Overview of Artificial Intelligence Systems in Ophthalmology

Ruamviboonsuk, Paisan; Kaothanthong, Natsuda; Theeramunkong, Thanaruk; Ruamviboonsuk, Varis

doi:10.1007/978-3-030-78601-4_3

Cited by 3 publications

(3 citation statements)

References 114 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The inception algorithm, like other CNN architectures, has been widely used in image classification to solve various challenges such as prediction, diagnosis, and recognition. The goal of this architecture is to increase the network's width and depth in order to achieve high accuracy [25]. The Inception-V3 was employed in this study.…”

Section: Inceptionnetmentioning

confidence: 99%

Comparing machine learning and deep learning classifiers for enhancing agricultural productivity: case study in Larache Province, Northern Morocco

Belattar

Abdoun

Haimoudi

2023

IJECE

View full text Add to dashboard Cite

<p><span lang="EN-US">The agriculture sector in the Tangier-Tetouan-Al-Hoceima-Region (Northern Morocco) contributes a significant percentage to the national revenue. The Larache Province is at the regional forefront in agriculture terms due to its large irrigated areas. Golden-Gogi is a biological farm located in the Larache Province, and its objective is to produce organic crops. Besides climate change, this farm suffers from biotic factors such as snails and insects. These problems cause diseases in plants, resulting in massive crop production losses. Early detection of disease and biotic factors in plants is a difficult task for farmers, but it is now possible thanks to artificial intelligence. For that reason, we aim to contribute to this Province by comparing the well-known models in machine learning (ML) and deep learning (DL) used in early plant disease detection to specify the best-classifier in terms of detecting mint plant diseases. Mint plant is a major crop on the Golden-Gogi farm, and its dataset was collected from there. As per findings, DL classifiers outperform ML classifiers in disease detection. The best-classifier is DenseNet201, with high accuracy of 94.12%. Hence, the system using DenseNet201 offers a solution for farmers of this Province in making urgent decisions to avoid mint yield losses.</span></p>

show abstract

Section: Inceptionnetmentioning

confidence: 99%

Comparing machine learning and deep learning classifiers for enhancing agricultural productivity: case study in Larache Province, Northern Morocco

Belattar

Abdoun

Haimoudi

2023

IJECE

View full text Add to dashboard Cite

show abstract

“… 19 , 21 – 25 Although various transfer learning approaches are used in the deep learning model scheme, concerns about the “black-box” issue and interpretability of these models in clinical research and application remain. 27 , 28 This motivates us to explore regression-based transfer learning, which enhances the prediction accuracy by borrowing information from other source data and enables the study of associations between parameters of interest and clinical outcomes.…”

Section: Introductionmentioning

confidence: 99%

Reducing Ophthalmic Health Disparities Through Transfer Learning: A Novel Application to Overcome Data Inequality

Lee,

Wollstein,

Madu

et al. 2023

Trans. Vis. Sci. Tech.

View full text Add to dashboard Cite

Purpose Race disparities in the healthcare system and the resulting inequality in clinical data among different races hinder the ability to generate equitable prediction results. This study aims to reduce healthcare disparities arising from data imbalance by leveraging advanced transfer learning (TL) methods. Method We examined the ophthalmic healthcare disparities at a population level using electronic medical records data from a study cohort (N = 785) receiving care at an academic institute. Regression-based TL models were usesd, transferring valuable information from the dominant racial group (White) to improve visual field mean deviation (MD) rate of change prediction particularly for data-disadvantaged African American (AA) and Asian racial groups. Prediction results of TL models were compared with two conventional approaches. Results Disparities in socioeconomic status and baseline disease severity were observed among the AA and Asian racial groups. The TL approach achieved marked to comparable improvement in prediction accuracy compared to the two conventional approaches as evident by smaller mean absolute errors or mean square errors. TL identified distinct key features of visual field MD rate of change for each racial group. Conclusions The study introduces a novel application of TL that improved reliability of the analysis in comparison with conventional methods, especially in small sample size groups. This can improve assessment of healthcare disparity and subsequent remedy approach. Translational Relevance TL offers an equitable and efficient approach to mitigate healthcare disparities analysis by enhancing prediction performance for data-disadvantaged group.

show abstract

“…The most common way to train a deep learning model for medical image classification purposes, including for ophthalmic images, involves supervised learning in which training data are manually labeled by trained human graders. Then, transfer learning may be applied to a pretrained “off-the-shelf” model backbone, such as VGG and ResNet, and model fine-tuning is performed with the labeled ophthalmic data. This common workflow is limited by the time-consuming and labor-intensive nature of training data annotation.…”

mentioning

confidence: 99%

Self-Supervised Deep Learning—The Next Frontier

Liu,

Koseoglu,

Jones

2024

JAMA Ophthalmol

View full text Add to dashboard Cite

The most common way to train a deep learning model for medical image classification purposes, including for ophthalmic images, involves supervised learning in which training data are manually labeled by trained human graders.Then, transfer learning may be applied to a pretrained "off-the-shelf" model backbone, such as VGG and ResNet, 1 and model fine-tuning is performed with the labeled ophthalmic data. This common workflow is limited by the time-consuming and laborintensive nature of training data annotation. One promising approach to address this limitation is self-supervised learning (SSL), which is the focus of the article by Gholami et al 2 in JAMA Ophthalmology. As the name suggests, SSL is a technique that obviates the need for human annotation of the training data.Gholami et al 2 explored how to incorporate SSL into model development for detection of macular telangiectasia in optical coherence tomography (OCT) images. Their study had 3 main findings. First, if a decent amount of labeled training data was available (using 100% of the available labeled training data in their simulation), then first finetuning the model via SSL could incrementally boost model performance compared with traditional supervised learning (TSL). Second, if only a small amount of labeled training data was available (using 10% of the available labeled training data in their simulation), then an SSL-first approach dramatically outperformed a TSL-only approach. Third, their SSL model fine-tuned with only 10% of available labeled training data achieved performance comparable to that of the best model in their study.

show abstract

Overview of Artificial Intelligence Systems in Ophthalmology

Cited by 3 publications

References 114 publications

Comparing machine learning and deep learning classifiers for enhancing agricultural productivity: case study in Larache Province, Northern Morocco

Comparing machine learning and deep learning classifiers for enhancing agricultural productivity: case study in Larache Province, Northern Morocco

Reducing Ophthalmic Health Disparities Through Transfer Learning: A Novel Application to Overcome Data Inequality

Self-Supervised Deep Learning—The Next Frontier

Contact Info

Product

Resources

About