26199 research outputs found
Sort by
Polyp Segmentation with the FCB-SwinV2 Transformer
Get PDFPolyp segmentation within colonoscopy video frames using deep learning models has the potential to automate the workflow of clinicians. This could help improve the early detection rate and characterization of polyps which could progress to colorectal cancer. Recent state-of-the-art deep learning polyp segmentation models have combined the outputs of Fully Convolutional Network architectures and Transformer Network architectures which work in parallel. In this paper we propose modifications to the current state-of-the-art polyp segmentation model FCBFormer. The transformer architecture of the FCBFormer is replaced with a SwinV2 Transformer-UNET and minor changes to the Fully Convolutional Network architecture are made to create the FCB-SwinV2 Transformer. The performance of the FCB-SwinV2 Transformer is evaluated on the popular colonoscopy segmentation bench-marking datasets Kvasir-SEG and CVC-ClinicDB. Generalizability tests are also conducted. The FCB-SwinV2 Transformer is able to consistently achieve higher mDice scores across all tests conducted and therefore represents new state-of-the-art performance. Issues found with how colonoscopy segmentation model performance is evaluated within literature are also re-ported and discussed. One of the most important issues identified is that when evaluating performance on the CVC-ClinicDB dataset it would be preferable to ensure no data leakage from video sequences occurs during the training/validation/test data partition
Digital Forensics
No full textDigital devices now pervade our lives. Very seldom does any aspect of our lives not interact with some element of the digital world. As both software and hardware develop in size, speed and complexity at an ever-increasing rate, so does the need for there be a means to investigate this forever growing area of science.
This chapter serves as an introduction to the complex and far-reaching world of cybercrime. It will introduce the reader to the detailed, complex and sometimes confusing world of Digital Forensics and explain some of the techniques and methods investigators use to tease out and make sense of fragments of data in a metaphorical mountain of information. We will, by means of a case study show the complexities involved in actually attributing the actions recorded on a computer to the individual sat at the keyboard, our cybercriminal.
Finally, looking at the predicted changes since the last edition of this chapter as our guide, we cast an eye to the future and consider what challenges the digital investigators of the future will face and perhaps how they may well face them
Drug-coated balloon angioplasty with rescue stenting versus intended stenting for the treatment of patients with de novo coronary artery lesions (REC-CAGEFREE I): an open-label, randomised, non-inferiority trial
No full textBackground
The long-term impact of drug-coated balloon (DCB) angioplasty for the treatment of patients with de novo coronary artery lesions remains uncertain. We aimed to assess the non-inferiority of DCB angioplasty with rescue stenting to intended drug-eluting stent (DES) deployment for patients with de novo, non-complex coronary artery lesions.
Methods
REC-CAGEFREE I was an open-label, randomised, non-inferiority trial conducted at 43 sites in China. After successful lesion pre-dilatation, patients aged 18 years or older with de novo, non-complex coronary artery disease (irrespective of target vessel diameter) and an indication for percutaneous coronary intervention were randomly assigned (1:1), via a web-based centralised system with block randomisation (block size of two, four, or six) and stratified by site, to paclitaxel-coated balloon angioplasty with the option of rescue stenting due to an unsatisfactory result (DCB group) or intended deployment of second-generation thin-strut sirolimus-eluting stents (DES group). The primary outcome was the device-oriented composite endpoint (DoCE; including cardiovascular death, target vessel myocardial infarction, and clinically and physiologically indicated target lesion revascularisation) assessed at 24 months in the intention-to-treat (ITT) population (ie, all participants randomly assigned to treatment). Non-inferiority was established if the upper limit of the one-sided 95% CI for the absolute risk difference was smaller than 2·68%. Safety was assessed in the ITT population. This study is registered with
ClinicalTrials.gov
,
NCT04561739
. It is closed to accrual and extended follow-up is ongoing.
Findings
Between Feb 5, 2021, and May 1, 2022, 2272 patients were randomly assigned to the DCB group (1133 [50%]) or the DES group (1139 [50%]). Median age at the time of randomisation was 62 years (IQR 54–69), 1574 (69·3%) of 2272 were male, 698 (30·7%) were female, and all patients were of Chinese ethnicity. 106 (9·4%) of 1133 patients in the DCB group received rescue DES after unsatisfactory DCB angioplasty. As of data cutoff (May 1, 2024), median follow-up was 734 days (IQR 731–739). At 24 months, the DoCE occurred in 72 (6·4%) of 1133 patients in the DCB group and 38 (3·4%) of 1139 in the DES group, with a risk difference of 3·04% in the cumulative event rate (upper boundary of the one-sided 95% CI 4·52; pnon-inferiority=0·65; two-sided 95% CI 1·27–4·81; p=0·0008); the criterion for non-inferiority was not met. During intervention, no acute vessel closures occurred in the DCB group and one (0·1%) of 1139 patients in the DES group had acute vessel closure. Periprocedural myocardial infarction occurred in ten (0·9%) of 1133 patients in the DCB group and nine (0·8%) in the DES group.
Interpretation
In patients with de novo, non-complex coronary artery disease, irrespective of vessel diameter, a strategy of DCB angioplasty with rescue stenting did not achieve non-inferiority compared with the intended DES implantation in terms of the DoCE at 2 years, which indicates that DES should remain the preferred treatment for this patient population
Chapter 10 – Playful practicals: breaking free from educational norms with online escape rooms
Get PDFRecreational escape rooms are becoming increasingly popular as social immersive experiences (Veldkamp et al., 2020). Escape rooms challenge players to work together in order to solve puzzles and complete tasks to ‘escape the room’ within a specific time. Play is well known to improve cognitive development in both children and adults (Piaget, 1962; Vygotsky, 1962) and gamification fosters the relationship between fun, focus and learning outcomes (Tulloch, 2014; van Gaalen et al., 2021). Escape rooms therefore offer a fun and novel way of developing interpersonal skills and effective team working and universities have begun to apply this approach across multiple disciplines. This chapter shares our experience of designing and delivering online escape rooms to develop practical skills in healthcare education and considers further playful applications across Higher Education
AN INVESTIGATION OF THE EFFECTS OF TRACTOR TYRE WIDTH ON SOIL COMPACTION AND CROP DAMAGE
Get PDFSoil compaction is a prevailing problem in the UK agricultural industry. This investigation focuses on the effect of tractor tyre width on a temporary grass crop used for both grazing and silage conservation. The tyres used were 650/75 R38 on the rear and 600/70R28 on the front, with wider tyres 900/70 R42 on the rear and 710/55 R30 on the front axles for comparison. Measurements identified the areas affected by the tyres, included the degree of soil compaction and damage to the crop. The results proved the wider the tyre, although creating a wider track, compacted a lower volume of soil when compared to the narrower tyre. The narrower tyre width compacted the soil to a greater depth where compaction is more difficult to relieve without disturbing the soil structure
A country that works for all children and young people: An evidence-based plan for improving children’s oral health with and through education settings
Get PDFD5.1. Intermediate report on user needs, SLICES services catalogue, access policies and training strategy
Get PDFThis deliverable is an intermediate report of the workpackage on user needs, services, access and training strategy. It presents the current view that will be further consolidated in
Deliverable D5.2 “Final report on user needs, SLICES service catalog, access policies and training strategy” (M40, December 2025) and that will provide the final view on these topics.
This deliverable starts by dealing with user needs by describing a methodology for identification of them. By applying it, three first blueprints have been identified and are
presented: the “post-5G” blueprint, the “cloud/edge” blueprint, and the "machine learning/federated learning" blueprint.
The second section of the deliverable recalls the three access types (trans-national/physical access, trans-national virtual access/remote access, and virtual access) and the three access modes (excellence-driven, market-driven, and wide access) that will be supported by SLICES.
The third section of the deliverable deals with the current vision of SLICES services catalogue. Services have been further divided into supporting and basic services. As a consequence of blueprints, a new category of services, SLICES Blueprint services, have been introduced. It aims at gathering the services provided by the blueprints, i.e., services specific to a particular research community. It is important to notice that this part also considered services from a pre-operation point of view and therefore also contains implementation considerations. This work has been carried out in collaboration with relevant workpackages: WP3 “Scientific and technical strategy and specifications”, WP6 “Operational framework”, and WP7 “Data management and ethics requirements”.
The fourth section focus on training activities in particular with respect to four objectives: i) to identify the training needs and training methodologies that will be followed; ii) to develop and provide the respective training material to organize SLICES-RI training events (training sessions, webinars, plugfests, hackathons) as well as shared teaching material used in SLICESRI for teaching basic skills at the convergence of computing and networking (SLICES Academy); iii) to provide guidelines, inter-site collaboration incentives and alignment with national programs pertaining to the teaching of key technical skills in the areas of interest of SLICES-RI (SLICES Academy); and iv) to facilitate researcher mobility, among the SLICES-RI member institutions and for the research community at large, for the exchange of know-how among the users of the facilities
Glucose influences endometrial receptivity to embryo implantation through O-GlcNAcylation-mediated regulation of the cytoskeleton
No full textPhenotypic changes to endometrial epithelial cells underpin receptivity to embryo implantation at the onset of pregnancy but the effect of hyperglycemia on these processes remains poorly understood. Here, we show that physiological levels of glucose (5 mM) abolished receptivity in the endometrial epithelial cell line, Ishikawa. However, embryo attachment was supported by 17 mM glucose as a result of glucose flux through the hexosamine biosynthetic pathway (HBP) and modulation of cell function via protein O-GlcNAcylation. Pharmacological inhibition of HBP or protein O-GlcNAcylation reduced embryo attachment in cocultures at 17 mM glucose. Mass spectrometry analysis of the O-GlcNAcylated proteome in Ishikawa cells revealed that myosin phosphatase target subunit 1 (MYPT1) is more highly O-GlcNAcylated in 17 mM glucose, correlating with loss of its target protein, phospho-myosin light chain 2, from apical cell junctions of polarized epithelium. Two-dimensional (2-D) and three-dimensional (3-D) morphologic analysis demonstrated that the higher glucose level attenuates epithelial polarity through O-GlcNAcylation. Inhibition of Rho (ras homologous)A-associated kinase (ROCK) or myosin II led to reduced polarity and enhanced receptivity in cells cultured in 5 mM glucose, consistent with data showing that MYPT1 acts downstream of ROCK signaling. These data implicate regulation of endometrial epithelial polarity through RhoA signaling upstream of actomyosin contractility in the acquisition of endometrial receptivity. Glucose levels impinge on this pathway through O-GlcNAcylation of MYPT1, which may impact endometrial receptivity to an implanting embryo in women with diabetes.
NEW & NOTEWORTHY Understanding how glucose regulates endometrial function will support preconception guidance and/or the development of targeted interventions for individuals living with diabetes wishing to embark on pregnancy. We found that glucose can influence endometrial epithelial cell receptivity to embryo implantation by regulating posttranslational modification of proteins involved in the maintenance of cell polarity. Impaired or inappropriate endometrial receptivity could contribute to fertility and/or early pregnancy complications caused by poor glucose control
CVAM-Pose: Conditional Variational Autoencoder for Multi-Object Monocular Pose Estimation
Get PDFEstimating rigid objects' poses is one of the fundamental problems in computer vision, with a range of applications across automation and augmented reality. Most existing approaches adopt one network per object class strategy, depend heavily on objects' 3D models, depth data, and employ a time-consuming iterative refinement, which could be impractical for some applications. This paper presents a novel approach, CVAM-Pose, for multi-object monocular pose estimation that addresses these limitations. The CVAM-Pose method employs a label-embedded conditional variational autoencoder network, to implicitly abstract regularised representations of multiple objects in a single low-dimensional latent space. This autoencoding process uses only images captured by a projective camera and is robust to objects' occlusion and scene clutter. The classes of objects are one-hot encoded and embedded throughout the network. The proposed label-embedded pose regression strategy interprets the learnt latent space representations utilising continuous pose representations. Ablation tests and systematic evaluations demonstrate the scalability and efficiency of the CVAM-Pose method for multi-object scenarios. The proposed CVAM-Pose outperforms competing latent space approaches. For example, it is respectively 25% and 20% better than AAE and Multi-Path methods, when evaluated using the ARVSD metric on the Linemod-Occluded dataset. It also achieves results somewhat comparable to methods reliant on 3D models reported in BOP challenges