56016 research outputs found
Sort by
Detection of incomplete atypical femoral fracture on anteroposterior radiographs via explainable artificial intelligence
No full textOne of the key aspects of the diagnosis and treatment of atypical femoral fractures is the early detection of incomplete fractures and the prevention of their progression to complete fractures. However, an incomplete atypical femoral fracture can be misdiagnosed as a normal lesion by both primary care physicians and orthopedic surgeons; expert consultation is needed for accurate diagnosis. To overcome this limitation, we developed a transfer learning-based ensemble model to detect and localize fractures. A total of 1050 radiographs, including 100 incomplete fractures, were preprocessed by applying a Sobel filter. Six models (EfficientNet B5, B6, B7, DenseNet 121, MobileNet V1, and V2) were selected for transfer learning. We then composed two ensemble models; the first was based on the three models having the highest accuracy, and the second was based on the five models having the highest accuracy. The area under the curve (AUC) of the case that used the three most accurate models was the highest at 0.998. This study demonstrates that an ensemble of transfer-learning-based models can accurately classify and detect fractures, even in an imbalanced dataset. This artificial intelligence (AI)-assisted diagnostic application could support decision-making and reduce the workload of clinicians with its high speed and accuracy
A Bipartite Graph Neural Network Approach for Scalable Beamforming Optimization
No full textDeep learning (DL) techniques have been intensively studied for the optimization of multi-user multiple-input single-output (MU-MISO) downlink systems owing to the capability of handling nonconvex formulations. However, the fixed computation structure of existing deep neural networks (DNNs) lacks flexibility with respect to the system size, i.e., the number of antennas or users. This paper develops a bipartite graph neural network (BGNN) framework, a scalable DL solution designed for multi-antenna beamforming optimization. The MU-MISO system is first characterized by a bipartite graph where two disjoint vertex sets, each of which consists of transmit antennas and users, are connected via pairwise edges. These vertex interconnection states are modeled by channel fading coefficients. Thus, a generic beamforming optimization process is interpreted as a computation task over a weighted bipartite graph. This approach partitions the beamforming optimization procedure into multiple suboperations dedicated to individual antenna vertices and user vertices. Separated vertex operations lead to scalable beamforming calculations that are invariant to the system size. The vertex operations are realized by a group of DNN modules that collectively form the BGNN architecture. Identical DNNs are reused at all antennas and users so that the resultant learning structure becomes flexible to the network size. Component DNNs of the BGNN are trained jointly over numerous MU-MISO configurations with randomly varying network sizes. As a result, the trained BGNN can be universally applied to arbitrary MU-MISO systems. Numerical results validate the advantages of the BGNN framework over conventional methods
Adaptive goal-switching and planning for sequential decision-making in dynamic environments
No full text
High Performance Pd catalysts Modified with MoOx for LOHC
No full textLiquid organic hydrogen carrier systems (LOHCs) are an efficient approach for storing hydrogen in organic molecules. Catalytic hydrogenation makes it possible to safely store and efficiently store and transport hydrogen. By dehydrogenation, hydrogen can be released. However, high temperatures are required for the dehydrogenation reaction, which reduces process efficiency and decomposes organic substances as a side reaction, limiting reuse. In this study, a low-temperature dehydrogenation reaction was performed using 1-methylindole (NMID)/Octahydro-1-methylindole (8H-NMID) as a heterocyclic LOHC compound containing a nitrogen atom in a ring structure. As a strategy to enhance the Pd-catalyzed dehydrogenation performance, MoOx was added to control the electronic structure of Pd. The addition of a small amount of MoOx to the Pd/Al2O3catalyst improved the Pd dispersion due to hydrogen spillover, resulting in a 1.52-fold increase in dehydrogenation activity compared to the original Pd/Al2O3
Chemical and physical characteristics of hybrid alkaline cement composite after laser interaction
No full textThis study investigated the resistance of a hybrid alkaline cement composite (HACC) to laser cutting. HACC, synthesized using geopolymer and cement paste, can be useful for rapid construction or repair due to its faster setting and higher compressive strength than cement paste. Also, laser cutting is a recently popular fabrication method because of low dust and noise production as well as fast speed. Although cement paste was proven to enhance the compressive strength of the geopolymer, its effect on the resistance against laser cutting has not been studied. Laser-cutting experiments were performed in this study on HACCs with five different cement paste contents. Thermogravimetric analysis, X-ray diffraction, and scanning electron microscope with energy dispersive microscope were conducted to investigate the structural changes after the laser treatment. Lastly, the microhardness of HACC was measured to reveal the correlation between physical hardness and thermal resistance. The results demonstrate that sodium aluminosilicate hydrate gels can resist laser cutting more than calcium (alkali) aluminosilicate hydrate and calcium silicate hydrate gels because of their chemical and physical characteristics
?????? ????????? ????????? ?????? ??????????????? ??????????????? ?????? ?????????????????? ???????????? ?????? ????????? ??? ????????? ?????????
No full textDepartment of Urban and Environmental Engineering (Environmental Science and Engineering)Gas hydrate formation in oil and gas pipelines presents substantial challenges to industry operations, with significant implications for safety and efficiency. Traditional mitigation methods include the application of thermodynamic hydrate inhibitors (THIs), kinetic hydrate inhibitors (KHIs), and anti-agglomerants (AAs), yet the quest for effective and environmentally benign inhibitors remains an active area of research. This dissertation contributes novel insights to this field through the exploration and development of eco-friendly hydrate inhibitors, utilizing an integrative approach of experimental and computational chemistry.
The first part of this thesis (Chapters 3.1 and 3.2) focuses on the development and evaluation of hydrophilic monomeric substances as potential gas hydrate inhibitors and synergists. The inhibitory effects on methane (CH4) and carbon dioxide (CO2) hydrates were investigated through a combination of experimental and computational methods, yielding promising results.
In the second part of this thesis (Chapters 4.1, 4.2, and 4.3), the potential of oligopeptide-based inhibitors (dipeptides and tripeptide), waterborne polyurethanes (WPUs), and biodegradable oligopeptides as efficient kinetic hydrate inhibitors were explored. These novel inhibitors were evaluated for their effectiveness against CH4 hydrate and a mixed CH4 (90%) and propane (C3H8, 10%) hydrate system, as well as CO2 hydrate in the case of biodegradable oligopeptides, contributing to the growing body of research surrounding gas hydrate inhibition.
Finally, the concluding chapter (Chapter 5) outlines potential future research directions, encompassing deeper exploration into the mechanisms of action of these inhibitors, further development of computational models, and the potential application of these inhibitors in the context of carbon capture, utilization, and storage (CCUS), particularly in the transmission and storage stages under deep-sea conditions.
The findings presented in this dissertation contribute significantly to the field of gas hydrate research, particularly in the context of developing new eco-friendly and efficient inhibitors. Furthermore, the study's results have important implications for the oil and gas industry, assisting in the ongoing efforts to design and implement more efficient and environmentally friendly strategies for the prevention of gas hydrate formation.clos
Understanding of organic aerosol processes in South Korea based on observations and chemical transport model outputs
No full textDepartment of Urban and Environmental Engineering (Environmental Science and Engineering)clos
Neural representation of mechanical tactile information processing in the somatosensory cortex
No full textDepartment of Biomedical Engineering (Human Factors Engineering)clos
Study on Highly Efficient Organic and Perovskite Solar Cells with Operational Stability
No full textDepartment of Materials Science and Engineeringclos