A5035370376- Retinal Imaging and Analysis
- Retinal Diseases and Treatments
- Adipose Tissue and Metabolism
- Glaucoma and retinal disorders
- Retinal and Optic Conditions
- Cardiovascular Disease and Adiposity
- Artificial Intelligence in Healthcare and Education
- Ophthalmology and Visual Impairment Studies
- Fatty Acid Research and Health
- CNS Lymphoma Diagnosis and Treatment
- Retinal and Macular Surgery
- Explainable Artificial Intelligence (XAI)
- Machine Learning in Healthcare
- Topic Modeling
- Vascular Malformations Diagnosis and Treatment
- Multimodal Machine Learning Applications
- COVID-19 diagnosis using AI
- AI in cancer detection
- Antioxidant Activity and Oxidative Stress
- Advanced Glycation End Products research
- Meta-analysis and systematic reviews
- Optical Coherence Tomography Applications
- Intraocular Surgery and Lenses
Tsinghua University
2024-2025
Chinese University of Hong Kong
2020-2025
Shantou University
2020-2025
Beijing Tsinghua Chang Gung Hospital
2024-2025
Shantou University Medical College
2021-2024
Background/aims Large language models (LLMs) have substantial potential to enhance the efficiency of academic research. The accuracy and performance LLMs in a systematic review, core part evidence building, has yet be studied detail. Methods We introduced two LLM-based approaches review: an LLM-enabled fully automated approach (LLM-FA) utilising three different GPT-4 plugins (Consensus GPT, Scholar GPT web browsing modes) LLM-facilitated semi-automated (LLM-SA) using GPT4’s Application...
Vitreoretinal lymphoma (VRL) remains a diagnostic challenge due to its scarce prevalence, and delayed diagnosis usually results in blindness even fatal outcomes. Herein, an artificial intelligence (AI) system is developed identify VRL among 16 retinal diseases conditions on optical coherence tomography (OCT) images with the cross‐subject meta‐transfer learning (CS‐MTL) algorithm. Extensive experiments of few‐shot recognition tasks prove robustness our model 1‐, 3‐, 5‐shot scenarios,...
To evaluate the anatomical and functional outcomes of internal limiting membrane (ILM) peeling inverted ILM flap technique for rhegmatogenous retinal detachment coexisting with macular hole.This retrospective study evaluated medical records 79 eyes concurrent hole received vitrectomy silicone oil tamponade, on 56 23 eyes.The Type 1 closure rate was greater in group than (82.6% vs. 55.4%, P = 0.038). Lines improvement were 7.8 ± 5.3 8.9 5.6 group. Postoperative epiretinal reattachment rates...
PurposeThis study aims to estimate the regional choroidal thickness from color fundus images convolutional neural networks in different network structures and task learning models.Method1276 photos their corresponding values healthy subjects were obtained Topcon DRI Triton optical coherence tomography machine. Initially, ten commonly used deployed identify most accurate model, which was subsequently selected for further training. This model then employed combination with single-, multiple-,...
Purpose: To investigate the topographic characters of inter-individual variations macular choroidal thickness (CT). Methods: This was a retrospective study. Macular CT data for 900 0.2 × 0.2-mm grids from 410 healthy eyes were collected swept-source optical coherence tomography. Following analysis factors associated with mean CT, β-coefficients included in each grid summarized changes analysis. Additionally, coefficient variance (CoV), determination (CoD), and unexplained (CoVU) calculated...
Automatic and accurate estimation of choroidal thickness plays a very important role in computer-aided system for eye diseases. One the most common methods automatic is segmentation-based methods, which boundaries choroid are first detected from optical coherence tomography (OCT) images. The then computed based on boundaries. A shortcoming that estimating precision greatly depends segmentation results. To avoid dependence step, this paper, we propose direct method convolutional neural...
The current retinal artificial intelligence models were trained using data with a limited category of diseases and knowledge. In this paper, we present vision-language foundation model (RetiZero) knowledge over 400 fundus diseases. Specifically, collected 341,896 images paired text descriptions from 29 publicly available datasets, 180 ophthalmic books, online resources, encompassing across multiple countries ethnicities. RetiZero achieved outstanding performance various downstream tasks,...
Abstract Background: Genetic association of uncoupling proteins (UCPs) variants with the susceptibility diabetic retinopathy (DR) in diabetes mellites (DM) patients has been reported but controversy. Here we aimed to conduct a meta-analysis confirm different UCPs DR. Methods: Three databases (Medline Ovid, Embase Ovid and CENTRAL) were applied literature search. Five genetic models, including allelic, homozygous, heterozygous, dominant recessive evaluated. Odds ratios (OR) estimated under...