DR-NTU (Digital Repository of NTU)
Not a member yet
116018 research outputs found
Sort by
Alginate-prebiotic based biopolymeric encapsulation system for improving probiotics performance
No full textWith growing emphasis on the importance of gut health, the market for probiotic, prebiotic,
and synbiotic products is expanding considerably. To confer its benefits to human hosts,
there must be sufficient viable cell count of probiotics that remain upon storage and
ingestion, such that the probiotics can colonise the intestines. Encapsulation of probiotics
in an alginate matrix is a common technique employed to improve the survivability of
probiotics. Due to the growing demand for synbiotics, this study investigates the effect of
adding prebiotics on the performance of the alginate encapsulation system. Alginate
supplemented with fructo-oligosaccharides (FOS) and inulin obtained commercially, and
hydrolysed exopolysaccharides (HEPS) derived from Bacillus subtilis SOM8 isolated
from sesame oil meal, were used to encapsulate the most commercially available probiotic,
Lacticaseibacillus rhamnosus GG (LGG). The influence of the addition of prebiotics were
characterised via the encapsulation efficiency, exposure to simulated gastric and intestinal
environments for 2 h, and viability upon storage at 4 °C for 28 days. The formulation
containing pure sodium alginate showed the best performance, with superior tolerance to
acidic pH and storage conditions. It also achieved similar survival rates under bile
conditions compared to the alginate-prebiotic formulations. The pure sodium alginate
formulation was able to maintain viable cell counts above 106 CFU/g for all the exposure
conditions, making it a suitable candidate for commercial use. Formulations with FOS and
inulin incorporated also met this criterion for all the tests, achieving even higher
encapsulation efficiency compared to the pure alginate encapsulation. Conversely, HEPS
caused the storage viability to drop below this minimum requirement at 28 days and
performed worse than the pure alginate encapsulation in all the parameters. The findings
show that the addition of prebiotics generally does not improve the alginate encapsulation
system. Moreover, it was found that HEPS can be fermented by LGG, but not FOS and
inulin, suggesting that utilisable prebiotics have a negative effect on storage. Collectively,
this highlights that the concentration of utilisable prebiotics must be optimised to prevent
the shelf life of LGG from being compromised, such that there will be sufficient surviving
LGG to ferment these prebiotics to exert positive health benefits to hosts upon ingestion.Bachelor's degre
Reducing green-house gas emissions from shipping: methods and applications
No full textThe maritime shipping industry, which facilitates over 80% of global trade, contributes
approximately 3% of the world’s greenhouse gas (GHG) emissions. In response to growing
environmental pressures and international climate targets, notably those set by the International
Maritime Organization (IMO) for 2050, the sector is under increasing pressure to decarbonise. The
purpose of this report is to examine and evaluate both current and emerging strategies for reducing
GHG emissions from shipping. This paper provides a comprehensive analysis of near-term and
long-term decarbonisation strategies, including the adoption of transitional fuels such as LNG,
biofuels, and methanol; the implementation of operational measures like slow steaming and digital
route optimisation; and the exploration of future technologies such as green hydrogen, ammonia,
carbon capture and storage (CCS), and renewable propulsion systems. It also critically assesses the
regulatory, financial, and technological barriers to adoption. The report concludes that no single
solution will suffice to achieve full decarbonisation. Instead, a phased, multi-dimensional approach
is required - one that combines immediate operational improvements with long-term technological
innovation, supported by international policy alignment and stakeholder collaboration. This paper
offers practical recommendations for stakeholders to accelerate the transition toward a low-emission, sustainable maritime future.Bachelor's degre
Enhancing medical image segmentation with adaptive prompting with foundation model in semi-supervised learning
No full textSemi-supervised learning has emerged as an effective strategy to mitigate the
challenge of limited labeled data in medical image segmentation, a domain where
obtaining high-quality labels can be both expensive and time-consuming. Despite its
promise, conventional approaches often struggle with instability and inconsistency
when handling unlabeled data, resulting in unreliable segmentation outputs. To
address these issues, we present Adaptive Prompting with Segment Anything Model
(ADP-SAM), a novel framework that integrates two key innovations: adaptive
prompting and a cross-teaching mechanism.
Adaptive prompting incorporates learnable prompts that direct the model’s attention
to task-relevant features, thereby facilitating more robust feature extraction from both
labeled and unlabeled images. This targeted focus enables the model to effectively
capture subtle anatomical details critical for accurate medical segmentation. The
cross-teaching mechanism complements this process by employing two models that
iteratively refine each other’s pseudo-label predictions. By sharing and updating
these labels in a feedback loop, each model learns to correct mistakes identified by
its counterpart, reinforcing consistency and improving overall segmentation
performance.
In extensive experiments on various medical imaging datasets, ADP-SAM
demonstrates superior performance compared to existing semi-supervised
approaches. It yields more accurate and stable segmentations, particularly in
challenging scenarios characterized by complex anatomical structures or severely
limited annotated data. Furthermore, the adaptive prompting strategy helps the model
generalize more effectively across different organs and imaging modalities, while the
cross-teaching mechanism provides a powerful regularization effect that reduces
uncertainty in unlabeled samples. These findings underscore the potential of ADPSAM as a robust and scalable solution for medical image segmentation in real-world
clinical settings, offering a pathway toward more reliable, efficient, and accurate
segmentation—even when labeled data is scarce.
The main work reported in this FYP report has been accepted for presentation at the
IEEE 2025 Conference on Engineering in Medicine and Biology Society (EMBC).Bachelor's degre
Path complex neural networks for sequential process activities classification
No full textProcess mining aims to uncover, track, and enhance real-world workflows by deriving insights from event logs commonly found in modern information systems. With the growing focus on improving productivity within complex business operations, recent research has looked into developing process models to improve business performance metrics. As such, this study aims to enhance process mining from event logs by proposing a novel path-complex construction based on process mining sequential data and a path-complex-based message-passing mechanism for higher-order structural information. We adopt path-complex representations for event logs and their temporal connections developed from instance graphs. Representations are identified and optimised for 0-paths (events), 1-paths (two events in chronological order) and 2-paths (three consecutive events) to characterise intrinsic higher-order information among events. The proposed framework, Path Complex Neural Networks (PCNN), leverages the advantages of topological deep learning and obtains representations for higher-order complexes inductively. Additionally, we evaluated the results with four real-world benchmark datasets and found that PCNN outperforms existing models in analysing sequential and complex process data.Ministry of Education (MOE)Economic Development Board (EDB)Published versionThis work was supported in part by (i) the Singapore Ministry of Education Academic Research fund Tier 1 grant RG16/23, Tier 2 grants MOE-T2EP20120-0010 and MOE-T2EP20221-0003, and (ii) the Singapore Economic Development Board and HP Inc. Singapore through the Industrial Postgraduate Programme with Nanyang Technological University.
Computational study to investigate the placement of impella device to reduce the risk of aki by improving renal blood flow to mitigate hemodynamic impacts
No full textThe placement of Impella device outflow grafts plays a critical role in shaping the downstream hemodynamics and significantly influences renal perfusion and the associated risk of acute kidney injury (AKI). In this study, three distinct anatomical sites for Impella outflow graft positioning were evaluated: the descending thoracic aorta, the subclavian artery, and the distal abdominal aorta. These placements were selected based on their clinical relevance, anatomical accessibility, and potential for optimizing renal blood flow while maintaining systemic perfusion. Computational Fluid Dynamics (CFD) simulations were performed to model blood flow and assess the hemodynamic performance of each graft placement, focusing specifically on velocity, pressure, and wall shear stress (WSS) near the renal arteries. The study aimed to identify which placement minimized adverse effects on renal perfusion and reduced the risk of AKI, thus improving the overall efficacy of the Impella device in patients with cardiovascular support needs. The findings underscore the importance of selecting optimal graft placement strategies to mitigate risks associated with renal dysfunction and highlight potential pathways for future device design and clinical interventions.Bachelor's degre
Design of an inorganic passive radiative cooling paint that achieves sub-ambient temperature in tropical daytime
No full textThere are rising demands for energy towards keeping a building cool in both developed and developing countries, especially in hot and humid countries like Singapore. With increasing costs and the production of harmful emissions, more research and development has been done to find a sustainable solution to improve the energy efficiency and cooling effectiveness of buildings and structures. Passive radiative cooling paint has emerged as a suitable solution for addressing this sustainability issue, with its emphasis on improving energy efficiency and increasing green buildings as part of the Singapore Green Plan 2030. Passive radiative cooling has been observed to give positive results in temperate climates, achieving sub-ambient temperatures. However, it has been unsuccessful in replicating the results in tropical climates like Singapore, where solar irradiance and humidity are high. A passive radiative cooling method was used to achieve sub-ambient cooling in Singapore’s tropical climate, characterised by high humidity, strong solar radiation, and frequent rainfall. This project aims to develop an inorganic passive radiative cooling paint specifically tailored to the tropical climate conditions of Singapore. The research focuses on utilising inorganic resin as a key component to optimise the thermal emissivity and solar reflectivity of passive radiative cooling paint. Preliminary experiments have revealed challenges in formulating a single-layer paint with the desired thermal and mechanical properties due to poor adhesion and undesirable interactions between components. To overcome this, the project shifted to a dual-layer approach. By formulating a two-layer coating system—an inorganic sealer as the base layer and an inorganic topcoat—the project addresses issues such as compatibility between both layers. In this experiment, we document the design of a prototype inorganic passive radiative cooling paint that demonstrated good passive radiative cooling performance, with acceptable mechanical properties for paint application. The designed passive radiative cooling paint demonstrated a solar reflectance of 95.6% within the 8–13 μm atmospheric window and achieved a sub-ambient temperature of 1.7°C below ambient temperatures during peak solar irradiation conditions. The success in designing a passive radiative cooling paint using inorganic components opens new possibilities for passive cooling technology catering to Singapore’s climate. This project not only aims to achieve sub-ambient cooling during the day but also aligns with global sustainability goals by reducing energy consumption for cooling. With further research, the outcome of this project is expected to advance the knowledge in the realm of passive radiative cooling and provide practical solutions for sustainable building applications in tropical regions.Bachelor's degre
Can understand or not? Foreigners' struggle with Singlish syntax
No full textPrevious research primarily explores foreigner’s attitudes and perceptions of Singlish, with limited focus on actual intelligibility issues. These studies do not explicitly compare understanding based on participants’ language proficiency, whether in English or other language varieties, depending on the Singlish features used. This highlights a gap in understanding whether foreigners, especially those from native-English speaking backgrounds with high proficiency in Standard English, face challenges comprehending Singlish. Specifically, it raises questions about the extent of Singlish's intelligibility to Standard English speakers and foreigners unfamiliar with it, as well as identifying key features that impede understanding. This study recruited ten foreign participants from the researcher’s personal network and are all university students currently residing in native-English-speaking countries. An online questionnaire, individual interviews, and focused-group discussions was used to gather data on their understanding of Singlish. Results found that foreigners attained a mean score of 11.4 out of 14, but their responses seemed to reflect attitudes and perceptions rather than actual intelligibility. The scores alone do not clearly indicate whether foreigners found Singlish intelligible. Notably, eight out of ten participants identified sentence structure as the main cause of challenge in understanding Singlish. Findings also revealed that the top two questions which were commonly answered incorrectly solely comprised of English lexicons. In conclusion, relying solely on questionnaire scores is insufficient to gauge intelligibility, as these findings contradict the challenges reported by most participants. Thus, a future study delving into the word order aspect of sentences could be done to explore other factors and its relations to intelligibility.Bachelor's degre
Online post-fault strategy of minimum copper loss in full torque operation range for dual three-phase pmsm under single open-phase fault
No full textFault-tolerant ability of dual three-phase permanent magnet synchronous motors (PMSMs) is crucial for safetycritical applications. When a single open-phase fault occurs, the current references in conventional methods must be adjusted for avoiding the generation of torque ripple. Meanwhile, their design should balance the tradeoff between the torque operation range (TOR) and the copper loss. However, these post-fault current references still need to be re-tailored under different open-phase fault locations; otherwise, more severe system damage may occur. Aiming to improve the robustness of the control algorithm, this paper has proposed an online post-fault strategy that can achieve minimum copper loss while maximizing the TOR for dual threephase PMSMs with sinusoidal back-EMF. Unlike existing literature, the current reference in the harmonic subspace can be uniformly determined by the intermediate variable of phase shift and amplitude ratio, independent of fault scenarios. Additionally, the required fault identification is reduced from locating a specific phase to identifying one set of three-phase windings. The post-fault performance of minimum copper loss can be achieved during the full TOR. Furthermore, the implementation of proposed algorithm does not require any optimization iterative process or look-up tables. Comprehensive experimental results have verified the effectiveness of the proposed method.National Research Foundation (NRF)Submitted/Accepted versionThis work was supported by National Research Foundation (NRF) Singapore under its NRF Fellowship Grant NRF-NRFF12-2020-0003.
Public transport service operation with modular vehicles
No full textRapid population growth and increasing traffic congestion in urban areas like Singapore create significant challenges for traditional public transport systems, often resulting in overcrowding and long waiting hours during peak periods. This project explores the feasibility and potential benefits of modular transit vehicles, a flexible and adaptive transport solution designed to dynamically adjust passenger capacity through attaching or detaching modular pods. A simulation-based modeling approach was adopted, comparing the operational performance of modular vehicles with traditional buses from a passenger-focused perspective. Three modular vehicle scenarios were simulated to determine the optimal operational scheme. The results indicate that dynamically adjusting modular vehicle capacity at every bus stop provides substantial improvements in terms of service performance and passenger experience, including reduced waiting times, enhanced boarding efficiency, and greater passenger comfort compared to traditional buses. Additionally, modular vehicles demonstrated significant cost-efficiency, operating at approximately 15% lower daily costs. Despite identified practical challenges, such as public acceptance and infrastructure adaptations, the findings highlight modular vehicles as a promising and sustainable alternative for enhancing the public transport experience in Singapore. Future research directions and recommendations for pilot implementations are discussed to support successful integration into existing transport systems.Bachelor's degre
Interseismic megathrust coupling at the Osa Peninsula, Costa Rica
No full textAt the Osa Peninsula in southern Costa Rica, magnitude >7 earthquakes have been generated along the Middle American trench in 1904, 1941, and 1983 following a ∼40-year recurrence interval, suggesting a rupture may be impending. However, regional interseismic coupling remains poorly constrained, largely due to sparse observations that are likely contaminated by aliasing effects of repeating shallow slow slip events (SSEs) that occur roughly every 4 years, but were only discovered recently. These SSEs, while likely reducing megathrust coupling near the trench, may load or trigger the next rupture of the 1983 asperity. Using new continuous Global Navigation Satellite System (GNSS) data from an updated and densified regional network, we derive inter-SSE rates of deformation and invert for slip deficit and megathrust coupling along the Middle American Trench, implementing block modeling to correct for the motion of the Panama microplate. We invert for slow slip and remove a time-averaged estimate of cumulative slow slip from our models. Our results indicate that the region of highest inter-SSE coupling (>0.8) corresponds with the spatial extent of SSE slip. We also find that SSEs are sufficient to release nearly all the elastic strain accumulated over their 4-year recurrence interval in localized regions. Accounting for this, in the region immediately downdip of the slow slip patch—the same region thought to have ruptured in the 1983 Mw 7.4 event—we estimate an interseismic coupling ratio of ∼0.5–0.7 corresponding to ∼1.75–2 m of accumulated slip deficit since 1983, sufficient to generate a similar magnitude rupture in the future.Published versio