A5032667101- AI in cancer detection
- Theoretical and Computational Physics
- Medical Image Segmentation Techniques
- Radiomics and Machine Learning in Medical Imaging
- Cell Image Analysis Techniques
- Adversarial Robustness in Machine Learning
- Anomaly Detection Techniques and Applications
- Magnetic properties of thin films
- Image Retrieval and Classification Techniques
- Corneal surgery and disorders
- Image Processing Techniques and Applications
- Glaucoma and retinal disorders
- Medical Imaging and Analysis
- Generative Adversarial Networks and Image Synthesis
- Advanced Vision and Imaging
- Neural Networks and Applications
- Digital Imaging for Blood Diseases
- Gaussian Processes and Bayesian Inference
- Cardiac Imaging and Diagnostics
- Digital Media Forensic Detection
- Advanced Malware Detection Techniques
- Quantum many-body systems
- Cancer Genomics and Diagnostics
- Model Reduction and Neural Networks
- Ocular Surface and Contact Lens
Agency for Science, Technology and Research
2015-2025
Bioinformatics Institute
2016-2025
National University of Singapore
2006-2024
Singapore Eye Research Institute
2018-2024
Image and Pervasive Access Laboratory
2016-2024
Singapore Institute for Clinical Sciences
2021-2022
Singapore General Hospital
2018
Tokyo Metropolitan University
2005-2017
Institute for Infocomm Research
2017
Sorbonne Université
2016
Currently, molecular markers are not used when determining the prognosis and treatment strategy for patients with hepatocellular carcinoma (HCC). In present study, we proposed that identification of common pro-oncogenic pathways in primary tumors (PT) adjacent non-malignant tissues (AT) typically to predict HCC patient risks may result biomarker discovery. We examined genome-wide mRNA expression profiles paired PT AT samples from 321 patients. The workflow integrated differentially expressed...
Spatial omics data are clustered to define both cell types and tissue domains. We present Building Aggregates with a Neighborhood Kernel Yardstick (BANKSY), an algorithm that unifies these two spatial clustering problems by embedding cells in product space of their own the local neighborhood transcriptome, representing state microenvironment, respectively. BANKSY's feature augmentation strategy improved performance on tasks when tested diverse RNA (imaging, sequencing) protein (imaging)...
Anomaly detection is a classical problem where the aim to detect anomalous data that do not belong normal distribution. Current state-of-the-art methods for anomaly on complex high-dimensional are based generative adversarial network (GAN). However, traditional GAN loss directly aligned with objective: it encourages distribution of generated samples overlap real and so resulting discriminator ineffective as an detector. In this paper, we propose modifications such lie at boundaries With our...
Purpose.: To compare anterior segment parameters, assessed by optical coherence tomography (ASOCT), in subjects categorized as primary angle closure suspect (PACS), (PAC), glaucoma (PACG), and previous acute PAC (APAC); to identify factors associated with APAC. Methods.: This was a prospective ASOCT study of 425 (176 PACS, 66 PAC, 125 PACG, 58 APAC). Customized software used measure including opening distance (AOD750), trabecular–iris space area (TISA750), chamber depth, width, volume (ACD,...
With the recent developments in machine learning, Carrasquilla and Melko have proposed a paradigm that is complementary to conventional approach for study of spin models. As an alternative investigating thermal average macroscopic physical quantities, they used configurations classification disordered ordered phases phase transition through learning. We extend generalize this method. focus on configuration long-range correlation function instead itself, which enables us provide same...
Different types of Convolutional Neural Networks (CNNs) have been applied to detect cancerous lung nodules from computed tomography (CT) scans. However, the size a nodule is very diverse and can range anywhere between 3 30 millimeters. The high variation sizes makes classifying them difficult challenging task. In this study, we propose novel CNN architecture called Gated-Dilated (GD) networks classify as malignant or benign. Unlike previous studies, GD network uses multiple dilated...
Abstract The study of neuronal morphology and neurite outgrowth has been enhanced by the combination imaging informatics high content screening, in which thousands images are acquired using robotic fluorescent microscopy. To understand process context neuroregeneration, we used mouse neuroblastoma N1E115 as our model cell. Six‐thousand cellular four different culture conditions were with two‐channel widefield We developed a software package called NeuronCyto . It is fully automatic solution...
Purpose Coronary CT angiography (CCTA) is well established for the diagnostic evaluation and prognostication of coronary artery disease (CAD). The growing burden CAD in Asia emergence novel CT-based risk markers highlight need an automated platform that integrates patient data with CCTA findings to provide tailored, accurate cardiovascular assessments. This study aims develop artificial intelligence (AI)-driven assessment using Singapore’s multiethnic population. We will conduct a hybrid...
We propose a general method of using the Fokker-Planck equation (FPE) to link Monte-Carlo (MC) and Langevin micromagnetic schemes. derive drift disusion FPE terms corresponding MC show that it is analytically equivalent stochastic Landau-Lifshitz-Gilbert (LLG) Langevin-based micromagnetics. Subsequent results such as time quantification factor for Metropolis can be rigorously derived from this mapping equivalence. The validity shown by close numerical convergence between LLG case single...
Analyzing cellular morphologies on a cell-by-cell basis is vital for drug discovery, cell biology, and many other biological studies. Interactions between cells in their culture environments cause to touch each acquired microscopy images. Because of this phenomenon, segmentation challenging task, especially when the are similar brightness highly variable shapes. The concept topological dependence maximum common boundary (MCB) algorithm presented our previous work (Yu et al., Cytometry Part A...
Infrared (IR) meibography is an imaging technique to capture the Meibomian glands in eyelids. These ocular surface structures are responsible for producing lipid layer of tear film which helps reduce evaporation. In a normal healthy eye, have similar morphological features terms spatial width, in-plane elongation, length. On other hand, eyes with gland dysfunction show visible structural irregularities that help diagnosis and prognosis disease. However, currently there no universally...
Nuclear pleomorphic patterns are essential for Fuhrman grading of clear cell renal carcinoma (ccRCC). Manual observation histopathologic slides may lead to subjective and inconsistent assessment between pathologists. An automated, image-based system that classifies ccRCC by quantifying nuclear in an objective consistent interpretable fashion can aid pathologists assessment.In the current study, tissue 59 patients with who underwent surgery at Singapore General Hospital were assembled...
Abstract Background The visual assessment and severity grading of acne vulgaris by physicians can be subjective, resulting in inter‐ intra‐observer variability. Objective To develop validate an algorithm for the automated calculation Investigator's Global Assessment (IGA) scale, to standardize outcome measurements. Materials Methods A total 472 photographs (retrieved 01/01/2004‐04/08/2017) frontal view from 416 patients were used training testing. Photographs labeled according IGA scale...
Abstract Aims To identify differences in CT-derived perivascular (PVAT) and epicardial adipose tissue (EAT) characteristics that may indicate inflammatory status between post-treatment acute myocardial infarction (AMI) stable coronary artery disease (CAD) patients. Methods Results A cohort of 205 post-AMI patients (age 59.8±9.2, 92.2% male) was propensity-matched with CAD 60.5±10.0, 90.2% male). Coronary CT angiography non-contrast scans were performed to assess PVAT mean attenuation across...
The diagnostic evaluation of ring-enhancing brain lesions (REBLs) is challenging, especially in immunocompromised patients. We conducted a retrospective study to describe the clinicodemographic and radiological features among patients presenting with REBLs tertiary referral center. Radiological reports all who underwent computed tomography or magnetic resonance (MR) imaging between 1 November 2013 31 October 2017 were filtered for terms indicative REBLs. Infectious diseases physicians...
Improving in silico compound-protein interaction (CPI) predictability is critical for productive drug discovery. Current deep learning approaches largely rely on end-to-end models trained limited labeled CPI data, overlooking preexisting, specialized input representations. We present Ligand Extra trees-Accelerated Docking (LEAD), a virtual screening framework that accelerates docking by integrating rapid first-pass prediction via ET-Screen. Unlike models, ET-Screen uses seven distinct...
Glandular structural features are important for the tumor pathologist in assessment of cancer malignancy prostate tissue slides. The varying shapes and sizes glands combined with tedious manual observation task can result inaccurate assessment. There also discrepancies low-level agreement among pathologists, especially cases Gleason pattern 3 4 adenocarcinoma. An automated gland segmentation system highlight various glandular structures further analysis by pathologist. These objective...