Skip to main content

Syed M S Islam

Edith Cowan University, School of Science, Faculty Member

Curtin University, Department of Mechanical Engineering, Faculty Member

The University of Western Australia, Dentistry, Faculty Member

Edith Cowan University, School of Computer and Security Science, Faculty Member

Followers

156

Following

13

Co-authors

5

Public Views

I completed PhD with Distinction in Computer Engineering from the University of Western Australia (UWA) in 2011. I was a Research A/Prof at UWA, a Research Fellow at Curtin University and a Lecturer at UWA. Currently, I am working as a Lecturer in Computer Science at Edith Cowan University. I have over 60 research articles and 15 public media releases. My research interests include AI and Computer Vision.
Supervisors: Prof Geoff West, Prof Mohammed Bennamoun, Associate Prof Tele Tan, Prof Mauro Farella, and Prof Emile Rossouw
Address: 270 Joondalup Drive, Joondalup, WA 6027

less

National University of Singapore

Hemantha Kulathilake

Rajarata University of Sri Lanka

Nor Aniza Abdullah

University of Malaya, Malaysia

University of Leeds

Syed M Shamsul Islam

Arnab Bhattacharjee

Abdullah Al Miraj 161-24-539

Johns Hopkins University

InterestsView All (13)

Uploads

Papers by Syed M S Islam

Boosting Deep Transfer Learning For Covid-19 Classification

2021 IEEE International Conference on Image Processing (ICIP)

COVID-19 classification using chest Computed Tomography (CT) has been found pragmatically useful ... more COVID-19 classification using chest Computed Tomography (CT) has been found pragmatically useful by several studies. Due to the lack of annotated samples, these studies recommend transfer learning and explore the choices of pre-trained models and data augmentation. However, it is still unknown if there are better strategies than vanilla transfer learning for more accurate COVID-19 classification with limited CT data. This paper provides an affirmative answer, devising a novel 'model' augmentation technique that allows a considerable performance boost to transfer learning for the task. Our method systematically reduces the distributional shift between the source and target domains and considers augmenting deep learning with complementary representation learning techniques. We establish the efficacy of our method with publicly available datasets and models, along with identifying contrasting observations in the previous studies.

SAM-SoS: A Stochastic Software Architecture Modeling and Verification Approach for Complex System-of-Systems

IEEE Access, 2020

A System-of-Systems (SoS) is a complex, dynamic system whose Constituent Systems (CSs) are not kn... more A System-of-Systems (SoS) is a complex, dynamic system whose Constituent Systems (CSs) are not known precisely at design time, and the environment in which they operate is uncertain. SoS behavior is unpredictable due to underlying architectural characteristics such as autonomy and independence. Although the stochastic composition of CSs is vital to achieving SoS missions, their unknown behaviors and impact on system properties are unavoidable. Moreover, unknown conditions and volatility have significant effects on crucial Quality Attributes (QAs) such as performance, reliability and security. Hence, the structure and behavior of a SoS must be modeled and validated quantitatively to foresee any potential impact on the properties critical for achieving the missions. Current modeling approaches lack the essential syntax and semantics required to model and verify SoS behaviors at design time and cannot offer alternative design choices for better design decisions. Therefore, the majority of existing techniques fail to provide qualitative and quantitative verification of SoS architecture models. Consequently, we have proposed an approach to model and verify Non-Deterministic (ND) SoS in advance by extending the current algebraic notations for the formal models as a hybrid stochastic formalism to specify and reason architectural elements with the required semantics. A formal stochastic model is developed using a hybrid approach for architectural descriptions of SoS with behavioral constraints. Through a model-driven approach, stochastic models are then translated into PRISM using formal verification rules. The effectiveness of the approach has been tested with an end-to-end case study design of an emergency response SoS for dealing with a fire situation. Architectural analysis is conducted on the stochastic model, using various qualitative and quantitative measures for SoS missions. Experimental results reveal critical aspects of SoS architecture model that facilitate better achievement of missions and QAs with improved design, using the proposed approach. INDEX TERMS Stochastic systems, system-of-systems, architecture modeling, quantitative verification, statistical model checking, system properties, formal modeling.

Airway morphology and its influence on OSA severity and surgical intervention: a retrospective study

Australasian Orthodontic Journal

Introduction: The aim was to assess the relationship between airway morphology and surgical inter... more Introduction: The aim was to assess the relationship between airway morphology and surgical intervention in a cohort of patients presenting with increased body mass index (BMI) and a confirmed diagnosis of obstructive sleep apnoea (OSA). A secondary aim was to revisit the relationship between morphology and OSA severity. Methods: A retrospective analysis was conducted of pre-operative maxillofacial 3D-CT scans of thirty-two patients with a confirmed diagnosis of OSA who received treatment from an ear nose and throat specialist (ENT). Lateral cephalograms were imported into Quick Ceph Studio (Quick Ceph Systems Inc, San Diego, CA, USA) after which linear and angular measurements of selected hard and soft tissues were obtained. 3D-CT images were loaded into the software program 3dMDVultus (3dMD) which permitted 3D visualisation of the airway. Measurements were repeated 3 times on the images of six patients after an interval of two weeks to establish the intraclass correlation coeffici...

Fast Semantic-Assisted Outlier Removal for Large-scale Point Cloud Registration

Cornell University - arXiv, Feb 21, 2022

With current trends in sensors (cheaper, more volume of data) and applications (increasing afford... more With current trends in sensors (cheaper, more volume of data) and applications (increasing affordability for new tasks, new ideas in what 3D data could be useful for); there is corresponding increasing interest in the ability to automatically, reliably, and cheaply, register together individual point clouds. The volume of data to handle, and still elusive need to have the registration occur fully reliably and fully automatically, mean there is a need to innovate further. One largely untapped area of innovation is that of exploiting the semantic information of the points in question. Points on a tree should match points on a tree, for example, and not points on car. Moreover, such a natural restriction is clearly human-like-a human would generally quickly eliminate candidate regions for matching based on semantics. Employing semantic information is not only efficient but natural. It is also timely-due to the recent advances in semantic classification capabilities. This paper advances this theme by demonstrating that state of the art registration techniques, in particular ones that rely on "preservation of length under rigid motion" as an underlying matching consistency constraint, can be augmented with semantic information. Semantic identity is of course also preserved under rigid-motion, but also under wider motions present in a scene. We demonstrate that not only the potential obstacle of cost of semantic segmentation, and the potential obstacle of unreliability of semantic segmentation; are both no impediment to achieving both speed and accuracy in fully automatic registration of large scale point clouds. In particular, we take a state of the art approach, that uses an (approximate) maximum clique heuristic (a clique on edges encapsulating matches that preserve length), to yield a version further enhanced with exploitation of semantic information. In essence, cheaply and reliably (for the purposes required) pruning the data that has to be dealt with.

A New Uncertainty-aware Deep Neuroevolution Model for Quantifying Tidal Prediction

2021 IEEE Industry Applications Society Annual Meeting (IAS), 2021

In this work, we propose a deep learning-based prediction interval framework in order to model th... more In this work, we propose a deep learning-based prediction interval framework in order to model the forecasting uncertainties of tidal current datasets. The proposed model develops optimum prediction intervals (PIs) focused on the deep learning-based CNN-LSTM model (CLSTM), and non-parametric approach termed as the lower upper bound estimation (LUBE) model. On the other hand, due to the high complexity raises in designing manually the deep learning architectures, as well as the enhancing the performance of the prediction intervals, we develop a novel deep neuroevolution algorithm based on the two-stage modification of the Gaining-Sharing Knowledge (GSK) optimization algorithm to optimize the architecture of the CLSTM automatically without the procedure of trial and error. We also utilize coverage width criterion (CWC) to establish an excellent correlation appropriately between both the the PI coverage probability (PICP) and PI normalized average width (PINAW). We also indicate the searching efficiency and high accuracy of our proposed framework named as MGSK-CLSTM-LUBE by examining over the practical collected tidal current datasets from the Bay of Fundy, NS, Canada. The performance of the proposed model is examined on the practical tidal current data collected from the Bay of Fundy, NS,Canada

1Evaluation of Assembly Tasks in Augmented

— ieee-THREE-conf A distributed telerobotic system consists of a client station (operator) and a ... more — ieee-THREE-conf A distributed telerobotic system consists of a client station (operator) and a server station (slave arm) interconnected by a computer network. The system is evaluated using (1) peg-in-hole insertion, and (2) assembly of a small water pump. Direct teleoperation is evaluated using following schemes: (1) stereo vision, (2) vision and force feedback, and (3) vision with active compliance. Space indexing and scalability tools are also used. Mapping of operator hand motion and force feedback to a convenient tool point reduces operator mental load and task time due to highly-coordinated motion and ease of understanding of force feedback. Operator is logically manipulating the remote tool both in motion and force and feels the exerted forces as they were exerted in the hand. With active compliance all tasks are done in the least task times and with the least contact force. Stereo vision may alone be used but with large peak forces and extended task time. Force feedback ha...

A New Ensemble Reinforcement Learning Strategy for Solar Irradiance Forecasting using Deep Optimized Convolutional Neural Network Models

2021 International Conference on Smart Energy Systems and Technologies (SEST), 2021

Solar irradiance forecasting is a major priority for the power transmission systems in order to g... more Solar irradiance forecasting is a major priority for the power transmission systems in order to generate and incorporate the performance of massive photovoltaic plants efficiently. As such, prior forecasting techniques that use classical modelling and single deep learning models that undertake feature extraction procedures manually were unable to meet the output demands in specific situations with dynamic variability. Therefore, in this study, we propose an efficient novel hybrid solar irradiance forecasting based on three steps. In step I, we employ a powerful variable input selection strategy named as partial mutual information (PMI) to calculate the linear and non-linear correlations of the original solar irradiance data. In step II, unlike the traditional deep learning models designing their architectures manually, we utilize several deep convolutional neural network (CNN) models optimized by a novel modified whale optimization algorithm in order to compute the forecasting results of the solar irradiance datasets. Finally in step III, we deploy a deep Qlearning reinforcement learning strategy for selecting the best subsets of the combined deep optimized CNN models. Through analysing the forecasting results over two USA solar irradiance stations, it can be inferred that the proposed optimized deep RL-ensemble framework (ODERLEN) outperforms other powerful benchmarked algorithms in different time-step horizons.

A Hybrid Approach to Recognising Activities of Daily Living from Object Use in the Home Environment

Informatics, 2018

Accurate recognition of Activities of Daily Living (ADL) plays an important role in providing ass... more Accurate recognition of Activities of Daily Living (ADL) plays an important role in providing assistance and support to the elderly and cognitively impaired. Current knowledge-driven and ontology-based techniques model object concepts from assumptions and everyday common knowledge of object use for routine activities. Modelling activities from such information can lead to incorrect recognition of particular routine activities resulting in possible failure to detect abnormal activity trends. In cases where such prior knowledge are not available, such techniques become virtually unemployable. A significant step in the recognition of activities is the accurate discovery of the object usage for specific routine activities. This paper presents a hybrid framework for automatic consumption of sensor data and associating object usage to routine activities using Latent Dirichlet Allocation (LDA) topic modelling. This process enables the recognition of simple activities of daily living from object usage and interactions in the home environment. The evaluation of the proposed framework on the Kasteren and Ordonez datasets show that it yields better results compared to existing techniques.

Doppler Radar-Based Non-Contact Health Monitoring for Obstructive Sleep Apnea Diagnosis: A Comprehensive Review

Big Data and Cognitive Computing, 2019

Today’s rapid growth of elderly populations and aging problems coupled with the prevalence of obs... more Today’s rapid growth of elderly populations and aging problems coupled with the prevalence of obstructive sleep apnea (OSA) and other health related issues have affected many aspects of society. This has led to high demands for a more robust healthcare monitoring, diagnosing and treatments facilities. In particular to Sleep Medicine, sleep has a key role to play in both physical and mental health. The quality and duration of sleep have a direct and significant impact on people’s learning, memory, metabolism, weight, safety, mood, cardio-vascular health, diseases, and immune system function. The gold-standard for OSA diagnosis is the overnight sleep monitoring system using polysomnography (PSG). However, despite the quality and reliability of the PSG system, it is not well suited for long-term continuous usage due to limited mobility as well as causing possible irritation, distress, and discomfort to patients during the monitoring process. These limitations have led to stronger deman...

Experimental Evaluation of Feedback Modalities for Five Teleoperation Tasks

IEEE Transactions on Instrumentation and Measurement, 2010

A distributed telerobotic system (DTS) is proposed based on a master arm station which is interco... more A distributed telerobotic system (DTS) is proposed based on a master arm station which is interconnected by a computer network to a slave arm station. DTS is evaluated using a set of teleoperated experiments which are (1) peg-inhole insertion, (2) assembly of a small water pump, (3) operating drawers, (4) pouring of water, and (5) wire-wrapping. Direct teleoperation is evaluated using the following schemes: (1) stereo vision (V), (2) vision and force feedback (VFF), and (3) vision with active compliance (VAC). Space indexing and scaling tools are also used. Operator hand is logically mapped to a remote tool both in position and force. The operator feels the forces exerted on the tool as they were exerted on the hand. Extensive experimental analysis showed that mapping of operator hand motion and force feedback (FF) to a convenient tool point reduces operator mental load and task time due to highly-coordinated motion. Stereo vision may alone be used at the cost of large peak forces and extended task time. VFF has nearly equal task time as compared to VAC but with a noticeable increase in contact forces. For a large majority of cases, the contact-based tasks done using VAC resulted in the least task times and the least contact forces. VAC is superior to VFF which is better than V. In other words, there is an enormous gain in stability if one removes the bilateral force feedback channel in teleoperation and relies on a slave arm active compliance.

High Latency Unmanned Ground Vehicle Teleoperation Enhancement by Presentation of Estimated Future through Video Transformation

Journal of Intelligent & Robotic Systems

Long-distance, high latency teleoperation tasks are difficult, highly stressful for teleoperators... more Long-distance, high latency teleoperation tasks are difficult, highly stressful for teleoperators, and prone to over-corrections, which can lead to loss of control. At higher latencies, or when teleoperating at higher vehicle speed, the situation becomes progressively worse. To explore potential solutions, this research work investigates two 2D visual feedback-based assistive interfaces (sliding-only and sliding-and-zooming windows) that apply simple but effective video transformations to enhance teleoperation. A teleoperation simulator that can replicate teleoperation scenarios affected by high and adjustable latency has been developed to explore the effectiveness of the proposed assistive interfaces. Three image comparison metrics have been used to fine-tune and optimise the proposed interfaces. An operator survey was conducted to evaluate and compare performance with and without the assistance. The survey has shown that a 900ms latency increases task completion time by up to 205%...

(will be inserted by the editor) Efficient Detection and Recognition of 3D Ears

This is pre-copy-editing, author-produced version of an article accepted for publication, followi... more

Resetting the baseline: CT-based COVID-19 diagnosis with Deep Transfer Learning is not as accurate as widely thought

2021 Digital Image Computing: Techniques and Applications (DICTA), 2021

Deep learning is gaining instant popularity in computer aided diagnosis of COVID-19. Due to the h... more Deep learning is gaining instant popularity in computer aided diagnosis of COVID-19. Due to the high sensitivity of Computed Tomography (CT) to this disease, CT-based COVID-19 detection with visual models is currently at the forefront of medical imaging research. Outcomes published in this direction are frequently claiming highly accurate detection under deep transfer learning. This is leading medical technologists to believe that deep transfer learning is the mainstream solution for the problem. However, our critical analysis of the literature reveals an alarming performance disparity between different published results. Hence, we conduct a systematic thorough investigation to analyze the effectiveness of deep transfer learning for COVID-19 detection with CT images. Exploring 14 state-of-the-art visual models with over 200 model training sessions, we conclusively establish that the published literature is frequently overestimating transfer learning performance for the problem, even in the prestigious scientific sources. The roots of overestimation trace back to inappropriate data curation. We also provide case studies that consider more realistic scenarios, and establish transparent baselines for the problem. We hope that our reproducible investigation will help in curbing hype-driven claims for the critical problem of COVID-19 diagnosis, and pave the way for a more transparent performance evaluation of techniques for CT-based COVID-19 detection.

A Fast and Fully Automatic Ear Recognition Approach Based on 3D Local Surface Features

Advanced Concepts for Intelligent Vision Systems, 2008

Sensitivity of global features to pose, illumination and scale variations encouraged researchers ... more Sensitivity of global features to pose, illumination and scale variations encouraged researchers to use local features for object representation and recognition. Availability of 3D scanners also made the use of 3D data (which is less affected by such variations compared to its 2D counterpart) very popular in computer vision applications. In this paper, an approach is proposed for human ear recognition based on robust 3D local features. The features are constructed on distinctive locations in the 3D ear data with an approximated surface around them based on the neighborhood information. Correspondences are then established between gallery and probe features and the two data sets are aligned based on these correspondences. A minimal rectangular subset of the whole 3D ear data only containing the corresponding features is then passed to the Iterative Closest Point (ICP) algorithm for final recognition. Experiments were performed on the UND biometric database and the proposed system achieved 90, 94 and 96 percent recognition rate for rank one, two and three respectively. The approach is fully automatic, comparatively very fast and makes on assumption about the localization of the nose or the ear pit, unlike previous works on ear recognition.

Automatic ear detection using deep learning

Ear based biometric identification can be the solution for instance such as surveillance where ot... more Ear based biometric identification can be the solution for instance such as surveillance where other biometric traits are simply very hard to access. Although many semi-automatic approaches have been made to detect ear and use it for human recognition purpose, most of them are based on feature extraction or shallow machine learning approaches. Very few approaches those used deep neural network architectures are either having less hidden layers, or a combination of deep neural network and feature extraction classifiers or already trained complex deep convolutional network. In this research approach, a deep but simple raw convolutional neural network have been used to detect ear from an ear and non-ear environment which is the initial part of ear based biometric implication. Data-Augmentation have also used and a comparative analysis have been done for original data set and Augmented dataset. Using this deep but from scratch architecture trained by 792 original images we achieved prom...

Exploring Intensity Invariance in Deep Neural Networks for Brain Image Registration

2020 Digital Image Computing: Techniques and Applications (DICTA), 2020

Image registration is a widely-used technique in analysing large scale datasets that are captured... more Image registration is a widely-used technique in analysing large scale datasets that are captured through various imaging modalities and techniques in biomedical imaging such as MRI, X-Rays, etc. These datasets are typically collected from various sites and under different imaging protocols using a variety of scanners. Such heterogeneity in the data collection process causes inhomogeneity or variation in intensity (brightness) and noise distribution. These variations play a detrimental role in the performance of image registration, segmentation and detection algorithms. Classical image registration methods are computationally expensive but are able to handle these artifacts relatively better. However, deep learning-based techniques are shown to be computationally efficient for automated brain registration but are sensitive to the intensity variations. In this study, we investigate the effect of variation in intensity distribution among input image pairs for deep learning-based image...

Imaging and Classification Techniques for Seagrass Mapping and Monitoring: A Comprehensive Survey

Monitoring underwater habitats is a vital part of observing the condition of the environment. The... more Monitoring underwater habitats is a vital part of observing the condition of the environment. The detection and mapping of underwater vegetation, especially seagrass has drawn the attention of the research community as early as the nineteen eighties. Initially, this monitoring relied on in situ observation by experts. Later, advances in remote-sensing technology, satellite-monitoring techniques and, digital photo- and video-based techniques opened a window to quicker, cheaper, and, potentially, more accurate seagrass-monitoring methods. So far, for seagrass detection and mapping, digital images from airborne cameras, spectral images from satellites, acoustic image data using underwater sonar technology, and digital underwater photo and video images have been used to map the seagrass meadows or monitor their condition. In this article, we have reviewed the recent approaches to seagrass detection and mapping to understand the gaps of the present approaches and determine further resear...

Discovering trends for the development of novel authentication applications for dementia patients

We aim at creating ease in authentication process through non-password-based authentication schem... more We aim at creating ease in authentication process through non-password-based authentication scheme for the Dementia patients. The chronic neuro-degenerative disease leaves the patients with memory recall/loss issues. With ever growing rich list of assistive technologies, that bring ease in patient’s daily life i.e. remote Electrocardiography and peripheral capillary oxygen saturation monitoring, remote blood glucose level monitoring applications etc. These assistive technologies are ubiquitous, seamless, immersed in the background, often remotely monitored, and the most intimate applications that run very close to the patient’s physiology. In this paper, we investigate the existing technologies and discover the trends to build Yet Another Authentication Method (YAAM). The YAAM is going to extract a distinctive image from a patient’s viewfinder and securely transform it into authentication token that are supported by the Geo-location, relative proximity of surrounding smart objects e...

The k-means Algorithm: A Comprehensive Survey and Performance Evaluation

Electronics, 2020

The k-means clustering algorithm is considered one of the most powerful and popular data mining a... more The k-means clustering algorithm is considered one of the most powerful and popular data mining algorithms in the research community. However, despite its popularity, the algorithm has certain limitations, including problems associated with random initialization of the centroids which leads to unexpected convergence. Additionally, such a clustering algorithm requires the number of clusters to be defined beforehand, which is responsible for different cluster shapes and outlier effects. A fundamental problem of the k-means algorithm is its inability to handle various data types. This paper provides a structured and synoptic overview of research conducted on the k-means algorithm to overcome such shortcomings. Variants of the k-means algorithms including their recent developments are discussed, where their effectiveness is investigated based on the experimental analysis of a variety of datasets. The detailed experimental analysis along with a thorough comparison among different k-means...

Deep Learning on Underwater Marine Object Detection: A Survey

Advanced Concepts for Intelligent Vision Systems, 2017

Deep learning, also known as deep machine learning or deep structured learning based techniques, ... more Deep learning, also known as deep machine learning or deep structured learning based techniques, have recently achieved tremendous success in digital image processing for object detection and classification. As a result, they are rapidly gaining popularity and attention from the computer vision research community. There has been a massive increase in the collection of digital imagery for the monitoring of underwater ecosystems, including seagrass meadows. This growth in image data has driven the need for automatic detection and classification using deep neural network based classifiers. This paper systematically describes the use of deep learning for underwater imagery analysis within the recent past. The analysis approaches are categorized according to the object of detection, and the features and deep learning architectures used are highlighted. It is concluded that there is a great scope for automation in the analysis of digital seabed imagery using deep neural networks, especially for the detection and monitoring of seagrass.

Boosting Deep Transfer Learning For Covid-19 Classification

2021 IEEE International Conference on Image Processing (ICIP)

COVID-19 classification using chest Computed Tomography (CT) has been found pragmatically useful ... more COVID-19 classification using chest Computed Tomography (CT) has been found pragmatically useful by several studies. Due to the lack of annotated samples, these studies recommend transfer learning and explore the choices of pre-trained models and data augmentation. However, it is still unknown if there are better strategies than vanilla transfer learning for more accurate COVID-19 classification with limited CT data. This paper provides an affirmative answer, devising a novel 'model' augmentation technique that allows a considerable performance boost to transfer learning for the task. Our method systematically reduces the distributional shift between the source and target domains and considers augmenting deep learning with complementary representation learning techniques. We establish the efficacy of our method with publicly available datasets and models, along with identifying contrasting observations in the previous studies.

SAM-SoS: A Stochastic Software Architecture Modeling and Verification Approach for Complex System-of-Systems

IEEE Access, 2020

A System-of-Systems (SoS) is a complex, dynamic system whose Constituent Systems (CSs) are not kn... more A System-of-Systems (SoS) is a complex, dynamic system whose Constituent Systems (CSs) are not known precisely at design time, and the environment in which they operate is uncertain. SoS behavior is unpredictable due to underlying architectural characteristics such as autonomy and independence. Although the stochastic composition of CSs is vital to achieving SoS missions, their unknown behaviors and impact on system properties are unavoidable. Moreover, unknown conditions and volatility have significant effects on crucial Quality Attributes (QAs) such as performance, reliability and security. Hence, the structure and behavior of a SoS must be modeled and validated quantitatively to foresee any potential impact on the properties critical for achieving the missions. Current modeling approaches lack the essential syntax and semantics required to model and verify SoS behaviors at design time and cannot offer alternative design choices for better design decisions. Therefore, the majority of existing techniques fail to provide qualitative and quantitative verification of SoS architecture models. Consequently, we have proposed an approach to model and verify Non-Deterministic (ND) SoS in advance by extending the current algebraic notations for the formal models as a hybrid stochastic formalism to specify and reason architectural elements with the required semantics. A formal stochastic model is developed using a hybrid approach for architectural descriptions of SoS with behavioral constraints. Through a model-driven approach, stochastic models are then translated into PRISM using formal verification rules. The effectiveness of the approach has been tested with an end-to-end case study design of an emergency response SoS for dealing with a fire situation. Architectural analysis is conducted on the stochastic model, using various qualitative and quantitative measures for SoS missions. Experimental results reveal critical aspects of SoS architecture model that facilitate better achievement of missions and QAs with improved design, using the proposed approach. INDEX TERMS Stochastic systems, system-of-systems, architecture modeling, quantitative verification, statistical model checking, system properties, formal modeling.

Airway morphology and its influence on OSA severity and surgical intervention: a retrospective study

Australasian Orthodontic Journal

Introduction: The aim was to assess the relationship between airway morphology and surgical inter... more Introduction: The aim was to assess the relationship between airway morphology and surgical intervention in a cohort of patients presenting with increased body mass index (BMI) and a confirmed diagnosis of obstructive sleep apnoea (OSA). A secondary aim was to revisit the relationship between morphology and OSA severity. Methods: A retrospective analysis was conducted of pre-operative maxillofacial 3D-CT scans of thirty-two patients with a confirmed diagnosis of OSA who received treatment from an ear nose and throat specialist (ENT). Lateral cephalograms were imported into Quick Ceph Studio (Quick Ceph Systems Inc, San Diego, CA, USA) after which linear and angular measurements of selected hard and soft tissues were obtained. 3D-CT images were loaded into the software program 3dMDVultus (3dMD) which permitted 3D visualisation of the airway. Measurements were repeated 3 times on the images of six patients after an interval of two weeks to establish the intraclass correlation coeffici...

Fast Semantic-Assisted Outlier Removal for Large-scale Point Cloud Registration

Cornell University - arXiv, Feb 21, 2022

With current trends in sensors (cheaper, more volume of data) and applications (increasing afford... more With current trends in sensors (cheaper, more volume of data) and applications (increasing affordability for new tasks, new ideas in what 3D data could be useful for); there is corresponding increasing interest in the ability to automatically, reliably, and cheaply, register together individual point clouds. The volume of data to handle, and still elusive need to have the registration occur fully reliably and fully automatically, mean there is a need to innovate further. One largely untapped area of innovation is that of exploiting the semantic information of the points in question. Points on a tree should match points on a tree, for example, and not points on car. Moreover, such a natural restriction is clearly human-like-a human would generally quickly eliminate candidate regions for matching based on semantics. Employing semantic information is not only efficient but natural. It is also timely-due to the recent advances in semantic classification capabilities. This paper advances this theme by demonstrating that state of the art registration techniques, in particular ones that rely on "preservation of length under rigid motion" as an underlying matching consistency constraint, can be augmented with semantic information. Semantic identity is of course also preserved under rigid-motion, but also under wider motions present in a scene. We demonstrate that not only the potential obstacle of cost of semantic segmentation, and the potential obstacle of unreliability of semantic segmentation; are both no impediment to achieving both speed and accuracy in fully automatic registration of large scale point clouds. In particular, we take a state of the art approach, that uses an (approximate) maximum clique heuristic (a clique on edges encapsulating matches that preserve length), to yield a version further enhanced with exploitation of semantic information. In essence, cheaply and reliably (for the purposes required) pruning the data that has to be dealt with.

A New Uncertainty-aware Deep Neuroevolution Model for Quantifying Tidal Prediction

2021 IEEE Industry Applications Society Annual Meeting (IAS), 2021

In this work, we propose a deep learning-based prediction interval framework in order to model th... more In this work, we propose a deep learning-based prediction interval framework in order to model the forecasting uncertainties of tidal current datasets. The proposed model develops optimum prediction intervals (PIs) focused on the deep learning-based CNN-LSTM model (CLSTM), and non-parametric approach termed as the lower upper bound estimation (LUBE) model. On the other hand, due to the high complexity raises in designing manually the deep learning architectures, as well as the enhancing the performance of the prediction intervals, we develop a novel deep neuroevolution algorithm based on the two-stage modification of the Gaining-Sharing Knowledge (GSK) optimization algorithm to optimize the architecture of the CLSTM automatically without the procedure of trial and error. We also utilize coverage width criterion (CWC) to establish an excellent correlation appropriately between both the the PI coverage probability (PICP) and PI normalized average width (PINAW). We also indicate the searching efficiency and high accuracy of our proposed framework named as MGSK-CLSTM-LUBE by examining over the practical collected tidal current datasets from the Bay of Fundy, NS, Canada. The performance of the proposed model is examined on the practical tidal current data collected from the Bay of Fundy, NS,Canada

1Evaluation of Assembly Tasks in Augmented

— ieee-THREE-conf A distributed telerobotic system consists of a client station (operator) and a ... more — ieee-THREE-conf A distributed telerobotic system consists of a client station (operator) and a server station (slave arm) interconnected by a computer network. The system is evaluated using (1) peg-in-hole insertion, and (2) assembly of a small water pump. Direct teleoperation is evaluated using following schemes: (1) stereo vision, (2) vision and force feedback, and (3) vision with active compliance. Space indexing and scalability tools are also used. Mapping of operator hand motion and force feedback to a convenient tool point reduces operator mental load and task time due to highly-coordinated motion and ease of understanding of force feedback. Operator is logically manipulating the remote tool both in motion and force and feels the exerted forces as they were exerted in the hand. With active compliance all tasks are done in the least task times and with the least contact force. Stereo vision may alone be used but with large peak forces and extended task time. Force feedback ha...

A New Ensemble Reinforcement Learning Strategy for Solar Irradiance Forecasting using Deep Optimized Convolutional Neural Network Models

2021 International Conference on Smart Energy Systems and Technologies (SEST), 2021

Solar irradiance forecasting is a major priority for the power transmission systems in order to g... more Solar irradiance forecasting is a major priority for the power transmission systems in order to generate and incorporate the performance of massive photovoltaic plants efficiently. As such, prior forecasting techniques that use classical modelling and single deep learning models that undertake feature extraction procedures manually were unable to meet the output demands in specific situations with dynamic variability. Therefore, in this study, we propose an efficient novel hybrid solar irradiance forecasting based on three steps. In step I, we employ a powerful variable input selection strategy named as partial mutual information (PMI) to calculate the linear and non-linear correlations of the original solar irradiance data. In step II, unlike the traditional deep learning models designing their architectures manually, we utilize several deep convolutional neural network (CNN) models optimized by a novel modified whale optimization algorithm in order to compute the forecasting results of the solar irradiance datasets. Finally in step III, we deploy a deep Qlearning reinforcement learning strategy for selecting the best subsets of the combined deep optimized CNN models. Through analysing the forecasting results over two USA solar irradiance stations, it can be inferred that the proposed optimized deep RL-ensemble framework (ODERLEN) outperforms other powerful benchmarked algorithms in different time-step horizons.

A Hybrid Approach to Recognising Activities of Daily Living from Object Use in the Home Environment

Informatics, 2018

Accurate recognition of Activities of Daily Living (ADL) plays an important role in providing ass... more Accurate recognition of Activities of Daily Living (ADL) plays an important role in providing assistance and support to the elderly and cognitively impaired. Current knowledge-driven and ontology-based techniques model object concepts from assumptions and everyday common knowledge of object use for routine activities. Modelling activities from such information can lead to incorrect recognition of particular routine activities resulting in possible failure to detect abnormal activity trends. In cases where such prior knowledge are not available, such techniques become virtually unemployable. A significant step in the recognition of activities is the accurate discovery of the object usage for specific routine activities. This paper presents a hybrid framework for automatic consumption of sensor data and associating object usage to routine activities using Latent Dirichlet Allocation (LDA) topic modelling. This process enables the recognition of simple activities of daily living from object usage and interactions in the home environment. The evaluation of the proposed framework on the Kasteren and Ordonez datasets show that it yields better results compared to existing techniques.

Doppler Radar-Based Non-Contact Health Monitoring for Obstructive Sleep Apnea Diagnosis: A Comprehensive Review

Big Data and Cognitive Computing, 2019

Today’s rapid growth of elderly populations and aging problems coupled with the prevalence of obs... more Today’s rapid growth of elderly populations and aging problems coupled with the prevalence of obstructive sleep apnea (OSA) and other health related issues have affected many aspects of society. This has led to high demands for a more robust healthcare monitoring, diagnosing and treatments facilities. In particular to Sleep Medicine, sleep has a key role to play in both physical and mental health. The quality and duration of sleep have a direct and significant impact on people’s learning, memory, metabolism, weight, safety, mood, cardio-vascular health, diseases, and immune system function. The gold-standard for OSA diagnosis is the overnight sleep monitoring system using polysomnography (PSG). However, despite the quality and reliability of the PSG system, it is not well suited for long-term continuous usage due to limited mobility as well as causing possible irritation, distress, and discomfort to patients during the monitoring process. These limitations have led to stronger deman...

Experimental Evaluation of Feedback Modalities for Five Teleoperation Tasks

IEEE Transactions on Instrumentation and Measurement, 2010

A distributed telerobotic system (DTS) is proposed based on a master arm station which is interco... more A distributed telerobotic system (DTS) is proposed based on a master arm station which is interconnected by a computer network to a slave arm station. DTS is evaluated using a set of teleoperated experiments which are (1) peg-inhole insertion, (2) assembly of a small water pump, (3) operating drawers, (4) pouring of water, and (5) wire-wrapping. Direct teleoperation is evaluated using the following schemes: (1) stereo vision (V), (2) vision and force feedback (VFF), and (3) vision with active compliance (VAC). Space indexing and scaling tools are also used. Operator hand is logically mapped to a remote tool both in position and force. The operator feels the forces exerted on the tool as they were exerted on the hand. Extensive experimental analysis showed that mapping of operator hand motion and force feedback (FF) to a convenient tool point reduces operator mental load and task time due to highly-coordinated motion. Stereo vision may alone be used at the cost of large peak forces and extended task time. VFF has nearly equal task time as compared to VAC but with a noticeable increase in contact forces. For a large majority of cases, the contact-based tasks done using VAC resulted in the least task times and the least contact forces. VAC is superior to VFF which is better than V. In other words, there is an enormous gain in stability if one removes the bilateral force feedback channel in teleoperation and relies on a slave arm active compliance.

High Latency Unmanned Ground Vehicle Teleoperation Enhancement by Presentation of Estimated Future through Video Transformation

Journal of Intelligent & Robotic Systems

Long-distance, high latency teleoperation tasks are difficult, highly stressful for teleoperators... more Long-distance, high latency teleoperation tasks are difficult, highly stressful for teleoperators, and prone to over-corrections, which can lead to loss of control. At higher latencies, or when teleoperating at higher vehicle speed, the situation becomes progressively worse. To explore potential solutions, this research work investigates two 2D visual feedback-based assistive interfaces (sliding-only and sliding-and-zooming windows) that apply simple but effective video transformations to enhance teleoperation. A teleoperation simulator that can replicate teleoperation scenarios affected by high and adjustable latency has been developed to explore the effectiveness of the proposed assistive interfaces. Three image comparison metrics have been used to fine-tune and optimise the proposed interfaces. An operator survey was conducted to evaluate and compare performance with and without the assistance. The survey has shown that a 900ms latency increases task completion time by up to 205%...

(will be inserted by the editor) Efficient Detection and Recognition of 3D Ears

This is pre-copy-editing, author-produced version of an article accepted for publication, followi... more

Resetting the baseline: CT-based COVID-19 diagnosis with Deep Transfer Learning is not as accurate as widely thought

2021 Digital Image Computing: Techniques and Applications (DICTA), 2021

Deep learning is gaining instant popularity in computer aided diagnosis of COVID-19. Due to the h... more Deep learning is gaining instant popularity in computer aided diagnosis of COVID-19. Due to the high sensitivity of Computed Tomography (CT) to this disease, CT-based COVID-19 detection with visual models is currently at the forefront of medical imaging research. Outcomes published in this direction are frequently claiming highly accurate detection under deep transfer learning. This is leading medical technologists to believe that deep transfer learning is the mainstream solution for the problem. However, our critical analysis of the literature reveals an alarming performance disparity between different published results. Hence, we conduct a systematic thorough investigation to analyze the effectiveness of deep transfer learning for COVID-19 detection with CT images. Exploring 14 state-of-the-art visual models with over 200 model training sessions, we conclusively establish that the published literature is frequently overestimating transfer learning performance for the problem, even in the prestigious scientific sources. The roots of overestimation trace back to inappropriate data curation. We also provide case studies that consider more realistic scenarios, and establish transparent baselines for the problem. We hope that our reproducible investigation will help in curbing hype-driven claims for the critical problem of COVID-19 diagnosis, and pave the way for a more transparent performance evaluation of techniques for CT-based COVID-19 detection.

A Fast and Fully Automatic Ear Recognition Approach Based on 3D Local Surface Features

Advanced Concepts for Intelligent Vision Systems, 2008

Sensitivity of global features to pose, illumination and scale variations encouraged researchers ... more Sensitivity of global features to pose, illumination and scale variations encouraged researchers to use local features for object representation and recognition. Availability of 3D scanners also made the use of 3D data (which is less affected by such variations compared to its 2D counterpart) very popular in computer vision applications. In this paper, an approach is proposed for human ear recognition based on robust 3D local features. The features are constructed on distinctive locations in the 3D ear data with an approximated surface around them based on the neighborhood information. Correspondences are then established between gallery and probe features and the two data sets are aligned based on these correspondences. A minimal rectangular subset of the whole 3D ear data only containing the corresponding features is then passed to the Iterative Closest Point (ICP) algorithm for final recognition. Experiments were performed on the UND biometric database and the proposed system achieved 90, 94 and 96 percent recognition rate for rank one, two and three respectively. The approach is fully automatic, comparatively very fast and makes on assumption about the localization of the nose or the ear pit, unlike previous works on ear recognition.

Automatic ear detection using deep learning

Ear based biometric identification can be the solution for instance such as surveillance where ot... more Ear based biometric identification can be the solution for instance such as surveillance where other biometric traits are simply very hard to access. Although many semi-automatic approaches have been made to detect ear and use it for human recognition purpose, most of them are based on feature extraction or shallow machine learning approaches. Very few approaches those used deep neural network architectures are either having less hidden layers, or a combination of deep neural network and feature extraction classifiers or already trained complex deep convolutional network. In this research approach, a deep but simple raw convolutional neural network have been used to detect ear from an ear and non-ear environment which is the initial part of ear based biometric implication. Data-Augmentation have also used and a comparative analysis have been done for original data set and Augmented dataset. Using this deep but from scratch architecture trained by 792 original images we achieved prom...

Exploring Intensity Invariance in Deep Neural Networks for Brain Image Registration

2020 Digital Image Computing: Techniques and Applications (DICTA), 2020

Image registration is a widely-used technique in analysing large scale datasets that are captured... more Image registration is a widely-used technique in analysing large scale datasets that are captured through various imaging modalities and techniques in biomedical imaging such as MRI, X-Rays, etc. These datasets are typically collected from various sites and under different imaging protocols using a variety of scanners. Such heterogeneity in the data collection process causes inhomogeneity or variation in intensity (brightness) and noise distribution. These variations play a detrimental role in the performance of image registration, segmentation and detection algorithms. Classical image registration methods are computationally expensive but are able to handle these artifacts relatively better. However, deep learning-based techniques are shown to be computationally efficient for automated brain registration but are sensitive to the intensity variations. In this study, we investigate the effect of variation in intensity distribution among input image pairs for deep learning-based image...

Imaging and Classification Techniques for Seagrass Mapping and Monitoring: A Comprehensive Survey

Monitoring underwater habitats is a vital part of observing the condition of the environment. The... more Monitoring underwater habitats is a vital part of observing the condition of the environment. The detection and mapping of underwater vegetation, especially seagrass has drawn the attention of the research community as early as the nineteen eighties. Initially, this monitoring relied on in situ observation by experts. Later, advances in remote-sensing technology, satellite-monitoring techniques and, digital photo- and video-based techniques opened a window to quicker, cheaper, and, potentially, more accurate seagrass-monitoring methods. So far, for seagrass detection and mapping, digital images from airborne cameras, spectral images from satellites, acoustic image data using underwater sonar technology, and digital underwater photo and video images have been used to map the seagrass meadows or monitor their condition. In this article, we have reviewed the recent approaches to seagrass detection and mapping to understand the gaps of the present approaches and determine further resear...

Discovering trends for the development of novel authentication applications for dementia patients

We aim at creating ease in authentication process through non-password-based authentication schem... more We aim at creating ease in authentication process through non-password-based authentication scheme for the Dementia patients. The chronic neuro-degenerative disease leaves the patients with memory recall/loss issues. With ever growing rich list of assistive technologies, that bring ease in patient’s daily life i.e. remote Electrocardiography and peripheral capillary oxygen saturation monitoring, remote blood glucose level monitoring applications etc. These assistive technologies are ubiquitous, seamless, immersed in the background, often remotely monitored, and the most intimate applications that run very close to the patient’s physiology. In this paper, we investigate the existing technologies and discover the trends to build Yet Another Authentication Method (YAAM). The YAAM is going to extract a distinctive image from a patient’s viewfinder and securely transform it into authentication token that are supported by the Geo-location, relative proximity of surrounding smart objects e...

The k-means Algorithm: A Comprehensive Survey and Performance Evaluation

Electronics, 2020

The k-means clustering algorithm is considered one of the most powerful and popular data mining a... more The k-means clustering algorithm is considered one of the most powerful and popular data mining algorithms in the research community. However, despite its popularity, the algorithm has certain limitations, including problems associated with random initialization of the centroids which leads to unexpected convergence. Additionally, such a clustering algorithm requires the number of clusters to be defined beforehand, which is responsible for different cluster shapes and outlier effects. A fundamental problem of the k-means algorithm is its inability to handle various data types. This paper provides a structured and synoptic overview of research conducted on the k-means algorithm to overcome such shortcomings. Variants of the k-means algorithms including their recent developments are discussed, where their effectiveness is investigated based on the experimental analysis of a variety of datasets. The detailed experimental analysis along with a thorough comparison among different k-means...

Deep Learning on Underwater Marine Object Detection: A Survey

Advanced Concepts for Intelligent Vision Systems, 2017

Deep learning, also known as deep machine learning or deep structured learning based techniques, ... more Deep learning, also known as deep machine learning or deep structured learning based techniques, have recently achieved tremendous success in digital image processing for object detection and classification. As a result, they are rapidly gaining popularity and attention from the computer vision research community. There has been a massive increase in the collection of digital imagery for the monitoring of underwater ecosystems, including seagrass meadows. This growth in image data has driven the need for automatic detection and classification using deep neural network based classifiers. This paper systematically describes the use of deep learning for underwater imagery analysis within the recent past. The analysis approaches are categorized according to the object of detection, and the features and deep learning architectures used are highlighted. It is concluded that there is a great scope for automation in the analysis of digital seabed imagery using deep neural networks, especially for the detection and monitoring of seagrass.

Human recognition using local 3D ear and face features

The field of Biometrics is rapidly gaining popularity due to increasing breaches of traditional s... more The field of Biometrics is rapidly gaining popularity due to increasing breaches of traditional security systems and the decreasing costs of sensors. Among the biometric traits, the ear and the face are considered to be the most socially accepted due to their easy and non-intrusive data acquisition. Furthermore, their feature richness and physical proximity make them good candidates for fusion. However, occlusions due to the presence of hair and ornaments and deformations due to facial expressions pose great challenges for real-life applications of these two biometrics. These challenges are addressed in this dissertation through the development of efficient and robust algorithms for ear detection, ear data representation and finally, the combination of ear and face biometrics using robust fusion techniques. The dissertation is organized as a set of papers already published and/or submitted to journals or internationally refereed conferences.

Three Dimensional Modelling for the Planning of OSA Treatment using Maxillomandibular Advancement Surgery