Big data for imaging assessment in glaucoma DOI Creative Commons

Douglas R. da Costa,

Felipe A. Medeiros

Taiwan Journal of Ophthalmology, Journal Year: 2024, Volume and Issue: 14(3), P. 299 - 318

Published: July 1, 2024

Glaucoma is the leading cause of irreversible blindness worldwide, with many individuals unaware their condition until advanced stages, resulting in significant visual field impairment. Despite effective treatments, over 110 million people are projected to have glaucoma by 2040. Early detection and reliable monitoring crucial prevent vision loss. With rapid development computational technologies, artificial intelligence (AI) deep learning (DL) algorithms emerging as potential tools for screening, diagnosing, progression. Leveraging vast data sources, these technologies promise enhance clinical practice public health outcomes enabling earlier disease detection, progression forecasting, deeper understanding underlying mechanisms. This review evaluates use Big Data AI research, providing an overview most relevant topics discussing various models diagnosis, progression, correlating structural functional changes, assessing image quality, exploring innovative such generative AI.

Language: Английский

FairCLIP: Harnessing Fairness in Vision-Language Learning DOI
Yan Luo, Min Shi, Muhammad Osama Khan

et al.

2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Journal Year: 2024, Volume and Issue: 30, P. 12289 - 12301

Published: June 16, 2024

Language: Английский

Citations

10
Harvard Glaucoma Fairness: A Retinal Nerve Disease Dataset for Fairness Learning and Fair Identity Normalization DOI
Yan Luo, Yu Tian, Min Shi

et al.

IEEE Transactions on Medical Imaging, Journal Year: 2024, Volume and Issue: 43(7), P. 2623 - 2633

Published: March 18, 2024

Fairness (also known as equity interchangeably) in machine learning is important for societal well-being, but limited public datasets hinder its progress. Currently, no dedicated medical with imaging data fairness are available, though underrepresented groups suffer from more health issues. To address this gap, we introduce Harvard Glaucoma (Harvard-GF), a retinal nerve disease dataset including 3,300 subjects both 2D and 3D balanced racial glaucoma detection. the leading cause of irreversible blindness globally Blacks having doubled prevalence than other races. We also propose fair identity normalization (FIN) approach to equalize feature importance between different groups. Our FIN compared various state-of-the-art methods superior performance racial, gender, ethnicity tasks data, demonstrating utilities our Harvard-GF learning. facilitate comparisons models, an equity-scaled measure, which can be flexibly used compare all kinds metrics context fairness. The code publicly accessible via https://ophai.hms.harvard.edu/datasets/harvard-gf3300/.

Language: Английский

Citations

7
Equitable artificial intelligence for glaucoma screening with fair identity normalization DOI Creative Commons
Min Shi, Yan Luo, Yu Tian

et al.

npj Digital Medicine, Journal Year: 2025, Volume and Issue: 8(1)

Published: Jan. 20, 2025

Language: Английский

Citations

1
Latest developments of generative artificial intelligence and applications in ophthalmology DOI Creative Commons
Xiaoru Feng,

Kezheng Xu,

Mingjie Luo

et al.

Asia-Pacific Journal of Ophthalmology, Journal Year: 2024, Volume and Issue: 13(4), P. 100090 - 100090

Published: July 1, 2024

The emergence of generative artificial intelligence (AI) has revolutionized various fields. In ophthalmology, AI the potential to enhance efficiency, accuracy, personalization and innovation in clinical practice medical research, through processing data, streamlining documentation, facilitating patient-doctor communication, aiding decision-making, simulating trials. This review focuses on development integration models into workflows scientific research ophthalmology. It outlines need for a standard framework comprehensive assessments, robust evidence, exploration multimodal capabilities intelligent agents. Additionally, addresses risks model application service including data privacy, bias, adaptation friction, over interdependence, job replacement, based which we summarized risk management mitigate these concerns. highlights transformative enhancing patient care, improving operational efficiency also advocates balanced approach its adoption.

Language: Английский

Citations

5
Explainable Deep Learning for Glaucomatous Visual Field Prediction: Artifact Correction Enhances Transformer Models DOI Creative Commons

Kornchanok Sriwatana,

Chanon Puttanawarut, Yanin Suwan

et al.

Translational Vision Science & Technology, Journal Year: 2025, Volume and Issue: 14(1), P. 22 - 22

Published: Jan. 23, 2025

Purpose: The purpose of this study was to develop a deep learning approach that restores artifact-laden optical coherence tomography (OCT) scans and predicts functional loss on the 24-2 Humphrey Visual Field (HVF) test. Methods: This cross-sectional, retrospective used 1674 visual field (VF)-OCT pairs from 951 eyes for training 429 345 testing. Peripapillary retinal nerve fiber layer (RNFL) thickness map artifacts were corrected using generative diffusion model. Three convolutional neural networks 2 transformer-based models trained original artifact-corrected datasets estimate 54 sensitivity thresholds HVF Results: Predictive performances calculated root mean square error (RMSE) absolute (MAE), with explainability evaluated through GradCAM, attention maps, dimensionality reduction techniques. Distillation No Labels (DINO) Vision Transformers (ViT) achieved highest accuracy (RMSE, 95% confidence interval [CI] = 4.44, CI 4.07, 4.82 decibel [dB], MAE 3.46, 3.14, 3.79 dB), greatest interpretability, showing improvements 0.15 dB in global RMSE (P < 0.05) compared performance maps. Feature maps visualization tools indicate compromise DINO-ViT's predictive ability but improve artifact correction. Conclusions: Combining self-supervised ViTs correction enhances correlation between glaucomatous structures functions. Translational Relevance: Our offers comprehensive tool glaucoma management, facilitates exploration structure-function correlations research, underscores importance addressing clinical interpretation OCT.

Language: Английский

Citations

0
Integrating Retinal Segmentation Metrics with Machine Learning for Predictions from Mouse SD-OCT Scans DOI Creative Commons

Maide Gözde İnam,

Onur İnam,

Xiangjun Yang

et al.

Current Eye Research, Journal Year: 2025, Volume and Issue: unknown, P. 1 - 10

Published: Jan. 23, 2025

Purpose This study aimed to initially test whether machine learning approaches could categorically predict two simple biological features, mouse age and species, using the retinal segmentation metrics.

Language: Английский

Citations

0
Machine learning-assisted image analysis techniques for glaucoma detection DOI Creative Commons
Vaibhav Yadav, Barnali Dey, Udayan Baruah

et al.

EURASIP Journal on Image and Video Processing, Journal Year: 2025, Volume and Issue: 2025(1)

Published: May 10, 2025

Language: Английский

Citations

0
FairDomain: Achieving Fairness in Cross-Domain Medical Image Segmentation and Classification DOI
Yu Tian, Congcong Wen, Min Shi

et al.

Lecture notes in computer science, Journal Year: 2024, Volume and Issue: unknown, P. 251 - 271

Published: Oct. 30, 2024

Language: Английский

Citations

2
Big data for imaging assessment in glaucoma DOI Creative Commons

Douglas R. da Costa,

Felipe A. Medeiros

Taiwan Journal of Ophthalmology, Journal Year: 2024, Volume and Issue: 14(3), P. 299 - 318

Published: July 1, 2024

Glaucoma is the leading cause of irreversible blindness worldwide, with many individuals unaware their condition until advanced stages, resulting in significant visual field impairment. Despite effective treatments, over 110 million people are projected to have glaucoma by 2040. Early detection and reliable monitoring crucial prevent vision loss. With rapid development computational technologies, artificial intelligence (AI) deep learning (DL) algorithms emerging as potential tools for screening, diagnosing, progression. Leveraging vast data sources, these technologies promise enhance clinical practice public health outcomes enabling earlier disease detection, progression forecasting, deeper understanding underlying mechanisms. This review evaluates use Big Data AI research, providing an overview most relevant topics discussing various models diagnosis, progression, correlating structural functional changes, assessing image quality, exploring innovative such generative AI.

Language: Английский

Citations

1