Clustering-based segmentation for flood area detection across different color spaces uses flood-affected images. Segmentation performance with multiple metrics and bootstrap confidence interval estimation identify most effective color space. doi.org/10.1109/FMLDS6… Salwa Al Garea

Parametric sensitivity of adaptive reinforcement learning algorithms for battery control in microgrids demonstrates that careful tuning of discretization/exploration parameters improves the performance over traditional ε-greedy methods in EMS. doi.org/10.1016/j.rene… Deepak Kumar Panda

A pipeline combines preprocessing, dimensionality reduction, deep learning visualization techniques to standardize & enhance underwater video datasets for improved image quality & feature analysis in ecological monitoring. doi: doi.org/10.1109/WiDS-P… Amani Homoud #Video #DataScience

A statistical feature learning approach for classifying microscopy images of butterfly wings and moth eyes, aims to enhance biomimicry-driven material design through exploratory analysis, feature extraction & deep learning. doi.org/10.1109/WiDS-P… Katie Shanks #Microscopy #Data

A novel analytical method for designing PID controllers via dominant pole placement for SOPTD systems, avoids Pade approximations by transforming the delay term into discrete poles. It leads to irregular, discontinuous, multi-modal stability regions. doi: doi.org/10.1016/j.ces.…

Reinforcement learning based energy management system for a PV-integrated battery storage system in Malaysian chemical manufacturing plant addresses the intermittency of solar energy & limitations of traditional optimization methods. doi.org/10.1007/978-98… #reinforcementlearning

Happy to serve as Session Chair in The 1st Int. Online Conf. on Fractal & Fractional. We invite conf. papers within 4-8 pages and will be published by Computer Sciences & Mathematics Forum, MDPI. Abstract deadline: 12 Dec 2025. Paper deadline: 28 May 2026. sciforum.net/event/IOCFF2026

Bayesian & frequentist implementation comparisons of RNN & LSTM neural networks for air quality time-series regression tasks uses identical structures/cross-validation. Model performance is evaluated through the R^2, Pearson, & Spearman correlation metrics doi.org/10.1007/978-98…

2026 China Scholarship Council and University of Exeter PhD Scholarships are now open. I have 2 First Supervision PhD projects under "Earth and Environmental Science" and also 2 projects as Second Supervisor under "Engineering" Theme exeter.ac.uk/study/pg-resea… Deadline: 1st Dec 2025

A PhD studentship is available with me through the EPSRC DLA scheme. Deadline: 12th January 2025. Project description: exeter.ac.uk/v8media/recrui… Application: exeter.ac.uk/study/funding/… Academic Co-supervisor: Dr Shuyue Lin , Industrial Co-supervisor: Dr Mohammad Abusara #PhDFunding

ESRC SWDTP invites PhD applications with proposal in the area of Economic & Social Sciences. I am happy to help developing research proposals. Deadline: 9th January 2026. Application web: exeter.ac.uk/study/pg-resea… ESRC SWDTP pathway details: swdtp.ac.uk/prospective-st… #PhDFunding

DenseNet, ResNet-LSTM, and Vision Transformer models classify underwater videos, revealing trade-offs between accuracy, computational cost, and memory usage. Publisher: International Conference on High Performance Computing and Communications doi: doi.org/10.1109/HPCC67… Amani Homoud

Supervised deep learning of SEM and confocal microscopy images predicts the surface roughness and stochastic texture features of PDMS polymers treated with different sandpaper grit sizes, using modality-specific preprocessing pipelines. doi: doi.org/10.1109/ICWITE… #Microscopy

The k-means clustering of geospatial data from the Global Terrorism Database identifies patterns and insights in terrorist attacks, with a focused case study on the UK. doi.org/10.1007/978-3-… #DataScience #Peace #UNSDG16 #History #geospatial #Pattern

A smart home energy prediction framework uses Gaussian Process Regression with Bayesian hyper-parameter optimization and domain-specific feature engineering to accurately forecast energy consumption and renewable production from heterogeneous data sources. doi.org/10.1007/978-3-…

Computational analysis of a 6D nonlinear epidemiological model with vaccination shows that increasing the vaccination rate beyond a critical threshold induces a Hopf bifurcation, leading to oscillatory epidemic waves. doi.org/10.1109/InCoWo… Dr. Lamia Alsaloom #Nonlinear #Dynamics

A hybrid ordinary differential equation solving approach combine forward Euler numerical integration scheme with Luenberger observer through a switching mechanism. This method significantly improves stability and accuracy. doi.org/10.1080/002071… Dr.Aisha Mosa Al Hayzea #Control #Mathematics

Underwater object detection for unknown marine species combines image enhancement techniques with the YOLOv8 model addressing visibility degradation and misclassification issues in support of UN SDG 14 (Life Below Water) doi.org/10.1109/MELECO… Amani Homoud #DeepLearning #Underwater