A5016291944- Radiomics and Machine Learning in Medical Imaging
- AI in cancer detection
- Biomedical Text Mining and Ontologies
- Lung Cancer Diagnosis and Treatment
- Colorectal Cancer Screening and Detection
- Topic Modeling
- Machine Learning in Healthcare
- Global Cancer Incidence and Screening
- Medical Imaging Techniques and Applications
- Advanced X-ray and CT Imaging
- Prostate Cancer Diagnosis and Treatment
- Scientific Computing and Data Management
- Semantic Web and Ontologies
- Prostate Cancer Treatment and Research
- Cardiac electrophysiology and arrhythmias
- Artificial Intelligence in Healthcare and Education
- Data Visualization and Analytics
- Digital Radiography and Breast Imaging
- COVID-19 diagnosis using AI
- Reinforcement Learning in Robotics
- MRI in cancer diagnosis
- Adversarial Robustness in Machine Learning
- Cloud Computing and Resource Management
- Cardiac pacing and defibrillation studies
- Advanced Image Processing Techniques
University of California, Los Angeles
2016-2025
Twitter (United States)
2025
Samueli Institute
2020-2024
Kansas State University
2010-2024
Artificial Intelligence in Medicine (Canada)
2024
National Taiwan Ocean University
2014-2023
Vanderbilt University Medical Center
2023
Johns Hopkins University
2023
University of Auckland
2019-2022
American Medical Informatics Association
2022
Computed tomography (CT) is a widely used screening and diagnostic tool that allows clinicians to obtain high-resolution, volumetric image of internal structures in non-invasive manner. Increasingly, efforts have been made improve the quality low-dose CT (LDCT) reduce cumulative radiation exposure patients undergoing routine exams. The resurgence deep learning has yielded new approach for noise reduction by training multi-layer convolutional neural networks (CNN) map normal-dose images....
Risk-stratified screening (RSS) scheduling may facilitate more effective use of same-day diagnostic testing for potentially abnormal mammograms, thereby reducing the need follow-up appointments ("recall"). Our simulation study assessed potential impact RSS on patients recommended diagnostics. We used a discrete event to model workflow at high-volume breast imaging center, incorporating artificial intelligence (AI)-triaged workups after mammograms. The design sequences in daily schedule using...
Given the costs of delivering care for men with prostate cancer remain poorly described, this article reports results time-driven activity-based costing (TDABC) competing treatments low-risk cancer. Process maps were developed each phase from initial urologic visit through 12 years follow-up robotic-assisted laparoscopic prostatectomy (RALP), cryotherapy, high-dose rate (HDR) and low-dose (LDR) brachytherapy, intensity-modulated radiation therapy (IMRT), stereotactic body (SBRT), active...
Plasma cell-free DNA (cfDNA) is a noninvasive biomarker for cell death of all organs. Deciphering the tissue origin cfDNA can reveal abnormal because diseases, which has great clinical potential in disease detection and monitoring. Despite promise, sensitive accurate quantification tissue-derived remains challenging to existing methods due limited characterization methylation reliance on unsupervised methods. To fully exploit cfDNA, here we present one largest comprehensive high-resolution...
Due to the increasingly data-intensive clinical environment, physicians now have unprecedented access detailed information from a multitude of sources. However, applying this guide medical decisions for specific patient case remains challenging. One issue is related presenting practitioner: displaying large (irrelevant) amount often leads overload. Next-generation interfaces electronic health record (EHR) should not only make data easily searchable and accessible, but also synthesize...
Corn (Zea mays L.) is one of the most sensitive crops to planting pattern and early-season uniformity. The common method determine number plants by visual inspection on ground but this field activity becomes time-consuming, labor-intensive, biased, may lead less profitable decisions farmers. objective study was develop a reliable, timely, unbiased for counting corn based ultra-high-resolution imagery acquired from unmanned aerial systems (UAS) automatically scout fields applied real...
The objective of this study was to explore radiomics features from longitudinal diffusion-weighted MRIs (DWIs) for pathologic treatment effect prediction in patients with localized soft tissue sarcoma (STS) undergoing hypofractionated preoperative radiotherapy (RT). Thirty STS treated RT were recruited imaging study. DWIs acquired at three time points using a 0.35 T MRI-guided system. Treatment score (TES) obtained the post-surgery pathology as surrogate outcome. Patients divided into two...
With a shortfall in fellowship-trained breast radiologists, mammography screening programs are looking toward artificial intelligence (AI) to increase efficiency and diagnostic accuracy. External validation studies provide an initial assessment of how promising AI algorithms perform different practice settings.To externally validate ensemble deep-learning model using data from high-volume, distributed program academic health system with diverse patient population.In this study, learning...
Globally, lung cancer is responsible for nearly one in five deaths. The National Lung Screening Trial (NLST) demonstrated the efficacy of low-dose computed tomography (LDCT) to identify early-stage disease, setting basis widespread implementation screening programs. However, specificity LDCT suboptimal, with a significant false positive rate. Representing this imaging-based process as sequential decision making problem, we combined multiple machine learning-based methods learn...
Quantitative 3-T multiparametric MRI parameters correlated with and helped predict the presence of aggressive large cribriform pattern intraductal carcinoma prostate cancer at whole-mount histopathology.
The study aims to systematically characterize the effect of CT parameter variations on images and lung radiomic deep features, evaluate ability different image harmonization methods mitigate observed variations. A retrospective in-house sinogram dataset 100 low-dose chest scans was reconstructed by varying radiation dose (100%, 25%, 10%) reconstruction kernels (smooth, medium, sharp). set processing, convolutional neural network (CNNs), generative adversarial network-based (GANs) were...
This paper presents a framework for detection and classification of cyber threat indicators in the Twitter stream. Contrary to bulk similar proposals that rely on manually-designed heuristics keywordbased filtering tweets, our provides data-driven approach modeling tweets are related cybersecurity events. We present cascaded Convolutional Neural Network (CNN) architecture, comprised binary classifier cyber-related multi-class model into multiple types threats. Furthermore, we an open-source...
Breast cancer screening policies attempt to achieve timely diagnosis by regularly healthy women via various imaging tests. Various clinical decisions are needed manage the process: selecting initial tests, interpreting test results, and deciding if further diagnostic tests required. Current guided practice guidelines (CPGs), which represent a "one-size-fits-all" approach, designed work well (on average) for population, can only offer coarse expert-based patient stratification that is not...
Although there is accumulating evidence regarding multimorbidity in Western countries, this information very limited Asian countries. This study aimed to estimate population-based, age-specific and gender-specific prevalence trends of the Taiwanese population.This was a cross-sectional based on claims data (National Health Insurance Research Database, Taiwan).The participants included subset National which contains for two million randomly selected beneficiaries (~10% total population) under...
High-Performance Computing (HPC) systems are resources utilized for data capture, sharing, and analysis. The majority of our HPC users come from other disciplines than Computer Science. including computer scientists have difficulties do not feel proficient enough to decide the required amount their submitted jobs on cluster. Consequently, encouraged over-estimate jobs, so will be killing due insufficient resources. This process waste devour resources; hence, this lead inefficient cluster...