157082 research outputs found
Sort by
Augmented Reality Navigation in Robot-Assisted Surgery
No full textIn the last few decades, major complications in surgery have emerged as a significant public health issue, and so the practical implementation of safety measures to prevent injuries and deaths in different phases of surgery is required. Augmented reality (AR) is considered one of the most promising solutions for safer procedures in several surgical specialties. Fusing patient-specific preoperative information, typically 3D models extracted from CT scans or MRI, with real-time surgical images allows the surgeon to have detailed information on the anatomical structure of the surgical target intraoperatively. The coupling of AR and robotics represents the next step toward introducing awareness into the surgical room, thus enhancing the surgeon’s perceptual, cognitive, and manipulative capabilities. This chapter describes the main areas involved in an AR navigation system integrated into a robotic teleoperated platform. It will describe the modalities to obtain a patient-specific virtual model, the methodologies to develop an AR navigation system, and the methods to implement it in a teleoperated surgical robotic platform. Recent advances in the field are also presented, providing as an example a novel integrated system for real-time AR navigation in robotic minimally invasive surgery (RMIS), composed of a robotic endoscopic camera, teleoperated implementing a software-based remote center of motion (RCM), and an AR navigation software based on an initial manual registration of virtual 3D models with the real anatomy
UHS Microfluidic Investigation of Hydrogen-Methane Mixtures at 58.6 bar and 20–50 °C
No full textHydrogen storage in depleted gas reservoirs has gained attention as a large-scale option for intermittent green energy. Few experimental studies have investigated the influence of its native natural gas on the storage process. We investigate the influence of methane on UHS through microfluidic experiments with pre-blended and port-injection mixtures at 58.6 bar and 20–50 °C. Our drainage results indicate a positive correlation between front stability and capillary number. A wider range of pores is invaded with increasing methane content. In-situ mixing exerts a negative effect on storage capacity (as high as 13 %) and flow connectivity. Conversely, a higher withdrawal efficiency is observed for hydrogen, with one-third of the remaining ganglia occupying pore corners and parts of pores. Presence of methane in the mixtures increases the percentage of multiple pore-spanning ganglia from 1.6 % to 23–25 % at the end of imbibition. While the presence of methane can enhance connectivity, the process of in-situ mixing may reduce storage capacity, indicating a complex effect on the overall UHS process which required further investigation especially over higher-pressure conditions
IO-K-Means:Iterative Optimization for Centroids in K-Means
No full textK-means is a widely used unsupervised learning algorithm, but its clustering performance can be heavily influenced by the choice of initial centroids and easily falls into local optima. To address these limitations, we propose IO-K-means, an improved K-means algorithm based on iterative centroid optimization. First, instead of random initialization, we employ reverse nearest-neighbor relationships (RNNs) to select higher-quality initial centroids, ensuring a more representative starting point. Second, we introduce an iterative refinement mechanism: in each iteration, a novel within-cluster compactness measure identifies which centroids require adjustment, and two operations—interconnection and perturbation—are applied to fine-tune their positions. Through multiple iterations, the algorithm progressively improves centroid placement, ultimately assigning data points to their nearest centroids for final clustering. To verify the proposed IO-K-means, we make experiments on 16 real-world datasets. The experimental results show that IO-K-means outperforms the state-of-the-art (SOTA) extensions of K-means especially including Nie’s work (TKDE2022).</p
Quantifiers for Differentiable Logics in Rocq
No full textThe interpretation of logical expressions into loss functions has given rise to so-called differentiable logics. They function as a bridge between formal logic and machine learning, offering a novel approach for property-driven training. The added expressiveness of these logics comes at the price of a more intricate semantics for first-order quantifiers. To ease their integration into machine-learning backends, we explore how to formalize semantics for first-order differentiable logics using the Mathematical Components library in the Rocq proof assistant. We seek to give rigorous semantics for quantifiers, verify their properties with respect to other logical connectives, as well as prove the soundness and completeness of the resulting logics.</p
Miniaturized mechanical design of a spiral-scanning laser ablation instrument for advanced surgical applications
No full textRecent developments in picosecond pulsed laser ablation show considerable promise for the minimally invasive removal of neoplastic tissue. This technology is particularly noteworthy for its selective targeting of diseased tissue while substantially reducing collateral damage to adjacent healthy areas. Such precision is critical for interventions in delicate and anatomically complex regions. However, integrating this advanced technique into constrained or intricate anatomical spaces presents a significant practical barrier. Overcoming these limitations is essential for the technology to achieve its full clinical potential and broader adoption. Building upon our earlier spiral-scanning laser ablation prototype, we have engineered a substantially enhanced probe. This next-generation probe features a smaller footprint and improved scanning capabilities. Our newly developed probe has a compact 12 mm outer diameter and 300 mm in length, making it well-suited for endoscopic and laparoscopic procedures. A key innovation is its distal scanning tip, designed to generate Archimedean spiral trajectories. This ensures an even distribution of laser energy across the target area. The system achieves a laser spot diameter of 20 microns and can completely scan a 0.6 mm radius area in approximately one and a half minutes by following a spiral trajectory at a speed of 2 mm per second. The development process required solving several key engineering challenges, including mechanical miniaturization, enabling perpendicular laser emission from the probe surface during scanning, and maintaining fiber tip stability throughout rapid spiral movements. Addressing these challenges has significantly enhanced the probe’s effectiveness and reliability for potential use in clinical scenarios. Conducted experimental results have confirmed the superior performance of our new probe, demonstrating consistent, tightly packed spiral patterns that achieve nearly complete target coverage with minimal unscanned regions. Furthermore, real-time imaging during experimental setups confirmed that the laser spot maintains a constant speed of 2 mm/s along the Archimedean spiral during scanning. Preliminary results underscore significant advancements in scanning sensitivity, speed, and coverage compared to the initial prototype, affirming the device’s strong potential for diverse clinical applications
Uncovering complementary information sharing in spider monkey collective foraging using higher-order spatial networks
No full textCollectives are often able to process information in a distributed fashion, surpassing each individual member’s processing capacity. In fission-fusion dynamics, where group members come together and split from others often, sharing complementary information about uniquely known foraging areas could allow a group to track a heterogenous foraging environment better than any group member on its own. We analyse the partial overlaps between individual spider monkey core ranges, which we assume represent the knowledge of an individual during a given season. Sets of individuals with complementary overlaps are identified, showing a balance between redundantly and uniquely known portions, and we use simplicial complexes to represent these higher-order interactions. The structures of the simplicial complexes show holes in various dimensions, revealing complementarity in the foraging information that is being shared. We propose that the complex spatial networks arising from fission-fusion dynamics allow for adaptive, collective processing of foraging information in dynamic environments
Building Dynamic Capabilities for Fashion Retailer Internationalisation
No full textThis chapter seeks to examine the role of dynamic capabilities (DCs) in fashion retail internationalisation. It explores the constantly changing global fashion retail landscape and how environmental turbulence and technological advances have made the possession of strong DCs a source of competitive advantage. The chapter is underpinned by Teece’s (2007) tripartite conceptualisation of DCs as comprising sensing, seizing and reconfiguring activities. Whilst Teece’s (2007) framework provides the basis for the chapter, it will also examine how this theory has been developed by other scholars to gain a comprehensive understanding of the significance of DCs for fashion firms operating or seeking to expand in international markets. Practical examples of how fashion brands sense and seize international opportunities and how they harness their resource base to leverage them are also provided. The chapter concludes by affirming the importance of managers and specifically, teams, in the development and maintenance of DCs within firms
Taking charge of my health in my way! Exploring consumers' continuous usage intention and brand loyalty toward health wearable devices: a hybrid method of PLS-SEM and artificial neural network
No full textFacing the low continuous use intention on health wearable devices (HWDs) today, this study evaluates the effects of both individual and technical attributes on perceived satisfaction, thus affecting brand loyalty and continuous use intention by extending the stimulus-organism-response (S-O-R) theory. An online self-administered questionnaire, consisting of 360 valid responses, is adopted for data collection and then conducts a hybrid analytical method that combines partial least squares structural equation modelling (PLS-SEM) and artificial neural network (ANN). The findings show that health-consciousness, convenience, measurement accuracy and affordance significantly affect perceived satisfaction, which in turn affects brand loyalty and continuous use intention. Besides, brand loyalty positively relates to ongoing use intention. This study strengthens the role of S-O-R theory in continuous behavioural intention on HWDs by integrating individual and technical attributes and provides practical suggestions to manufacturers for enhancing consumer satisfaction and brand loyalty.</p
Multi-Modal Multi-Stage Multi-Task Learning for Occlusion-Aware Facial Landmark Localisation
No full textThermal facial imaging enables non-contact measurements of face heat patterns that are valuable for healthcare and affective computing, but common occluders (glasses, masks, scarves) and the single-channel, texture-poor nature of thermal frames make robust landmark localisation and visibility estimation challenging. We propose M3MSTL, a multi-modal, multi-stage, multi-task framework for occlusion-aware landmarking on thermal faces. M3MSTL pairs a ResNet-50 backbone with two lightweight heads: a compact fully connected landmark regressor and a Vision Transformer occlusion classifier that explicitly fuses per-landmark temperature cues. A three-stage curriculum (mask-based backbone pretraining, head specialisation with a frozen trunk, and final joint fine-tuning) stabilises optimisation and improves generalisation from limited thermal data. On the TFD68 dataset, M3MSTL substantially improves both visibility and localisation: the occlusion accuracy reaches 91.8% (baseline 89.7%), the mean NME reaches 0.246 (baseline 0.382), the ROC–AUC reaches 0.974, and the AP is 0.966. Paired statistical tests confirm that these gains are significant. Our approach aims to improve the reliability of temperature-based biometric and clinical measurements in the presence of realistic occluders
The main determinants of global portfolio flows dynamics
No full textAny episode of global financial turbulence can lead to the freezing and significant reversal of portfolio flows across different countries, emphasizing the need for adequate pre-emptive financial policy responses. This study investigates the sensitivities of global portfolio flows dynamics to a variety of global (push) and domestic (pull) determinants in a sample of 43 countries for the period from Q1 2005 to Q4 2020. Using panel regressions incorporating country fixed effects, we corroborate previous empirical evidence that push determinants remain the most important in driving portfolio inflows/outflows. The analysis shows that portfolio inflows/outflows decrease with the level of the expected change in the US central bank policy rate, world inflation surprise index, macro-risk index, and increase with better economic sentiment expectations, and investors’ confidence index. The biggest difference in exposures of global portfolio flows dynamics comes with the level of the short-term world interest rates, implying that they may discourage portfolio outflows but not inflows
Access Repository Dashboard
Do you manage Heriot Watt Pure? Access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard!