85000 research outputs found
Sort by
GIFCOS-DT: One Stage Detection of Gastrointestinal Tract Lesions From Endoscopic Images With Distance Transform
No full textThis study aims at developing a computer-aided diagnostic system based on deep learning techniques for detecting various typical lesions during endoscopic examinations in the human gastrointestinal tract. We propose a lesion detection model, namely called GIFCOS-DT, that is built upon a one-stage backbone for object detection (Fully Convolutional One-Stage Object Detection - FCOS). For the proposed model, to deal with the diverse shapes and appearance of the lesions, we introduce a new loss function based on Distance Transform, that better describes the elongated or curved shapes of lesions than the common loss functions like Intersection of Union or centroid loss. We then deploy the detection model on an embedded device that connects to the endoscopic machine to assist endoscopists during examinations. A multithread technique is employed to accelerate the processing times of all steps of the system. Extensive experiments have been conducted on two challenging datasets, the benchmark dataset (Kvasir-SEG) and our newly collected dataset (IGH_GIEndoLesion-SEG), which include various typical lesions of the gastrointestinal (GI) tract (reflux esophagitis, esophageal cancer, helicobacter pylori negative gastritis, helicobacter pylori positive gastritis, gastric cancer, duodenal ulcer, and colorectal polyps). Experimental results show that our proposed methods outperform the original FCOS by 4.2% and 7.2% on Kvasir-SEG and our collected dataset respectively in terms of the average AP50 score. On the Kvasir-SEG dataset, the GIFCOS-DT outperforms state-of-the-art detectors such as Faster R-CNN, DETR, YOLOv3, and YOLOv4. Our developed supporting system for lesion detection can run at 14.85 FPS on an embedded Jetson AGX Xavier or 31.92 FPS on an RTX 3090. The detection results of various types of lesions are promising, mostly on malignant lesions such as gastric cancers. The proposed system can be deployed as an assistant tool in endoscopy to reduce missed detection of lesions. Our code is available at https://github.com/hanhtran201/GIFCOS-DT.</p
[De]Compose: Conversations between the dead, the soil and the trees
No full textBeginnings that are endings that are beginnings: where do you end and where do I begin?
We all know about the symbiotic relationship between our expirations and those of green living organisms.
When we breathe, we take in the atoms that were once connected to another being.
This begs the question: are we really separate from each other and the ‘natural’ world around us?
Following the death of a human, their burial results in an entwining with the soil and their elemental transmigration from one into the other.
Through the practices of tracing around our bodies,
letter-writing and the arrangement of found materials, we will address our future (dead) selves.</p
Structure, Properties, and Applications of Silica Nanoparticles: Recent Theoretical Modeling Advances, Challenges, and Future Directions
No full textSilica nanoparticles (SNPs), one of the most widely researched materials in modern science, are now commonly exploited in surface coatings, biomedicine, catalysis, and engineering of novel self-assembling materials. Theoretical approaches are invaluable to enhancing fundamental understanding of SNP properties and behavior. Tremendous research attention is dedicated to modeling silica structure, the silica-water interface, and functionalization of silica surfaces for tailored applications. In this review, the range of theoretical methodologies are discussed that have been employed to model bare silica and functionalized silica. The evolution of silica modeling approaches is detailed, including classical, quantum mechanical, and hybrid methods and highlight in particular the last decade of theoretical simulation advances. It is started with discussing investigations of bare silica systems, focusing on the fundamental interactions at the silica-water interface, following with a comprehensively review of the modeling studies that examine the interaction of silica with functional ligands, peptides, ions, surfactants, polymers, and carbonaceous species. The review is concluded with the perspective on existing challenges in the field and promising future directions that will further enhance the utility and importance of the theoretical approaches in guiding the rational design of SNPs for applications in engineering and biomedicine.</p
Understanding the needs and preferences for cancer care among First Nations people: An integrative review
No full textAIM: This systematic review aimed to identify the needs and preferences for cancer care services among Australian First Nations people.
DESIGN: Integrative review.
DATA SOURCES: An integrative review was conducted. A wide range of search terms were used to increase the sensitivity and specificity of the searches in electronic databases. Methodological quality assessment, data extraction, was conducted independently by two reviewers, and a narrative synthesis was conducted.
RESULTS: Forty-two studies were included. A total of 2965 Australian First Nations adults, both men and women of various ages across the lifespan, were represented; no First Nations children affected by cancer were represented in the studies. Three themes emerged which included: (1) discrimination, racism and trauma, resulting from colonization, directly impacted First National people's cancer care experience; (2) cultural ways of knowing, being and doing are fundamental to how First Nations people engage with cancer care services; and (3) First Nations people need culturally safe person-centred cancer care services that address practical needs.
CONCLUSION: Most participants represented in this review experienced discrimination, racism and trauma, resulting from colonization, which directly negatively impacted Aboriginal peoples' cancer care experience. While the Optimal Cancer Pathway (OCP) was launched in Australia several years ago, people with cancer may continue to experience distressing unmet care needs.
PATIENT OR PUBLIC CONTRIBUTION: Our team includes both First Nations people, non-First Nations researchers and healthcare professionals with expertise in cancer care. The researchers employed decolonizing restorative approaches to ensure voice, respect, accountability and reciprocity in this review work.
IMPLICATIONS FOR NURSING PRACTICE: Members of the multidisciplinary team including nurses and policymakers should reflect on these findings, ensure that they have up-to-date cultural safety training and stand together with Indigenous and non-Indigenous cancer leaders to take proactive steps to stamp out and dismantle oppression in health, and safely implement the OCP.</p
Invisible Designing: Emotional and Affective Labour in Relational Participatory Practices
No full textThis paper presents emotional and affective labour as invisible designing required to nurture social relationships in participatory practices. Participatory Design (PD) is often framed through visible acts of designing, such as envisioning, co-creating and prototyping. Yet little is shared of the emotional and affective labour and its association with design that enables the condition of safe and comfortable participation. Initiating, nurturing and sustaining social relationships requires ongoing long-term commitment and care that extends beyond acknowledged practices of designing. Examples of emotional labour required for participatory practice are illustrated here through reflecting on our experience in infrastructuring an online transcultural peer mentoring programme for socially-engaged women creatives working in four Asia-Pacific countries. By making these invisible acts explicit, we hope to prompt collective consciousness of the labour involved in structuring social relationships and to support PD practitioners to acknowledge and account for this work in their practice.</p
Process-Induced Molecular-Level Protein–Carbohydrate–Polyphenol Interactions in Milk–Tea Blends: A Review
No full textThe rapid increase in the production of powdered milk-tea blends is driven by a growing awareness of the presence of highly nutritious bioactive compounds and consumer demand for convenient beverages. However, the lack of literature on the impact of heat-induced component interactions during processing hinders the production of high-quality milk-tea powders. The production process of milk-tea powder blends includes the key steps of pasteurization, evaporation, and spray drying. Controlling heat-induced interactions, such as protein-protein, protein-carbohydrate, protein-polyphenol, carbohydrate-polyphenol, and carbohydrate-polyphenol, during pasteurization, concentration, and evaporation is essential for producing a high-quality milk-tea powder with favorable physical, structural, rheological, sensory, and nutritional qualities. Adjusting production parameters, such as the type and the composition of ingredients, processing methods, and processing conditions, is a great way to modify these interactions between components in the formulation, and thereby, provide improved properties and storage stability for the final product. Therefore, this review comprehensively discusses how molecular-level interactions among proteins, carbohydrates, and polyphenols are affected by various unit operations during the production of milk-tea powders.</p
Climate change, Murray Valley encephalitis virus, and the imperative of one health: navigating the challenge in Western Australia
No full textThe Murray Valley encephalitis virus (MVEV), a rare but fatal mosquito-borne flavivirus, is a significant public health concern in endemic regions of Australia and Papua New Guinea. Major outbreaks of MVE virus were documented in southern and eastern Australia in the 1950s, 1970s, and 2010s, while other instances were mainly confined to Northern Australia. The virus, named after its initial detection in the Murray Valley of Southeastern Australia in 1917–182. Transmitted through the bite of the Culex annulirostris mosquito, commonly known as the common banded mosquito, the virus is closely related to regional flaviviruses like – Japanese encephalitis virus (JEV), and Kunjin virus, which share similar transmission characteristics. Water birds such as herons and egrets (Order – Pelecaniformes, Ciconiiformes) are considered to be reservoirs for MVEV. Due to the absence of specific treatments or vaccines, MVEV poses a growing challenge. This article will focus on the recent resurgence of MVEV in Western Australia (WA) in 2024, highlighting the interconnection between climate change and the virus, and measures implemented by local authorities to mitigate its impact within the framework of One Health.</p
Caralluma fimbriata Extract Improves Vascular Dysfunction in Obese Mice Fed a High-Fat Diet
No full textBackground: Obesity is a risk factor for developing cardiovascular diseases (CVDs) by impairing normal vascular function. Natural products are gaining momentum in the clinical setting due to their high efficacy and low toxicity. Caralluma fimbriata extract (CFE) has been shown to control appetite and promote weight loss; however, its effect on vascular function remains poorly understood. This study aimed to determine the effect that CFE had on weight loss and vascular function in mice fed a high-fat diet (HFD) to induce obesity, comparing this effect to that of lorcaserin (LOR) (an anti-obesity pharmaceutical) treatment. Methods: C57BL/6J male mice (n = 80) were fed a 16-week HFD to induce obesity prior to being treated with CFE and LOR as standalone treatments or in conjunction. Body composition data, such as weight gain and fat mass content were measured, isometric tension analyses were performed on isolated abdominal aortic rings to determine relaxation responses to acetylcholine, and immunohistochemistry studies were utilized to determine the expression profiles on endothelial nitric oxide synthase (eNOS) and cell stress markers (nitrotyrosine (NT) and 78 kDa glucose-regulated protein (GRP78)) in the endothelial, medial and adventitial layers of aortic rings. Results: The results demonstrated that CFE and CFE + LOR treatments significantly reduced weight gain (17%; 24%) and fat mass deposition (14%; 16%). A HFD markedly reduced acetylcholine-mediated relaxation (p < 0.05, p < 0.0001) and eNOS expression (p < 0.0001, p < 0.01) and significantly increased NT (p < 0.05, p < 0.0001) and GRP78 (p < 0.05, p < 0.01, p < 0.001). Obese mice treated with CFE exhibited significantly improved ACh-induced relaxation responses, increased eNOS (p < 0.05, p < 0.01) and reduced NT (p < 0.01) and GRP78 (p < 0.05, p < 0.01) expression. Conclusions: Thus, CFE alone or in combination with LOR could serve as an alternative strategy for preventing obesity-related cardiovascular diseases.</p
Enhancing Generative Artificial Intelligence Algorithms through Evolutionary Computing
No full textGenerative Artificial Intelligence (AI) has emerged as a transformative technology, enabling the creation of realistic data and facilitating numerous applications in image generation, text synthesis, and predictive modeling. Despite significant progress, challenges persist in optimizing the performance and efficiency of generative algorithms, particularly in the context of Generative Adversarial Networks(GANs) and diffusion models.
This thesis addresses these challenges through a comprehensive investigation into evolutionary approaches for enhancing the effectiveness and efficiency of generative AI algorithms. Beginning with a thorough review of foundational concepts in generative AI, transfer learning, and knowledge distillation,the thesis identifies key research gaps and outlines the research questions driving the inquiry.
The primary research questions explored in this thesis are:
• How can the training time of EvolutionaryGANs(E-GANs) be shortened without compromising performance?
• How can knowledge distillation techniques be integrated with E-GANs to improve convergence speedand resource efficiency?
• What novel approaches can be developed to optimize the noises cheduler in diffusion models, thereby enhancing sampling speed and quality?
The study begins with an extensive review of foundational concepts in generative AI, transfer learning, and knowledge distillation, identifying critical research gaps and areas for improvement. Through a series of innovative methodologies and rigorous experiments, this thesis introduces several groundbreaking solutions. It presents the Partial Transfer Training-based E-GAN (PT-EGAN) and the Knowledge Distillation Evolutionary GAN (KDE-GAN), which significantly reduce training time and improve convergence speed of Evolutionary GANs while maintaining high-quality output. Additionally, a novel evolutionary approach is developed to optimize the noise scheduler in diffusion models,resulting in enhanced sampling efficiency and performance.
Overall, this thesis contributes to advancing the field of generative AI by introducing novel evolutionary approaches that improve both the performance and efficiency of state-of-the-art generative algorithms. These findings offer valuable insights for researchers and practitioners seeking to leverage generativeAI for various real-world applications, from image synthesis to predictive modeling.</p
Achieving Diverse and Competitive Designs Using a Parallel Bi-directional Evolutionary Structural Optimisation Framework
No full textTopology optimisation is increasingly used as a computer-aided design tool. It enables designers to create elegant and efficient structures. The primary goal of topology optimisation is to identify the optimal design that satisfies prescribed requirements for structural performance subject to certain constraints. However, it is often challenging to incorporate various aesthetic considerations or functional requirements due to their complexities and uncertainties. For this reason, it is desirable to have an optimisation approach that can generate diverse solutions for the same loading and boundary conditions, allowing the designer to make selections based on their personal preferences or other requirements. Therefore, this thesis aims to develop automatic design tools that can effectively generate diverse, freeform and optimised solutions, providing architects and structural engineers with multiple choices in their conceptual design phase.
The thesis consists of three main parts. Firstly, recognising the critical importance of computational efficiency in large-scale design problems, the traditional Bi-directional Evolutionary Structural Optimisation (BESO) method is enhanced with modern parallel computing techniques. Several methods are integrated to boost computational efficiency, including the iterative solver, the reanalysis approach, the partial differential equation (PDE)-based filter scheme, and the hard-kill BESO option. The finite element analysis is facilitated by FEniCS, an open-source platform that automates the solution of complex PDEs in a simple implementation. The developed method can efficiently solve the high-resolution topology optimisation problems, serving as the foundation for the remaining two parts of this thesis.
In the second part, two perturbation approaches are integrated into the parallel BESO framework, aiming to create diverse and structurally competitive designs. By applying perturbations to load directions and altering material properties near supports during interim iterations, noise is introduced into the sensitivity analysis, thereby resulting in a variety of solutions. Furthermore, the approaches hold the potential to control the similarity of diverse designs by adjusting the perturbation functions. This component offers a powerful tool for designers to explore the design space, paving the way for innovative design solutions.
In the third part, a three-field BESO method with a variable-radius filter is proposed to further investigate the generation of diverse designs. In this study, both the density of elements and the filter radius are considered variables in the design process. Compared to the second part of this thesis, the size constraints brought by the filter scheme are relaxed, thus enabling the achievement of designs with multiple morphologies. To enhance the range of possible designs, this method includes the use of both isotropic and anisotropic filters. By applying hard or soft constraints to the filter radius, the proposed approach enables the creation of a variety of structural forms, such as extruded shapes, plate-like structures, and a blend of free-form structures with the aforementioned types. Moreover, by manipulating the filter radius values across different regions within the design domain, it is possible to generate composite designs featuring a variety of structural morphologies. This adaptability allows designers to customise filter radius values in targeted regions to develop structures that meet both aesthetic preferences and functional requirements.
In summary, this thesis establishes novel approaches to achieving diverse and competitive designs based on the parallel BESO method. The main contributions of this thesis are outlined below. First, a parallel BESO framework is proposed and implemented, equipping designers with the capability to obtain optimised solutions with unprecedented speed and resolution. Second, perturbation approaches are proposed and tested, enabling designers to use topology optimisation to create diverse and efficient designs. Third, a three-field BESO method with a variable-radius filter scheme is proposed and verified, forging a new path for creating diverse designs with multiple morphologies via topology optimisation. The research outcomes hold great potential for practical applications in architecture and engineering, where diverse and competitive solutions are in high demand.</p