A5077265318- EEG and Brain-Computer Interfaces
- Gaze Tracking and Assistive Technology
- Neuroscience and Neural Engineering
- Interactive and Immersive Displays
- Neural dynamics and brain function
- Advanced Vision and Imaging
- Advanced Image and Video Retrieval Techniques
- Design Education and Practice
- Augmented Reality Applications
- Multimodal Machine Learning Applications
- Infrared Thermography in Medicine
- Blind Source Separation Techniques
- Architecture and Computational Design
- Tactile and Sensory Interactions
- Advanced Memory and Neural Computing
- Hand Gesture Recognition Systems
- Robotics and Sensor-Based Localization
- 3D Surveying and Cultural Heritage
- Glaucoma and retinal disorders
- Thermoregulation and physiological responses
- Face recognition and analysis
- Image Retrieval and Classification Techniques
- Medical Image Segmentation Techniques
- Diabetic Foot Ulcer Assessment and Management
- Advanced Image Processing Techniques
University of Malta
2016-2025
University of Edinburgh
2021
University of Ulster
2019
Solent NHS Trust
2019
University of Southampton
2019
Universidade Federal dos Vales do Jequitinhonha e Mucuri
2019
Institute of Linguistics
2017
Heinz Optical Engineering (United States)
2014
Abstract When a recurrent neural network (RNN) language model is used for caption generation, the image information can be fed to either by directly incorporating it in RNN – conditioning ‘injecting’ features or layer following ‘merging’ features. While both options are attested literature, there as yet no systematic comparison between two. In this paper, we empirically show that not especially detrimental performance whether one architecture another. The merge does have practical...
Objectives. To collect normative baseline data and identify any significant differences between hand foot thermographic distribution patterns in a healthy adult population. Design. A single-centre, randomized, prospective study. Methods. Thermographic was acquired using FLIR camera for the acquisition of both plantar dorsal aspects feet, volar hands, anterior lower limbs under controlled climate conditions. Results. There is general symmetry skin temperature same regions contralateral limbs,...
Image captioning has evolved into a core task for Natural Language Generation and also proved to be an important testbed deep learning approaches handling multimodal representations. Most contemporary rely on combination of convolutional network handle image features, recurrent encode linguistic information. The latter is typically viewed as the primary “generation” component. Beyond this high-level characterisation, CNN+RNN model supports variety architectural designs. dominant in...
Objective. Quantitative neurophysiological signal parameters are of value in predicting motor recovery after stroke. The novel role EEG-derived brain symmetry index for function prognostication the subacute phase stroke is explored. Methods . Ten male patients and ten matched healthy controls were recruited. Motor was first assessed clinically using MRC score, its derivative Motricity Index, Fugl–Meyer assessment score. EEG subsequently recorded with subjects at rest then during hand...
Aim . To evaluate the potential of thermography as an assessment tool for detection foot complications by understanding variations in temperature that occur type 2 diabetes mellitus (DM). Methods Participants were categorized according to a medical examination, ankle brachial index, doppler waveform analysis, and 10-gram monofilament testing into five groups: healthy adult, DM with no complications, peripheral neuropathy, neuroischaemia, arterial disease (PAD) groups. Thermographic imaging...
The application of non-destructive process analytical technologies in the area food science got a lot attention past years. In this work we used hyperspectral imaging to detect mould on milk agar and cheese. Principal component analysis is applied data localise visualise mycelia samples' surface. It also shown that PCA loadings obtained from set training samples can be new test presence these. For both cheeselets, first three principal components contained more than 99 % total variance....
Abstract Electroencephalogram (EEG) during pinprick stimulation has the potential to unveil neural mechanisms underlying sensorimotor impairments post-stroke. A proof-of-concept study explored event-related peak amplitude and oscillatory responses in healthy controls people with acute subuacute motor stroke, their relationship, what extent EEG somatosensory can predict impairment. In this study, 26 individuals participated, 10 an early subacute 6 age-matched controls. Pinpricks were applied...
One of the most important stages in a brain-computer interface (BCI) system is that extracting features can reliably discriminate data recorded during different user states. A popular technique used for feature extraction BCIs common spatial patterns (CSP) method, which provides set filters optimally between two classes least-squares sense. The method also yields are associated with relevant activity distinguishing classes. high recognition rates have been achieved led to its widespread...
We propose a multi-layer RNN for sign language detection. The system uses features extracted automatically from 2-stream convolutional neural network (CNN) that takes video image data and motion as input. also created dataset of videos containing signing "in the wild" to be used training evaluation purposes. compare our against state-of-the-art, attain an improvement around 18%, indicating is able leverage dynamic information hand during
In this paper we propose a biometric recognition system based on steady-state visual evoked potentials (SSVEPs), exploiting brain signals elicited by repetitive stimuli having constant frequency as identifiers. EEG responses to SSVEP flickering at different frequencies are recorded, and both mel-frequency cepstral coefficients (MFCCs) autoregressive (AR) reflection used discriminative features of the enrolled users. An analysis permanence across time response is also performed, data acquired...
Abstract Sketching is a natural and intuitive communication tool used for expressing concepts or ideas which are difficult to communicate through text speech alone. therefore variety of purposes, from the expression on two-dimensional (2D) physical media, object creation, manipulation, deformation in three-dimensional (3D) immersive environments. This sketching activities brings about range technologies which, while having similar scope, namely that recording interpreting sketch gesture...
For asymptomatic patients suffering from carotid stenosis, the assessment of plaque morphology is an important clinical task which allows monitoring risk rupture and future incidents stroke. Ultrasound Imaging provides a safe non-invasive modality for this, segmentation media-adventitia boundaries lumen-intima Carotid artery form essential part in this process. In paper, we propose novel Deep Neural Network as fully automated tool, its application delineating both boundary boundary. We...
Brain-computer interface (BCI) systems based on steady-state visual evoked potentials (SSVEPs) have gained considerable popularity because of the robustness and high information transfer rate these can provide. Typical SSVEP setups make use targets flashing at different frequencies, where a user's choice is determined from SSVEPs elicited by user gazing specific target. The range stimulus frequencies available for such limited variety factors, including strength as well comfort safety with...
The segmentation of media-adventitia and lumen-intima boundaries the Carotid Artery forms an essential part in assessing plaque morphology Ultrasound Imaging. Manual methods are tedious prone to variability thus, developing automated algorithms is preferable. In this paper, we propose use deep convolutional networks for boundary transverse longitudinal sections carotid ultrasound images. Deep have recently been employed with good success on image tasks, thus their application data, using...
This work develops a method for automatically extracting temperature data from prespecified anatomical regions of interest thermal images human hands, feet, and shins the monitoring peripheral arterial disease in diabetic patients. Binarisation, morphological operations, geometric transformations are applied cascade to extract required 44 predefined interest. The implemented algorithms region extraction were tested on 395 participants. A correct around 90% was achieved. process performed...
Purpose Brain-computer interface (BCI)-controlled wheelchairs have the potential to improve independence of people with mobility impairments. The low uptake BCI devices has been linked a lack knowledge among researchers needs end-users that should influence development.Materials and methods This study used semi-structured interviews learn about perceptions, needs, expectations spinal cord injury (SCI) patients regards BCI-controlled wheelchair. Topics discussed in interview include:...