Procter & Gamble (United Kingdom)
Open Access Institutional Repository at Robert Gordon UniversityNot a member yet
9563 research outputs found
Sort by
Prototype-guided spatial-spectral interaction network for hyperspectral anomaly detection.
Get PDFIn recent years, deep learning has emerged as one of the most widely utilized techniques in hyperspectral anomaly detection (HAD) with an impressive detection accuracy. However, the investigation into the diverse background representation and the spatial-spectral interaction remains underexplored. To tackle with this, we propose a novel framework namely the prototype-guided and spatial-spectral interaction network (PSSIN) for HAD in this paper. Specifically, an adaptive anomaly mask module is utilized to mitigate the interference of the background reconstruction caused by the blending of potential anomalies. Subsequently, we design a background-guided prototype autoencoder (BP-AE) to represent the backgrounds with various land cover types, incorporating two critical components: the background prototype module (BPM) and the spatial spectral interaction block (SSIB). To characterize different typical background features by a global perspective, BPM utilizes a prototype learning strategy with a self-attention mechanism, and a multivariate ensemble loss is employed for BPM to optimize the transformation of background features and the updating of a prototype codebook. To enhance the spatial-spectral utilization of window-based approach, SSIB first introduce a spatial-spectral interaction paradigm for HAD. The window-based self-attention branch is to mine spatial features characteristics, while the depth-wise convolution branch is to extract spectral features. These two branches in a parallel configuration interact with each other's features and then perform feature fusion. SSIB architecture not only broadens the receptive fields by concurrently modeling the intra-window and cross-window relationships but also facilitates bi-directional interactions between the spatial and spectral branches. Furthermore, the comprehensive experiments conducted on six authentic datasets have fully validated its superior performance
Blind quality assessment using channel-based structural, dispersion rate scores, and overall saturation and hue for underwater images.
Get PDFIn underwater subsea environments light attenuation, water turbidity, and limitations of the optical devices make the captured images suffer from poor contrast and quality, proportional degradation, low visibility, and low color richness. In recent years, various image enhancement techniques have been applied to improve the image quality, resulting in a new challenge, i.e., the quality assessment of the underwater images. In this study, we introduce an innovative and versatile blind quality assessment method for underwater images without using any references. Our approach leverages structural and contour-based metrics, combined with dispersion rate analysis, to quantify image degradation and color richness within an opponent color space. Specifically, we measure the proportional degradation by computing the edge magnitude using the directional Kirsch kernels, strengthened by image contour and saliency maps. To assess the color quality, chrominance dispersion rates and the overall saturation and hue are used to capture color distortions introduced by enhancement methods. The final quality score is obtained via a multiple linear regression model trained on extensive data sets. Experiments on three benchmark data sets have demonstrated the superior accuracy, consistency, and computational efficiency of the proposed method for both raw and enhanced underwater images
Blue hydrogen in the United Kingdom: a policy and environmental case study.
Get PDFBlue hydrogen is one of the energy carriers to be adopted by the United Kingdom to reduce emissions to net Zero by 2050 and its use is majorly influenced by policy and technological innovations. With more than 10 blue hydrogen facilities planning productive offtake from 2025, there is an urgent need to confirm the viability of these proposed facilities to aid decarbonisation and the path to conformity to policy regulation. This study discovers that the Acorn blue hydrogen facility can produce blue hydrogen within the low carbon hydrogen standard set by the United Kingdom's government. In this study, a detailed examination of hydrogen production techniques is conducted using lifecycle assessment (LCA) approach aimed to understand the environmental impact of producing 144 tons of hydrogen per day using Acorn hydrogen facility as a case study. This was followed on with sensitive analysis embracing steam and oxygen consumption and methane leakages, the ability of the facility meeting the low carbon hydrogen standard, economics, and the externality-priced production costs that embody the environmental impact. A gate-to-gate LCA shows that the Acorn hydrogen plant must aim at carbon capture rates of >90% to meet the set UK target of 20 gCO2e/MJLHV. The study further identifies from literature that the autothermal reforming (ATR) system with integrated carbon capture and storage (CCS) production technology as the most environmentally sustainable technology at present in comparison to commercially available options studied. This assessment helps to appraise potentially unintended causes and effects of the production of blue hydrogen that should aid future policy guidance and investments
Advanced DDoS attack detection and mitigation in software-defined networking (SDN) environments: an integrated machine learning approach.
Get PDFThe increasing sophistication of Distributed Denial of Service (DDoS) attacks poses critical challenges to network security, necessitating advanced detection and mitigation strategies. This research presents a machine learning-based framework that effectively distinguishes between normal and malicious traffic using engineered features such as unique source counts, flow counts, and packet rates. Among the models evaluated, Random Forest demonstrated the highest accuracy at 95.3%, showcasing its effectiveness in identifying diverse attack patterns. The framework incorporates a dynamic mitigation module that adapts in real-time to block or redirect malicious traffic while minimizing disruption to legitimate operations. Comprehensive evaluation confirms its scalability and relevance to real-world network environments. Despite its strengths, limitations include reliance on synthetic datasets and computational demands. Future work will address these challenges by integrating real-world traffic data, exploring advanced learning techniques, and enhancing resource efficiency. This study offers a scalable and adaptive solution to evolving DDoS threats
Improving geothermal resource assessment: a data-driven approach to chemical geothermometry using deep learning.
Get PDFThis study presents a deep learning model trained on a dataset of 674 water samples from Nevada to predict geothermal reservoir temperatures. The model outperforms traditional geothermometers and other machine learning models, achieving high accuracy and demonstrating global applicability when tested on samples from different geothermal fields around the world
Diffusion limit and the reactivity/affinity conundrum: implications for optimization and hit finding for irreversible modulators.
No full textIrreversible inhibition as a therapeutic modality has come of age over the previous decade. With minimal theoretical guidance for the design of an irreversible modulator, empirical optimization efforts often involve increasing the affinity of the small molecule while reducing the reactivity of the electrophile. The latter, as per prevalent opinion, is to ensure that binding dictates engagement and the reactive electrophile does not pose a safety liability arising from off-target reactivity. Here I argue that, like the second-order kinetic rate constant kcat/Km, the parameter kinact/KI is limited by the upper physical limit imposed by the rate of diffusion. This capping ensures that any attempt to improve the affinity of the electrophile-containing small-molecule at the limit will come with an equivalent trade-off in their reactivity. This has implications for both hit finding and lead optimization within targeted irreversible inhibition, especially for intractable targets with shallow pockets where the interactions are collision-induced second-order processes
Designing gender-inclusive sustainability curricula in engineering education.
Get PDFIncorporating sustainability into engineering education is essential for tackling global environmental issues. Nonetheless, the persistent gender imbalance in engineering remains a significant obstacle to achieving equitable and sustainable outcomes. This paper investigates the creation and execution of gender-inclusive sustainability curricula within engineering education. Through a thorough review of existing curricula, institutional policies, and current gender diversity initiatives, the study highlights effective strategies for integrating gender inclusivity into sustainability-oriented engineering courses. It specifically analyses how course content, teaching methods, and assessment practices influence the engagement of women and underrepresented genders in sustainability fields. By showcasing case studies and exemplary practices from institutions that have successfully adopted gender-sensitive approaches to sustainability, this research puts forward a framework for engineering programs to follow. This framework emphasizes intersectional perspectives on sustainability, illustrating how gendered experiences of environmental challenges, like climate change or resource scarcity, can inform engineering solutions. Additionally, it offers recommendations for promoting inclusive learning environments, ensuring diverse representation in sustainability initiatives, and advancing gender equity in engineering leadership. The findings indicate that gender-inclusive curricula not only empower underrepresented groups within engineering but also enhance the creation of more comprehensive, diverse, and innovative responses to sustainability challenges. This paper calls for a fundamental rethinking of engineering education to nurture a more inclusive and sustainable future for everyone
Resistance training beyond momentary failure: the effects of past-failure partials on muscle hypertrophy in the gastrocnemius.
Get PDFMuscle hypertrophy is often a desired goal of resistance training, and strategies that extend training beyond momentary failure may enhance muscular adaptations. Thus, the objective of this study was to assess whether performing additional past-failure partial repetitions beyond momentary failure increased muscle hypertrophy. A total of 23 untrained men completed a 10-week within-participant intervention study. This study comprised two weekly resistance training sessions of four sets of standing Smith machine calf raises. One limb was randomly allocated to the control condition performing sets to momentary failure (PLANTARMF), and the other limb was allocated to the test intervention that included additional past-failure partial repetitions in the lengthened position (DORSIvf). Muscle thickness of the medial gastrocnemius muscle was measured both pre- and post-intervention via ultrasound. Data were analysed within a Bayesian framework using a mixed-effect model with random effects to account for the within-participant design. The average treatment effect (ATE) was measured to assess any difference in condition and inferences made based on the ATE posterior distribution and associated Bayes Factor (BF). The main findings were that the PLANTARMF and DORSIVF legs increased medial gastrocnemius hypertrophy by 6.7 and +9.6%, respectively. The results identified an ATE favouring the inclusion of additional partial repetitions (0.62 [95%CrI: 0.21–1.0 mm; p(>0) = 0.998]) with ‘strong’ evidence (BF = 13.3) supporting a priori hypothesis. Therefore, when the goal is to train for maximum gastrocnemius hypertrophy over a relatively short time period, we suggest performing sets beyond momentary failure as a likely superior option
Developing an advanced multifuctional type.
Get PDFType IV hydrogen pressure vessels require liners with exceptional hydrogen barrier properties and robust thermo-mechanical performance. This research develops an advanced multifunctional liner through a novel hybrid architecture. We first fabricate a structural substrate from a High-Density Polyethylene (HDPE) and Polyamide 6 (PA6) blend, reinforced with functionalised clay nanofillers. A high-performance epoxy resin is then 3D-printed as a continuous, seamless inner layer onto this substrate, forming the primary permeation seal in direct contact with the hydrogen gas. This work characterizes the synergistic performance of this hybrid structure, demonstrating a significant reduction in permeability alongside enhanced thermo-mechanical strength for safer and more efficient hydrogen storage
Wikatoni: an agentic AI system for energy engineering workflows.
No full textCapturing expertise and enabling efficient information retrieval are critical in the energy sector, where high staff turnover can lead to significant knowledge loss. Retrieval Augmented Generation (RAG) offers a solution by grounding Large Language Model (LLM) outputs in documented sources, but its effectiveness is limited by reliance on general purpose embeddings. We present Wikatoni, an agentic AI system for energy engineering workflows that integrates a novel domain-specific embedding model. Wikatoni combines fine-tuned embeddings with agentic RAG, metadata filtering, and hybrid retrieval to improve document search, automated reporting, and workflow efficiency. Evaluation on internal enterprise offshore energy data shows that the domain-adapted embedding improves recall by 10%, and Wikatoni agentic RAG further increases answer accuracy by 14% compared to vanilla RAG with the base embedding model, achieving the best overall performance in context recall, faithfulness, and answer accuracy
Data sources:
Access Repository Dashboard
Do you manage Open Access Institutional Repository at Robert Gordon University? Access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard!