A5048591818- COVID-19 epidemiological studies
- SARS-CoV-2 and COVID-19 Research
- SARS-CoV-2 detection and testing
- Video Surveillance and Tracking Methods
- Advanced Neural Network Applications
- COVID-19 diagnosis using AI
- Infrared Target Detection Methodologies
- Virology and Viral Diseases
- Advanced X-ray and CT Imaging
- Rabies epidemiology and control
- COVID-19 and Mental Health
- Marine and fisheries research
- Domain Adaptation and Few-Shot Learning
- Robotics and Sensor-Based Localization
- Speech Recognition and Synthesis
- Fish Ecology and Management Studies
- Microbial infections and disease research
- Biosensors and Analytical Detection
- Medical Imaging Techniques and Applications
- Vaccine Coverage and Hesitancy
- Music and Audio Processing
- COVID-19 Pandemic Impacts
- Avian ecology and behavior
- Advanced Vision and Imaging
- AI-based Problem Solving and Planning
University of Glasgow
2021-2025
Glasgow Centre for Population Health
2024
World Health Organization
2018
Durham University
2014-2018
We consider the use of deep convolutional neural networks (CNNs) with transfer learning for image classification and detection problems posed within context X-ray baggage security imagery. The CNN approach requires large amounts data to facilitate a complex end-to-end feature extraction process. Within screening, limited availability object interest examples can thus pose problem. To overcome this issue, we employ paradigm such that pre-trained CNN, primarily trained generalized tasks where...
We consider the use of transfer learning, via deep Convolutional Neural Networks (CNN) for image classification problem posed within context X-ray baggage security screening. The a multi-layer CNN approach, traditionally requires large amounts training data, in order to facilitate construction complex complete end-to-end feature extraction, representation and process. Within screening, limited availability particular items interest can thus pose problem. To overcome this issue, we employ...
Here we explore the use of various feature point descriptors as visual word variants within a Bag-of-Visual-Words (BoVW) representation scheme for image classification based threat detection baggage security X-ray imagery.Using classical BoVW model with range detectors and descriptors, supported by both Support Vector Machine (SVM) Random Forest classification, illustrate current performance capability approaches following this paradigm over large imagery data set.An optimal statistical...
Abstract For migratory species, successful navigation is critical to fitness. In Atlantic salmon, for example, there evidence that during migration from natal streams the sea, passage through waters with poorly defined or mixed water velocity patterns may constrain directional navigation, causing individuals become trapped delayed in lakes other bodies slowly flowing water. this study, we determined minimum velocities needed elicit a behavioural response, case change direction of holding...
Abstract Diagnostics for COVID-19 detection are limited in many settings. Syndromic surveillance is often the only means to identify cases but lacks specificity. Rapid antigen testing inexpensive and easy-to-deploy can lack sensitivity. We examine how combining these approaches improve guiding interventions low-income communities Dhaka, Bangladesh. Rapid-antigen-testing with PCR validation was performed on 1172 symptomatically-identified individuals their homes. Statistical models were...
Background: Despite national elimination efforts, dog-mediated rabies remains endemic in the Philippines. Free provision of post-exposure prophylaxis (PEP) through widespread establishment Animal Bite Treatment Centers (ABTCs) has improved accessibility; however, resulting upsurge PEP demand is not sustainable, and human deaths continue. Dog vaccination coverage also inadequate, it unclear whether surveillance effective. Methods: Here, we used Integrated Case Management (IBCM) to collect...
Target tracking within conventional video imagery poses a significant challenge that is increasingly being addressed via complex algorithmic solutions. The complexity of this problem can be fundamentally attributed to the ambiguity associated with actual 3D scene position given tracked object in relation its observed 2D image space. We propose an approach challenges current trend solutions by addressing fundamental head-on. In contrast prior work field, we leverage key advantages...
Real-time classification of vehicles into sub-category types poses a significant challenge within infra-red imagery due to the high levels intra-class variation in thermal vehicle signatures caused by aspects design, current operating duration and ambient conditions. Despite these challenges, sensing offers generalized target object detection advantages terms all-weather operation invariance visual camouflage techniques. This work investigates accuracy number real-time approaches for this...
Non-pharmaceutical interventions (NPIs) used to limit SARS-CoV-2 transmission vary in their feasibility, appropriateness and effectiveness different contexts. In Bangladesh a national lockdown implemented March 2020 exacerbated poverty was untenable long-term. A resurgence 2021 warranted renewed NPIs. We sought identify NPIs that were feasible this context explore potential synergies between interventions.
Abstract Reverberation is a common problem for many speech technologies, such as automatic recognition (ASR) systems. This paper investigates the novel combination of precedence, binaural and statistical independence cues enhancing reverberant speech, prior to ASR, under these adverse acoustical conditions when two microphone signals are available. Results enhancement evaluated in terms relevant signal measures accuracy both English Polish ASR tasks. These show inconsistencies between...
Objective To evaluate the diagnostic performance and feasibility of rapid antigen testing for SARS-CoV-2 detection in low-income communities. Design We conducted a cross-sectional community-based accuracy study. Community health workers, who were trained supervised by medical technicians, performed tests on symptomatic individuals, up to two additional household members their households results calibrated against gold standard RT-PCR. Setting Low-income communities Dhaka, Bangladesh....
Abstract A dramatic resurgence of COVID-19 cases and deaths in Bangladesh March 2021 coincided with the SARS-CoV-2 B.1.351 (501Y.V2) variant concern rapidly becoming dominant circulating variant. Concurrently, increasing numbers reinfections have been detected effective Reproductive number, R t , has doubled, despite high levels prior infection Dhaka city. These data support prediction that acquired immunity from past provides reduced protection against B.1.351, highlights major public...
Currently, most land Intelligence, Surveillance and Reconnaissance (ISR) assets (e.g. EO/IR cameras) are simply data collectors. Understanding, decision making sensor control performed by the human operators, involving high cognitive load. Any automation in system has traditionally involved bespoke design of centralised systems that highly specific for assets/targets/environment under consideration, resulting complex, non-flexible exhibit poor interoperability. We address a concept...
Rabies is a viral zoonosis that kills thousands of people annually in low- and middle-income countries across Africa Asia where domestic dogs are the reservoir. 'Zero by 30', global strategy to end dog-mediated human rabies, promotes One Health approach underpinned mass dog vaccination, post-exposure vaccination bite victims, robust surveillance community engagement. Using Integrated Bite Case Management (IBCM) whole genome sequencing (WGS), we enhanced rabies detect an outbreak formerly...
Target tracking complexity within conventional video imagery can be fundamentally attributed to the ambiguity associated with actual 3D scene position of a given tracked object in relation its observed 2D image space. Recent work, thermal-band infrared imagery, has tackled this challenge head on by returning classical photogrammetry as means recovering true pedestrian targets. A key limitation such approaches is assumption posture – that at full height stance scene. Whilst prior work shown...
Abstract Background Non-pharmaceutical interventions (NPIs) used to limit SARS-CoV-2 transmission vary in their feasibility, appropriateness and effectiveness different contexts. In Bangladesh a national lockdown implemented March 2020 exacerbated poverty was untenable long-term, resurgence 2021 warranted renewed NPIs. Methods We developed an SEIR model for Dhaka District, parameterised from literature values calibrated data Bangladesh. discussed scenarios parameterisations with policymakers...
Abstract Diagnostics for COVID-19 detection are limited in many settings. Syndromic surveillance is often the only means to identify cases, but lacks specificity. Rapid antigen testing inexpensive and easy-to-deploy concerns remain about sensitivity. We examine how combining these approaches can improve guiding interventions low-income communities Dhaka, Bangladesh. Rapid-antigen-tests PCR validation was performed on 1172 symptomatically-identified individuals at home. Statistical models...
Background: Diagnostics for COVID-19 detection are limited in many settings. Syndromic surveillance is often the only means to identify cases, despite lack of specificity and asymptomatic cases. Rapid antigen testing inexpensive easy-to-deploy but concerns remain about sensitivity. We examine how combining these approaches can improve guiding interventions.Methods: Community-based volunteers were trained syndromically assess potential cases low-income communities Dhaka, Bangladesh. tests PCR...