (IoT) technology has led to an increase in the development of heterogeneous smart applications. Smart applications may involve a collaborative participation between IoT devices. Participation of IoT devices for specific application requires a tamper-proof identity to be generated and stored, in order to completely represent the device, as well as to eliminate the possibility of identity spoofing and presence of rogue devices in a network. In this paper, we present a composite Identity-of-Things (IDoT) approach on IoT devices with permissioned blockchain implementation for distributed identity management model. Our proposed approach considers both application and device domains in generating the composite identity. In addition, the use of permissioned blockchain for identity storage and verification allows the identity to be immutable. A simulation has been carried out to demonstrate the application of the proposed identity management model.
2019
Sinayobye, Janvier Omar; Kaawaase, Kyanda Swaib; Kiwanuka, Fred N; Musabe, Richard
Hybrid model of correlation based filter feature selection and machine learning classifiers applied on smart meter data set Proceedings Article
In: 2019 IEEE/ACM Symposium on Software Engineering in Africa (SEiA), pp. 1–10, IEEE 2019.
Abstract | Links | BibTeX | Tags:
@inproceedings{sinayobye2019hybrid,
title = {Hybrid model of correlation based filter feature selection and machine learning classifiers applied on smart meter data set},
author = {Janvier Omar Sinayobye and Kyanda Swaib Kaawaase and Fred N Kiwanuka and Richard Musabe},
url = {https://ieeexplore.ieee.org/abstract/document/8818678},
year = {2019},
date = {2019-01-01},
urldate = {2019-01-01},
booktitle = {2019 IEEE/ACM Symposium on Software Engineering in Africa (SEiA)},
pages = {1–10},
organization = {IEEE},
abstract = {Feature selection is referred to the process of obtaining a subset from an original feature set according to certain feature selection criterion, which selects the relevant features of the dataset. It plays a role in compressing the data processing scale, where the redundant and irrelevant features are removed. Feature selection techniques show that more information is not always good in machine learning applications. Apply different algorithms for the data at hand and with baseline classification performance values we can select a final feature selection algorithm. In this paper, we propose a hybrid classification model, which has correlation based filter feature selection algorithm and Machine learning as classifiers. The objective of this study is to select relevant features and analyze the outperform machine learning algorithms in order to train our model, predict and compare their classification performance. In this method, features are ordered according to their Absolute correlation value with respect to the class attribute. Then top K Features are selected from ordered list of features to form a reduced dataset. This proposed classifier model is applied to our smart meter datasets. To measure the performance of these selected features; seven benchmark classifier are used; Random Forest (RF), Logistic Regression (LR), k-Nearest Neighbor (kNN), Naïve Bayes (NB), Decision Tree (DT), Linear Discriminant Analysis (LDA) and Support Vector Machine (SVM). This paper then analyzes the performance of all classifiers with feature selection in term of accuracy, sensitivity, F-Measure, Specificity, Precision, and MCC. From our experiment, we found that Random Forest classifier performed higher than other used classifiers.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Feature selection is referred to the process of obtaining a subset from an original feature set according to certain feature selection criterion, which selects the relevant features of the dataset. It plays a role in compressing the data processing scale, where the redundant and irrelevant features are removed. Feature selection techniques show that more information is not always good in machine learning applications. Apply different algorithms for the data at hand and with baseline classification performance values we can select a final feature selection algorithm. In this paper, we propose a hybrid classification model, which has correlation based filter feature selection algorithm and Machine learning as classifiers. The objective of this study is to select relevant features and analyze the outperform machine learning algorithms in order to train our model, predict and compare their classification performance. In this method, features are ordered according to their Absolute correlation value with respect to the class attribute. Then top K Features are selected from ordered list of features to form a reduced dataset. This proposed classifier model is applied to our smart meter datasets. To measure the performance of these selected features; seven benchmark classifier are used; Random Forest (RF), Logistic Regression (LR), k-Nearest Neighbor (kNN), Naïve Bayes (NB), Decision Tree (DT), Linear Discriminant Analysis (LDA) and Support Vector Machine (SVM). This paper then analyzes the performance of all classifiers with feature selection in term of accuracy, sensitivity, F-Measure, Specificity, Precision, and MCC. From our experiment, we found that Random Forest classifier performed higher than other used classifiers.
Kiwanuka, Fred N; Alqatawna, Ja'far; Amin, Anang Hudaya Muhamad; Paul, Sujni; Faris, Hossam
Towards Automated Comprehensive Feature Engineering for Spam Detection. Proceedings Article
In: ICISSP, pp. 429–437, 2019.
Abstract | Links | BibTeX | Tags:
@inproceedings{kiwanuka2019towards,
title = {Towards Automated Comprehensive Feature Engineering for Spam Detection.},
author = {Fred N Kiwanuka and Ja'far Alqatawna and Anang Hudaya Muhamad Amin and Sujni Paul and Hossam Faris},
url = {https://www.scitepress.org/Papers/2019/73930/73930.pdf},
year = {2019},
date = {2019-01-01},
urldate = {2019-01-01},
booktitle = {ICISSP},
pages = {429–437},
abstract = {Everyday billions of emails are passed or processed through online servers of which about 59% is spam according to a recent research. Spam emails have increasingly contained viruses or other harmful malware and are a security risk to computer systems. The importance of spam filtering and the security of computer systems has become more essential than ever. The rate of evolution of spam nowadays is so high and hence previously successful spam detection methods are failing to cope. In this paper, we propose a comprehensive and automated feature engineering framework for spam classification. The proposed framework enables first, the development of a large number of features from any email corpus, and second extracting automated features using feature transformation and aggregation primitives. We show that the performance of classification of spam improves between 2% to 28% for almost all conventional machine learning classifiers when using automated feature engineering. As a by product of our comprehensive automated feature engineering, we develop a Python-based open source tool, which incorporates the proposed framework.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Everyday billions of emails are passed or processed through online servers of which about 59% is spam according to a recent research. Spam emails have increasingly contained viruses or other harmful malware and are a security risk to computer systems. The importance of spam filtering and the security of computer systems has become more essential than ever. The rate of evolution of spam nowadays is so high and hence previously successful spam detection methods are failing to cope. In this paper, we propose a comprehensive and automated feature engineering framework for spam classification. The proposed framework enables first, the development of a large number of features from any email corpus, and second extracting automated features using feature transformation and aggregation primitives. We show that the performance of classification of spam improves between 2% to 28% for almost all conventional machine learning classifiers when using automated feature engineering. As a by product of our comprehensive automated feature engineering, we develop a Python-based open source tool, which incorporates the proposed framework.
Amin, Anang Hudaya Muhamad; Alqatawna, Ja'far; Paul, Sujni; Kiwanuka, Fred N; Akhtar, Imtiaz Ahmad
Improving Event Monitoring in IoT Network Using an Integrated Blockchain-Distributed Pattern Recognition Scheme. Proceedings Article
In: BLOCKCHAIN, pp. 134–144, 2019.
Abstract | Links | BibTeX | Tags:
@inproceedings{amin2019improving,
title = {Improving Event Monitoring in IoT Network Using an Integrated Blockchain-Distributed Pattern Recognition Scheme.},
author = {Anang Hudaya Muhamad Amin and Ja'far Alqatawna and Sujni Paul and Fred N Kiwanuka and Imtiaz Ahmad Akhtar},
url = {https://link.springer.com/chapter/10.1007/978-3-030-23813-1_17},
year = {2019},
date = {2019-01-01},
urldate = {2019-01-01},
booktitle = {BLOCKCHAIN},
pages = {134–144},
abstract = {The application of blockchain technology for data storage and verification has been expanding from financial applications to other fields such as asset management and event monitoring in Internet-of-Things (IoT). This expansion consequently intensifies the problem of an increasing size of data stored in the blockchain, especially in event monitoring application where streams of data need to be stored and verified accordingly. In this paper, we propose an IoT-blockchain event monitoring framework that utilizes a distributed pattern recognition scheme for event data processing. Event data are treated as patterns comprising individual data retrieved from interconnected IoT sensors within a network composition. Preliminary results obtained indicate that the proposed scheme is capable of reducing the number of data blocks generated in the blockchain network, hence minimizing the needs for intensive storage and verification.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
The application of blockchain technology for data storage and verification has been expanding from financial applications to other fields such as asset management and event monitoring in Internet-of-Things (IoT). This expansion consequently intensifies the problem of an increasing size of data stored in the blockchain, especially in event monitoring application where streams of data need to be stored and verified accordingly. In this paper, we propose an IoT-blockchain event monitoring framework that utilizes a distributed pattern recognition scheme for event data processing. Event data are treated as patterns comprising individual data retrieved from interconnected IoT sensors within a network composition. Preliminary results obtained indicate that the proposed scheme is capable of reducing the number of data blocks generated in the blockchain network, hence minimizing the needs for intensive storage and verification.
2016
Kiwanuka, Fred N; Wilkinson, Michael HF
Cluster based vector attribute filtering Journal Article
In: Mathematical Morphology-Theory and Applications, vol. 1, no. 1, 2016.
Abstract | Links | BibTeX | Tags:
@article{kiwanuka2016cluster,
title = {Cluster based vector attribute filtering},
author = {Fred N Kiwanuka and Michael HF Wilkinson},
url = {https://www.degruyter.com/document/doi/10.1515/mathm-2016-0007/html},
year = {2016},
date = {2016-01-01},
urldate = {2016-01-01},
journal = {Mathematical Morphology-Theory and Applications},
volume = {1},
number = {1},
publisher = {De Gruyter Open Access},
abstract = {Morphological attribute lters operate on images based on properties or attributes of connectedcomponents. Until recently, attribute ltering was based on a single global threshold on a scalar property toremove or retain objects. A single threshold struggles in case no single property or attribute value has a suit-able, usually multi-modal, distribution. Vector-attribute ltering allows better description of characteristicfeatures for 2D images. In this paper, we apply vector-attribute ltering to 3D and incorporate unsupervisedpattern recognition, where connected components are classi ed based on the similarity of feature vectors.Using a single attribute allows multi-thresholding for attribute lters where more than two classes of struc-tures of interest can be selected. In vector-attribute lters automatic clustering avoids the need for eithersetting very many attribute thresholds, or nding suitable class prototypes in 3D and setting a dissimilaritythreshold. Explorative visualization reduces to visualizing and selecting relevant clusters. We show that theperformance of these new lters is better than those of regular attribute lters in enhancement of objects inmedical images.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Morphological attribute lters operate on images based on properties or attributes of connectedcomponents. Until recently, attribute ltering was based on a single global threshold on a scalar property toremove or retain objects. A single threshold struggles in case no single property or attribute value has a suit-able, usually multi-modal, distribution. Vector-attribute ltering allows better description of characteristicfeatures for 2D images. In this paper, we apply vector-attribute ltering to 3D and incorporate unsupervisedpattern recognition, where connected components are classi ed based on the similarity of feature vectors.Using a single attribute allows multi-thresholding for attribute lters where more than two classes of struc-tures of interest can be selected. In vector-attribute lters automatic clustering avoids the need for eithersetting very many attribute thresholds, or nding suitable class prototypes in 3D and setting a dissimilaritythreshold. Explorative visualization reduces to visualizing and selecting relevant clusters. We show that theperformance of these new lters is better than those of regular attribute lters in enhancement of objects inmedical images.
Kiwanuka, Fred N; Wilkinson, Michael HF
Automatic attribute threshold selection for morphological connected attribute filters Journal Article
In: Pattern recognition, vol. 53, pp. 59–72, 2016.
Abstract | Links | BibTeX | Tags:
@article{kiwanuka2016automatic,
title = {Automatic attribute threshold selection for morphological connected attribute filters},
author = {Fred N Kiwanuka and Michael HF Wilkinson},
url = {https://www.sciencedirect.com/science/article/abs/pii/S003132031500432X},
year = {2016},
date = {2016-01-01},
urldate = {2016-01-01},
journal = {Pattern recognition},
volume = {53},
pages = {59–72},
publisher = {Elsevier},
abstract = {Attribute filters allow enhancement and extraction of features without distorting their borders, and never introduce new image features. In attribute filters, till date setting the attribute-threshold parameters has to be done manually. This research explores novel, simple, fast and automated methods of computing attribute threshold parameters based on image segmentation, thresholding and data clustering techniques in medical image enhancement. A performance analysis of the different methods is carried out using various 3D medical images of different modalities. Though several techniques perform well on these images, the choice of technique appears to depend on the imaging mode.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Attribute filters allow enhancement and extraction of features without distorting their borders, and never introduce new image features. In attribute filters, till date setting the attribute-threshold parameters has to be done manually. This research explores novel, simple, fast and automated methods of computing attribute threshold parameters based on image segmentation, thresholding and data clustering techniques in medical image enhancement. A performance analysis of the different methods is carried out using various 3D medical images of different modalities. Though several techniques perform well on these images, the choice of technique appears to depend on the imaging mode.
2014
Quinn, John A; Andama, Alfred; Munabi, Ian; Kiwanuka, Fred N
Automated blood smear analysis for mobile malaria diagnosis Journal Article
In: Mobile Point-of-Care Monitors and Diagnostic Device Design, vol. 31, pp. 115, 2014.
Abstract | Links | BibTeX | Tags:
@article{quinn2014automated,
title = {Automated blood smear analysis for mobile malaria diagnosis},
author = {John A Quinn and Alfred Andama and Ian Munabi and Fred N Kiwanuka},
url = {https://jquinn.air.ug/files/Quinn_2014_MobilePointOfCare.pdf},
year = {2014},
date = {2014-01-01},
urldate = {2014-01-01},
journal = {Mobile Point-of-Care Monitors and Diagnostic Device Design},
volume = {31},
pages = {115},
publisher = {CRC Press},
abstract = {he gold standard test for malaria is the hundred-year-old method of preparing a blood smear on a glass slide, staining it, and examining it under a microscope to look for the parasite genus Plasmodium. While several rapid diagnostic tests are also currently available, they still have shortcomings compared to microscopic analysis [18]. In the regions worst affected by malaria, reliable diagnoses are often difficult to obtain, and treatment is routinely prescribed based only on symptoms. Accurate diagnosis is clearly important, since false negatives can be fatal and false positives lead to increased drug resistance, unnecessary economic burden, and possibly the failure to treat diseases with similar early symptoms such as meningitis or typhoid.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
he gold standard test for malaria is the hundred-year-old method of preparing a blood smear on a glass slide, staining it, and examining it under a microscope to look for the parasite genus Plasmodium. While several rapid diagnostic tests are also currently available, they still have shortcomings compared to microscopic analysis [18]. In the regions worst affected by malaria, reliable diagnoses are often difficult to obtain, and treatment is routinely prescribed based only on symptoms. Accurate diagnosis is clearly important, since false negatives can be fatal and false positives lead to increased drug resistance, unnecessary economic burden, and possibly the failure to treat diseases with similar early symptoms such as meningitis or typhoid.
2013
Kiwanuka, Fred N.
Exploring morphological attribute filters in medical image enhancement PhD Thesis
2013.
@phdthesis{nokey,
title = {Exploring morphological attribute filters in medical image enhancement},
author = {Fred N. Kiwanuka},
url = {https://drive.google.com/file/d/0B5TWIz-7XSWCSExzTVVPMS1RX0U/view?usp=drive_link&resourcekey=0--Wxcqq-8NrWwjbYjfiLnqA},
year = {2013},
date = {2013-09-11},
keywords = {},
pubstate = {published},
tppubtype = {phdthesis}
}
2012
Kiwanuka, Fred N; Wilkinson, Michael HF
Radial moment invariants for attribute filtering in 3D Proceedings Article
In: Applications of Discrete Geometry and Mathematical Morphology: First International Workshop, WADGMM 2010, Istanbul, Turkey, August 22, 2010, Revised Selected Papers, pp. 68–81, Springer 2012.
Abstract | Links | BibTeX | Tags:
@inproceedings{kiwanuka2012radial,
title = {Radial moment invariants for attribute filtering in 3D},
author = {Fred N Kiwanuka and Michael HF Wilkinson},
url = {https://link.springer.com/chapter/10.1007/978-3-642-32313-3_5},
year = {2012},
date = {2012-01-01},
urldate = {2012-01-01},
booktitle = {Applications of Discrete Geometry and Mathematical Morphology: First International Workshop, WADGMM 2010, Istanbul, Turkey, August 22, 2010, Revised Selected Papers},
pages = {68–81},
organization = {Springer},
abstract = {The edge or shape preservation property of connected attribute filters is a desirable feature for biomedical imaging and makes them a suitable tool for problems in which accurate shape analysis is of importance. However, there are still comparatively few attributes for 3D filtering upon which to select features of interest besides, efficient and fast computation of attributes from volumetric data is still a daunting challenge. In particular, whereas a vast literature on 2D moment invariants exist, far fewer 3D moment invariants are available. In this study we introduce a new, radial-moment based roundness attribute in 3D, and provide a memory-efficient algorithm to compute it, even for very high moment orders. It satisfies similarity transformations of translation, rotation and scaling invariance and be generalised to higher order moments without performance degradation. We show the utility of the new attribute in the isolation of kidney stones and other structures in 3D CT and MRI images.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
The edge or shape preservation property of connected attribute filters is a desirable feature for biomedical imaging and makes them a suitable tool for problems in which accurate shape analysis is of importance. However, there are still comparatively few attributes for 3D filtering upon which to select features of interest besides, efficient and fast computation of attributes from volumetric data is still a daunting challenge. In particular, whereas a vast literature on 2D moment invariants exist, far fewer 3D moment invariants are available. In this study we introduce a new, radial-moment based roundness attribute in 3D, and provide a memory-efficient algorithm to compute it, even for very high moment orders. It satisfies similarity transformations of translation, rotation and scaling invariance and be generalised to higher order moments without performance degradation. We show the utility of the new attribute in the isolation of kidney stones and other structures in 3D CT and MRI images.
Kiwanuka, Fred N; Wilkinson, Michael HF
Cluster-based vector-attribute filtering for ct and mri enhancement Proceedings Article
In: Proceedings of the 21st International Conference on Pattern Recognition (ICPR2012), pp. 3112–3115, IEEE 2012.
Abstract | Links | BibTeX | Tags:
@inproceedings{kiwanuka2012cluster,
title = {Cluster-based vector-attribute filtering for ct and mri enhancement},
author = {Fred N Kiwanuka and Michael HF Wilkinson},
url = {https://ieeexplore.ieee.org/abstract/document/6460823},
year = {2012},
date = {2012-01-01},
urldate = {2012-01-01},
booktitle = {Proceedings of the 21st International Conference on Pattern Recognition (ICPR2012)},
pages = {3112–3115},
organization = {IEEE},
abstract = {Morphological attribute filters modify images based on properties or attributes of connected components. Usually, attribute filtering is based on a scalar property which has relatively little discriminating power. Vector-attribute filtering allow better description of characteristic features for 2D images. In this paper, we extend vector attribute filtering by incorporating unsupervised pattern recognition, where connected components are clustered based on the similarity of feature vectors. We show that the performance of these new filters is better than those of scalar attribute filters in enhancement of objects in medical volumes.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Morphological attribute filters modify images based on properties or attributes of connected components. Usually, attribute filtering is based on a scalar property which has relatively little discriminating power. Vector-attribute filtering allow better description of characteristic features for 2D images. In this paper, we extend vector attribute filtering by incorporating unsupervised pattern recognition, where connected components are clustered based on the similarity of feature vectors. We show that the performance of these new filters is better than those of scalar attribute filters in enhancement of objects in medical volumes.
2011
Tushabe, Florence; Mwebaze, ERNEST; Kiwanuka, F
An image-based diagnosis of virus and bacterial skin infections Proceedings Article
In: The International Conference on Complications in Interventional Radiology, pp. 1–7, 2011.
Abstract | Links | BibTeX | Tags:
@inproceedings{tushabe2011image,
title = {An image-based diagnosis of virus and bacterial skin infections},
author = {Florence Tushabe and ERNEST Mwebaze and F Kiwanuka},
url = {https://www.researchgate.net/profile/Fred-Kiwanuka/publication/268241732_An_image-based_diagnosis_of_virus_and_bacterial_skin_infections/links/5512de910cf20bfdad523b8f/An-image-based-diagnosis-of-virus-and-bacterial-skin-infections.pdf},
year = {2011},
date = {2011-01-01},
urldate = {2011-01-01},
booktitle = {The International Conference on Complications in Interventional Radiology},
pages = {1–7},
abstract = {Skin diseases in sub-saharan Africa tend to be prevalent due to climatic as well as the living situation of the vast majority of people. The situation is compounded by the low numbers of trained medical personnel to diagnose and treat these diseases effectively. A lot of the rural medical personnel use whatever past experience they have to diagnose the general cause of the skin disease and as such give a blanket treatment. In this study we propose an image-based diagnosis method where images of skin disorders are used to classify whether the skin disease falls in the broad category of virus infections or bacterial infections. We show that with a few training images we can get very good performance results with accuracy precision of as much as 100%},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Skin diseases in sub-saharan Africa tend to be prevalent due to climatic as well as the living situation of the vast majority of people. The situation is compounded by the low numbers of trained medical personnel to diagnose and treat these diseases effectively. A lot of the rural medical personnel use whatever past experience they have to diagnose the general cause of the skin disease and as such give a blanket treatment. In this study we propose an image-based diagnosis method where images of skin disorders are used to classify whether the skin disease falls in the broad category of virus infections or bacterial infections. We show that with a few training images we can get very good performance results with accuracy precision of as much as 100%
2010
Kiwanuka, Fred N; Wilkinson, Michael HF
Automatic attribute threshold selection for blood vessel enhancement Proceedings Article
In: 2010 20th International Conference on Pattern Recognition, pp. 2314–2317, IEEE 2010.
Abstract | Links | BibTeX | Tags:
@inproceedings{kiwanuka2010automatic,
title = {Automatic attribute threshold selection for blood vessel enhancement},
author = {Fred N Kiwanuka and Michael HF Wilkinson},
url = {https://ieeexplore.ieee.org/abstract/document/5595741},
year = {2010},
date = {2010-01-01},
urldate = {2010-01-01},
booktitle = {2010 20th International Conference on Pattern Recognition},
pages = {2314–2317},
organization = {IEEE},
abstract = {Attribute filters allow enhancement and extraction of features without distorting their borders, and never introduce new image features. These are highly desirable properties in biomedical imaging, where accurate shape analysis is paramount. However, setting the attribute-threshold parameters has to date only been done manually. This paper explores simple, fast and automated methods of computing attribute threshold parameters based on image segmentation, thresholding and data clustering techniques. Though several techniques perform well on blood-vessel filtering, the choice of technique appears to depend on the imaging mode.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Attribute filters allow enhancement and extraction of features without distorting their borders, and never introduce new image features. These are highly desirable properties in biomedical imaging, where accurate shape analysis is paramount. However, setting the attribute-threshold parameters has to date only been done manually. This paper explores simple, fast and automated methods of computing attribute threshold parameters based on image segmentation, thresholding and data clustering techniques. Though several techniques perform well on blood-vessel filtering, the choice of technique appears to depend on the imaging mode.
2009
Kiwanuka, Fred N; Ouzounis, Georgios K; Wilkinson, Michael HF
Surface-area-based attribute filtering in 3d Proceedings Article
In: Mathematical Morphology and Its Application to Signal and Image Processing: 9th International Symposium, ISMM 2009 Groningen, The Netherlands, August 24-27, 2009 Proceedings 9, pp. 70–81, Springer 2009.
Abstract | Links | BibTeX | Tags:
@inproceedings{kiwanuka2009surface,
title = {Surface-area-based attribute filtering in 3d},
author = {Fred N Kiwanuka and Georgios K Ouzounis and Michael HF Wilkinson},
url = {https://link.springer.com/chapter/10.1007/978-3-642-03613-2_7},
year = {2009},
date = {2009-01-01},
urldate = {2009-01-01},
booktitle = {Mathematical Morphology and Its Application to Signal and Image Processing: 9th International Symposium, ISMM 2009 Groningen, The Netherlands, August 24-27, 2009 Proceedings 9},
pages = {70–81},
organization = {Springer},
abstract = {In this paper we describe a rotation-invariant attribute filter based on estimating the sphericity or roundness of objects by efficiently computing surface area and volume of connected components. The method is based on an efficient algorithm to compute all iso-surfaces of all nodes in a Max-Tree. With similar properties to moment-based attributes like sparseness, non-compactness, and elongation, our sphericity attribute can supplement these in finding blood-vessels in time-of-flight MR angiograms. We compare the method to a discrete surface area method based on adjacency, which has been used for urinary stone detection. Though the latter is faster, it is less accurate, and lacks rotation invariance.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
In this paper we describe a rotation-invariant attribute filter based on estimating the sphericity or roundness of objects by efficiently computing surface area and volume of connected components. The method is based on an efficient algorithm to compute all iso-surfaces of all nodes in a Max-Tree. With similar properties to moment-based attributes like sparseness, non-compactness, and elongation, our sphericity attribute can supplement these in finding blood-vessels in time-of-flight MR angiograms. We compare the method to a discrete surface area method based on adjacency, which has been used for urinary stone detection. Though the latter is faster, it is less accurate, and lacks rotation invariance.
Kiwanuka, Fred N
Multi-scale angiography filters: techniques today Journal Article
In: Strengthening the Role of ICT in Development, vol. 5, pp. 337–346, 2009.
Abstract | Links | BibTeX | Tags:
@article{kiwanuka2009multi,
title = {Multi-scale angiography filters: techniques today},
author = {Fred N Kiwanuka},
url = {https://nru.uncst.go.ug/bitstream/handle/123456789/4978/Web%20Content%20Filtration%20according%20to%2011,%20209.pdf?sequence=2#page=357},
year = {2009},
date = {2009-01-01},
urldate = {2009-01-01},
journal = {Strengthening the Role of ICT in Development},
volume = {5},
pages = {337–346},
abstract = {Vessel Enhancement and extraction in Angiography is still in developing state as many important problems still remain to be solved. The computation procedure of vessels in angiography is a very important due to limited computational resources. Computational procedure based on multi-scale has received considerable attention from scientists. Multiscale approaches perform enhancement based on image resolutions and structure sizes. Many contributions have been made on the problem of vessel enhancement multiscale computing of volume data sets but there has never been a head to head evaluation of these approaches. In this paper the various multi-scale vessel enhancement approaches are put in perspective through a head to head comparison of algorithms of the existing research and a generic framework for linear multiscale is presented.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Vessel Enhancement and extraction in Angiography is still in developing state as many important problems still remain to be solved. The computation procedure of vessels in angiography is a very important due to limited computational resources. Computational procedure based on multi-scale has received considerable attention from scientists. Multiscale approaches perform enhancement based on image resolutions and structure sizes. Many contributions have been made on the problem of vessel enhancement multiscale computing of volume data sets but there has never been a head to head evaluation of these approaches. In this paper the various multi-scale vessel enhancement approaches are put in perspective through a head to head comparison of algorithms of the existing research and a generic framework for linear multiscale is presented.