A5049231383- Retinal Imaging and Analysis
- Retinal Diseases and Treatments
- Retinal and Optic Conditions
- Retinal Development and Disorders
- Artificial Intelligence in Games
- Ocular Diseases and Behçet’s Syndrome
- Reinforcement Learning in Robotics
- Topic Modeling
- Corruption and Economic Development
- Human Pose and Action Recognition
- Multimodal Machine Learning Applications
- Generative Adversarial Networks and Image Synthesis
- Digital Imaging for Blood Diseases
- Computational Physics and Python Applications
- Expert finding and Q&A systems
- Genomics and Rare Diseases
- COVID-19 diagnosis using AI
- Natural Language Processing Techniques
University College London
2022-2024
Moorfields Eye Hospital NHS Foundation Trust
2022-2024
Moorfields Eye Hospital
2022-2024
Institute of Ophthalmology
2022-2023
York University
2017-2023
University of Manchester
2018
PurposeRare disease diagnosis is challenging in medical image-based artificial intelligence due to a natural class imbalance datasets, leading biased prediction models. Inherited retinal diseases (IRDs) are research domain that particularly faces this issue. This study investigates the applicability of synthetic data improving AI-enabled IRDs using Generative Adversarial Networks (GANs).DesignDiagnostic gene-labeled fundus autofluorescence (FAF) IRD images Deep Learning...
Introduction Inherited retinal diseases (IRD) are a leading cause of visual impairment and blindness in the working age population. Mutations over 300 genes have been found to be associated with IRDs identifying affected gene patients by molecular genetic testing is first step towards effective care patient management. However, diagnosis currently slow, expensive not widely accessible. The aim current project address evidence gap IRD an AI algorithm, Eye2Gene, accelerate democratise service....
Abstract Purpose To quantify relevant fundus autofluorescence (FAF) image features cross-sectionally and longitudinally in a large cohort of inherited retinal diseases (IRDs) patients. Design Retrospective study imaging data (55-degree blue-FAF on Heidelberg Spectralis) from Participants Patients with clinical molecularly confirmed diagnosis IRD who have undergone FAF 55-degree at Moorfields Eye Hospital (MEH) the Royal Liverpool (RLH) between 2004 2019. Methods Five interest were defined:...
Background: Inherited Retinal Diseases (IRDs) are the leading cause of blindness in young people UK. Despite significant improvements genomics medicine, diagnosis these conditions remains challenging, with many patients enduring lengthy diagnostic odysseys and even after genetic testing around 40% them do not receive a definite diagnosis. This survey aims to explore experience individuals affected IRDs, their relatives, friends caregivers, potential acceptability an AI technology, such as...
Abstract Rare eye diseases such as inherited retinal (IRDs) are challenging to diagnose genetically. IRDs typically monogenic disorders and represent a leading cause of blindness in children working-age adults worldwide. A growing number now being targeted clinical trials, with approved treatments increasingly available. However, access requires genetic diagnosis be established sufficiently early. Critically, the timely identification remains challenging. We demonstrate that deep-learning...
To quantify relevant fundus autofluorescence (FAF) features cross-sectionally and longitudinally in a large cohort of patients with inherited retinal diseases (IRDs). Retrospective study imaging data. Patients clinical molecularly confirmed diagnosis IRD who have undergone 55° FAF at Moorfields Eye Hospital (MEH) the Royal Liverpool between 2004 2019. Five interest were defined: vessels, optic disc, perimacular ring increased signal (ring), relative hypo-autofluorescence (hypo-AF),...
Abstract Purpose To develop an automated system for assessing the quality of Fundus Autofluorescence (FAF) images in patients with inherited retinal diseases (IRD). Methods We annotated a dataset 2445 FAF from Inherited Retinal Dystrophies which were assessed by three different expert graders. Graders marked as either gradable (acceptable quality) or ungradable (poor quality), following strict grading protocol. This was used to train Convolutional Neural Network (CNN) classification model...
Purpose To evaluate the application of Retrieval-Augmented Generation (RAG), a technique that combines information retrieval with text generation, to benchmark performance open-source and proprietary generative large language models (LLMs) in medical question-answering tasks within ophthalmology domain. Methods Our dataset comprised 260 multiple-choice questions sourced from two question-answer banks designed assess ophthalmic knowledge: American Academy Ophthalmology's Basic Clinical...
Deep reinforcement learning (DRL) has proven to be an effective tool for creating general video-game AI. However most current DRL agents learn end-to-end from the video-output of game, which is superfluous many applications and creates a number additional problems. More importantly, directly working on pixel-based raw video data substantially distinct what human player does. In this paper, we present novel method enables object information. This obtained via use embedding network (OEN) that...
While learning models are typically studied for inputs in the form of a fixed dimensional feature vector, real world data is rarely found this form. In order to meet basic requirement traditional models, structural generally have be converted into fix-length vectors handcrafted manner, which tedious and may even incur information loss. A common structured what we term "semantic tree-structures", corresponding where rich semantic encoded compositional such as those expressed JavaScript Object...
Deep reinforcement learning (DRL) has proven to be an effective tool for creating general video-game AI. However most current DRL agents learn end-to-end from the video-output of game, which is superfluous many applications and creates a number additional problems. More importantly, directly working on pixel-based raw video data substantially distinct what human player does.In this paper, we present novel method enables object information. This obtained via use embedding network (OEN) that...
Abstract Purpose: Inherited retinal diseases (IRDs) are single‐gene disorders caused by genetic mutations in any one of over 270 genes. Identifying the causative gene through testing is crucial for targeted treatments, recruitment to clinical trials, prognosis and family planning. The prescription interpretation results requires phenotype–genotype recognition that only IRD experts can provide hence this has motivated AI approaches able predict probable from scans suspected patients. However,...