48080 research outputs found
Sort by
Additive Manufacturing Technology in Fabricating Tissue-Equivalent Materials and Dosimetry Equipment for Advanced Radiotherapy Medical Physics Application
No full textRadiation therapy technology advances greatly in the past years because of the development in engineering and computing. This results in an improved dose delivery while decreasing the possibility of treatment complications. Given the increasing complexity and variety of treatment machines and techniques, an improvement in the quality assurance equipment should be considered. Conventional phantom equipment used in radiation therapy are commonly limited to simplified designs and homogenous characteristic. With the medical trend transitioning into personalized treatment, it is important to consider patient-specific quality assurance and customizable dosimetry equipment.Three-dimensional (3D) printing or additive manufacturing (AM) is an advanced technology used to manufacture products based on computer-aided design models using a unit material that allows customizable, and patient-specific fabrication. Due to its advantages compared to traditional manufacturing such as design flexibility and multimaterial manufacturing, it has become popular in different industries and fields including medical physics. Bioprinting is another innovative 3D printing technique that uses bioengineering methods to fabricate 3D biological constructs using biomaterials and cells. It can create tissue and tumor constructs that closely mimic some of the structural characteristics and physiological responses of a real tissue and tumor.This study characterized the radiological properties and other features of various additive manufacturing technologies and materials such as polymers, ceramics, metals, and hydrogels. Various 3D printed phantoms were fabricated such as a Computed Tomography (CT)-Electron Density phantom, a pediatric head phantom, a rat phantom, and an adult head phantom. Radiological imaging and therapy modalities were used in characterizing different AM materials and phantoms such as clinical CT, synchrotron CT, clinical linear accelerator (Linac), and synchrotron broadbeam and microbeam radiotherapy. Furthermore, the applications of 3D bioprinting in the fabrication of 3D in-vitro brain cancer tumor (glioma) constructs were investigated for synchrotron microbeam radiation therapy experiments. A Gelatin Methacryloyl (GelMA) and a compact 3D REDI bioprinter were used in fabricating bioprinted brain tumors.The CT number and attenuation measurements were utilized for tissue-equivalence characterization and printing evaluation. In addition, treatment beams were delivered to the 3D printed materials and phantoms to assess its capabilities as dosimetry equipment. Different dosimeters were used which includes an ionization chamber, microDiamond detector, radiochromic films, and a silicon single strip detector. Furthermore, a validated Geant4 Monte Carlo simulation and a treatment planning system were used to verify the measured doses with calculations for synchrotron radiotherapy and Linac radiotherapy, respectively. For bioprinting experiments, a cell viability assay and a fluorescence imaging method were performed to assess the biological effect of synchrotron radiotherapy delivery. The response of 3D bioprinted tumors was compared to a cell monolayer culture and a 3D spheroid culture.The results of this study provided further characterization data, 3D printed phantom fabrication protocols, and 3D bioprinting methods which are important in establishing guidelines in using 3D printing technology for advanced medical radiation physics and radiobiology applications. Overall, this research project demonstrated the capabilities of AM in fabricating customizable, and radiologically tissue equivalent medical physics dosimetry equipment.</p
The Efficacy and Mechanisms of Synergistic Multi-Modal Combinational Radiotherapies in the Treatment of Brain Cancer
No full textThe abstract for this item has not been populated.</p
Hybrid LDA-CNN Framework for Robust End-to-End Myoelectric Hand Gesture Recognition Under Dynamic Conditions
No full textGesture recognition based on conventional machine learning is the main control approach for advanced prosthetic hand systems. Its primary limitation is the need for feature extraction, which must meet real-time control requirements. On the other hand, deep learning models could potentially overfit when trained on small datasets. For these reasons, we propose a hybrid Linear Discriminant Analysis–convolutional neural network (LDA-CNN) framework to improve the gesture recognition performance of sEMG-based prosthetic hand control systems. Within this framework, 1D-CNN filters are trained to generate latent representation that closely approximates Fisher’s (LDA’s) discriminant subspace, constructed from handcrafted features. Under the train-one-test-all evaluation scheme, our proposed hybrid framework consistently outperformed the 1D-CNN trained with cross-entropy loss only, showing improvements from 4% to 11% across two public datasets featuring hand gestures recorded under various limb positions and arm muscle contraction levels. Furthermore, our framework exhibited advantages in terms of induced spectral regularization, which led to a state-of-the-art recognition error of 22.79% with the extended 23 feature set when tested on the multi-limb position dataset. The main novelty of our hybrid framework is that it decouples feature extraction in regard to the inference time, enabling the future incorporation of a more extensive set of features, while keeping the inference computation time minimal.</p
Real-time lithology identification while drilling based on drill cuttings image analysis with ensemble learning
No full textAccurate lithology identification through geological exploration is crucial for hazard risk management during deep underground operations. Artificial intelligence has advanced in image recognition but using it to analyze underground drill cuttings for accurate lithology remains challenging. Issues include imprecise sampling control, harsh environments, and inconsistent image acquisition procedures, all leading to poor image quality. To address these issues, a lithology identification while drilling method was proposed. A cuttings sampling, testing, and transporting system was developed and deeply integrated with the drilling rig, achieving automation in cuttings sampling operations while standardizing the timing, procedures, and environment for sampling. A cuttings image preprocessing method was proposed, which meets the requirements of machine learning for image dimensions while enabling the automatic calculation of the proportions of different lithological particles. This is highly significant for accurately determining stratigraphic interfaces. An ensemble learning method was applied to enhance the identification accuracy. Underground trials were conducted at a coal mine in Huainan, China, involving the construction of four boreholes and the acquisition of more than a thousand cuttings images. During the trials, the system cooperated with the drilling rig to realize the accurate identification of lithology information during drilling, with an accuracy of 97.42% and an average processing time of less than 0.11 s per image. The results showed that the proposed lithology identification method can accurately obtain formation lithology in real time during drilling. This study guides drilling operations, ensuring target area coverage, effective hazard management, and supporting unmanned drilling technology development.</p
Experimental study on the deflagration to detonation transition process of gas explosions in engineering-scale pipelines
No full textTo investigate the deflagration to detonation transition (DDT) process of methane gas in large-scale, unobstructed pipelines with weak ignition sources, experimental studies were conducted using pipelines with diameters of 500 mm and 700 mm, and lengths of 66.5 m and 93.1 m, respectively. The experiments were performed with methane concentrations of 7.5 %, 8.5 %, and 9.5 %. The explosion pressure and flame propagation speed during the DDT process were measured and theoretically calculated for different methane concentrations. The experimental and theoretical analysis results indicate that at concentrations of 7.5 %, 8.5 %, and 9.5 %, the gas remained in a quasi-detonation state. For the same pipeline diameter, the DDT process occurs earlier as the methane concentration approaches the stoichiometric value. For the same methane concentration, larger pipeline diameters result in later DDT occurrences. The DDT positions for DN500 mm and DN700 mm pipelines were found to be 53 m, 51.5 m, and 46 m, and 63.5 m, 53.5 m, and 48.5 m, respectively, for the three concentrations. A fitting analysis provided a coupling relationship between pipeline diameter, methane concentration, and DDT distance. These findings offer theoretical support for predicting and controlling gas explosions in long-distance methane transport pipelines, contributing valuable insights for practical engineering applications.</p
Development of Bismuth-Based Oxide Photocatalysts for Simultaneous Photo(electro)catalytic Reactions
No full textToxic waste and the growing demand for renewable energy have driven research into sustainable solutions. Advanced oxidation processes (AOPs) are gaining attention for their reduced environmental impact compared to conventional technologies. Photoelectrocatalysis and photocatalysis are particularly studied for their low operational costs, minimal energy needs, and reduced by-products. Bismuth-based oxide photocatalysts are among the semiconductors that have received more attention due to visible-light harvesting capabilities and stability in driving chemical reactions. However, issues like poor charge transfer, high recombination rates, and wide band gaps limit their efficiency. The development of efficient and sustainable photocatalysts is crucial for addressing environmental and energy challenges. This thesis explores the development of bismuth-based oxide photo(electro)catalyst for organic substant synthesis and the simultaneous degradation of pollutant and hydrogen generation under visible light irradiation. The synthesized bismuth-based photocatalysts were systematically studied to enhance charge separation and improve overall efficiency.Firstly, the study focuses on BiVO4 with tuneable (110)/(010) facet ratios, synthesized via ethanolamine (ETA)-assisted hydrothermal synthesis for photocatalytic oxidative benzylamine coupling under visible light. The (110)-dominant BiVO4 demonstrated superior photocatalytic activity at room temperature due to its highly exposed oxidative (110) facets and enhanced charge generation and migration. The catalyst achieved >85% selectivity in converting benzylamine derivatives to imines, showcasing its broad substrate applicability. Mechanistic insights suggest a superoxide radical-assisted (•O2−) oxidative coupling pathway. This work underscores ETA's role in tailoring BiVO4 facets, directly impacting oxidative amine coupling, and offers a strategy for designing high-performance catalysts.The second part of the study explores a novel Bi2MoO6-based photoelectrode (BMO/GO/CC), synthesized via solvothermal methods and supported on carbon cloth with graphene oxide (GO) as a binder. Physical and chemical characterisations confirmed that this composite exhibits the superior photoelectrocatalytic (PEC) performance for malachite green degradation compared to individual photocatalytic (PC) or electrocatalytic (EC) processes. The enhanced PEC performance is attributed to GO's role in improving conductivity and charge migration, reducing electron-hole recombination, and ensuring uniform catalyst deposition for better stability and recyclability. Mechanistic studies confirm that Bi2MoO6 facilitates the generation of various reactive oxygen species, driving efficient dye mineralization.The final part of the thesis studies the development of p-CuBi2O4/n-Bi2MoO6 heterostructured photoanode for the PEC degradation of Rhodamine B (RhB) and simultaneous hydrogen (H2) production in alkaline electrolyte and natural seawater. The heterostructured photoanode achieved higher RhB degradation and higher H2 production than pristine CuBi2O4 and Bi2MoO6. The performance enhancement was attributed to the internal electric field, improving charge separation, electron migration, and suppressing recombination. In seawater, RhB degradation was further improved due to its high ionic conductivity, although H2 production was limited due to competitive chloride oxidation and Mg(OH)2/Ca(OH)2 deposition at the Pt cathode. To the best of our knowledge, this study is among the first to demonstrate the PEC potential of the CuBi2O4/Bi2MoO6 heterostructure in seawater. Mechanistic investigations revealed that direct hole oxidation played a dominant role in RhB degradation, while RhB presence enhanced H2 production.This thesis demonstrates the potential of bismuth-based oxide photocatalysts for enhancing photo(electro)catalytic performance by various development strategies for environmental remediation and renewable energy production. The findings provide valuable insights for designing high-performance photocatalysts and advancing sustainable technologies.</p
Adapting Generative Large Language Models for Information Extraction from Unstructured Electronic Health Records in Residential Aged Care: A Comparative Analysis of Training Approaches
No full textInformation extraction (IE) of unstructured electronic health records is challenging due to the semantic complexity of textual data. Generative large language models (LLMs) offer promising solutions to address this challenge. However, identifying the best training methods to adapt LLMs for IE in residential aged care settings remains underexplored. This research addresses this challenge by evaluating the effects of zero-shot and few-shot learning, both with and without parameter-efficient fine-tuning (PEFT) and retrieval-augmented generation (RAG) using Llama 3.1-8B. The study performed named entity recognition (NER) to nursing notes from Australian aged care facilities (RACFs), focusing on agitation in dementia and malnutrition risk factors. Performance evaluation includes accuracy, macro-averaged precision, recall, and F1 score. We used non-parametric statistical methods to compare if the differences were statistically significant. Results show that zero-shot and few-shot learning, whether combined with PEFT or RAG, achieve comparable performance across the clinical domains when the same prompting template is used. Few-shot learning significantly outperforms zero-shot learning when neither PEFT nor RAG is applied. Notably, PEFT significantly improves model performance in both zero-shot and few-shot learning; however, RAG significantly improves performance only in few-shot learning. After PEFT, the performance of zero-shot learning reaches a comparable level with few-shot learning. However, few-shot learning with RAG significantly outperforms zero-shot learning with RAG. We also found a similar level of performance between few-shot learning with RAG and zero-shot learning with PEFT. These findings provide valuable insights for researchers, practitioners, and stakeholders to optimize the use of generative LLMs in clinical IE.</p
High Throughput Analysis of Single Cell - Material Interactions: Toward Novel Applications in Cancer Diagnostics and Treatments
No full textThe abstract for this item has not been populated.</p
Ongo (sounding/hearing/feeling), mate (death), fonua (people and place) and tā-vā (time-space): new foundations for Tongan music composition, performance and sound art
No full textThis thesis identifies and examines key Indigenous aesthetic concepts and practices of faiva fasi (Tongan music) to establish new theoretical and methodological foundations for contemporary composition, performance and sound art. It asks: ‘what are the key Indigenous aesthetic concepts and practices of faiva fasi, and how can they be employed as Indigenousled theoretical and methodological approaches to contemporary composition, performance and sound art?’. These aims are led by the global project of decolonisation, Indigenous selfdetermination and social justice. The research focuses on ongo fa'ahikehe (sound of the dead or ancestral sound) and tu'akautā (to beat outside, behind or beyond the beat) to acknowledge Tongan ontologies and epistemologies of ongo (sound) in relation to mate (death), fonua (people and place) and tā-vā (time-space). The thesis employs the Indigenous Tongan Tā-Vā (Time-Space) Philosophy of Art, and a combination of talanoa (talking critically yet harmoniously) with specialists in Tongan music, arts and culture, and music analysis of sound archives and museum collections using case studies of the tangilaulau (crying and reciting poetry), fakatangi (chanting in the style of crying) and fangufangu (bamboo noseflute). The research is accompanied by a creative exhibition, which is grounded in the three Tongan arts genres: faiva (performance arts), tufunga (material arts) and nimamea'a (fine arts). It includes five new works in composition, performance and sound art—as well as photography, video, sculpture and print—which are also examined in contribution to the research inquiry. The written and creative components of the thesis employ a method of ‘resounding’ (in the manner of ‘re-reading’ and ‘re-writing’ as post- and decolonial methodologies) existing archives on Tongan music by walking forward into the past and backwards into the future as a Moana Oceanian framework. This includes a critique of existing scholarship in which key aesthetic concepts and practices of Tongan music are disregarded or misinterpreted. It also shines a light on the colonial legacies of Eurocentrism and heteropatriarchy on Tongan music and adjacent art forms.</p
Comparison of RayStation Collapsed Cone Convolution and Monte Carlo Algorithms in Calculation of Treatment Dose to Base-of-Skull Metastases
No full textFor radiotherapy treatment planning, type-B algorithms such as collapsed cone convolution have been the clinical standard due to their fast calculation compared to type-C algorithms such as Monte Carlo. As computational speeds increase, type-C algorithms have become increasingly clinically viable, and the advent of software such as RayStation that uses Monte Carlo means that it’s important to create a base of research to understand how these types of algorithms calculate dose differently. Although disadvantages of type-B algorithms are known, one being that they become inaccurate when considering the presence of inhomogeneous tissue such as bone, research is needed for clearer evidence as to which type of algorithm will be best in a clinical setting.In this thesis, RayStation’s Monte Carlo (type-C) and collapsed cone (type-B) algorithms were used in the calculation of dose in a treatment targeting a tumour partially covering the base of the skull in the occipital bone of a patient. Collapsed cone estimates dose analytically and transports particles rectilinearly, while Monte Carlo uses pseudorandom number generation to simulate particle interactions at given intervals. For each algorithm, three beam plans were considered: 4-beam IMRT, 5-beam IMRT and 2-arc VMAT. Additionally, there was alternation between using material override for different regions of the patient head such as the compact bone, soft bone and airways, and no material override. What was observed was that, for IMRT plans, Monte Carlo generally provided a slightly higher estimation of dose to the planning target volume compared to collapsed cone, while also more accurately characterising the reduced dose to bone. In comparison, VMAT plans showed negligible differences between the use of algorithms. This suggests that the collapsed cone algorithm does not describe dose with inhomogeneous tissue as effectively, though this difference becomes more trivial with VMAT plan. Further research can better clarify this difference with the use of a golden standard model such as EGSnrc or Geant4’s Monte Carlo algorithms.</p