Feature subset selection through nature inspired computing for efficient glaucoma classification from fundus images

Multimedia Tools and Applications, Год журнала: 2024, Номер 83(32), С. 77873 - 77944

Опубликована: Фев. 23, 2024

Язык: Английский

---

A snake optimization algorithm-based feature selection framework for rapid detection of cardiovascular disease in its early stages

Biomedical Signal Processing and Control, Год журнала: 2024, Номер 102, С. 107417 - 107417

Опубликована: Дек. 24, 2024

Язык: Английский

---

Prediction of sepsis mortality in ICU patients using machine learning methods

BMC Medical Informatics and Decision Making, Год журнала: 2024, Номер 24(1)

Опубликована: Авг. 16, 2024

Abstract Problem Sepsis, a life-threatening condition, accounts for the deaths of millions people worldwide. Accurate prediction sepsis outcomes is crucial effective treatment and management. Previous studies have utilized machine learning prognosis, but limitations in feature sets model interpretability. Aim This study aims to develop that enhances accuracy using reduced set features, thereby addressing previous enhancing Methods analyzes intensive care patient MIMIC-IV database, focusing on adult cases. Employing latest data extraction tools, such as Google BigQuery, following stringent selection criteria, we selected 38 features this study. also informed by comprehensive literature review clinical expertise. Data preprocessing included handling missing values, regrouping categorical variables, Synthetic Minority Over-sampling Technique (SMOTE) balance data. We evaluated several models: Decision Trees, Gradient Boosting, XGBoost, LightGBM, Multilayer Perceptrons (MLP), Support Vector Machines (SVM), Random Forest. The Sequential Halving Classification (SHAC) algorithm was used hyperparameter tuning, both train-test split cross-validation methodologies were employed performance computational efficiency. Results Forest most effective, achieving an area under receiver operating characteristic curve (AUROC) 0.94 with confidence interval ±0.01. significantly outperformed other models new benchmark literature. provided detailed insights into importance various Organ Failure Assessment (SOFA) score average urine output being highly predictive. SHAP (Shapley Additive Explanations) analysis further enhanced model’s interpretability, offering clearer understanding impacts. Conclusion demonstrates significant improvements predicting model, supported advanced techniques thorough preprocessing. Our approach key impacting mortality, making accurate interpretable. By practical utility settings, offer valuable tool healthcare professionals make data-driven decisions, ultimately aiming minimize sepsis-induced fatalities.

Язык: Английский

---

An Explainable Artificial Intelligence Software System for Predicting Diabetes

Heliyon, Год журнала: 2024, Номер 10(16), С. e36112 - e36112

Опубликована: Авг. 1, 2024

Implementing diabetes surveillance systems is paramount to mitigate the risk of incurring substantial medical expenses. Currently, blood glucose measured by minimally invasive methods, which involve extracting a small sample and transmitting it meter. This method deemed discomforting for individuals who are undergoing it. The present study introduces an Explainable Artificial Intelligence (XAI) system, aims create intelligible machine capable explaining expected outcomes decision models. To this end, we analyze abnormal levels utilizing Bi-directional Long Short-Term Memory (Bi-LSTM) Convolutional Neural Network (CNN). In regard, acquired through oxidase (GOD) strips placed over human body. Later, signal data converted spectrogram images, classified as low glucose, average levels. labeled images then used train individualized monitoring model. proposed XAI model track real-time uses XAI-driven architecture in its feature processing. model's effectiveness evaluated analyzing performance several evolutionary metrics confusion matrix. revealed demonstrate that effectively identifies with elevated

Язык: Английский

---

A three-stage novel framework for efficient and automatic glaucoma classification from retinal fundus images

Multimedia Tools and Applications, Год журнала: 2024, Номер unknown

Опубликована: Июнь 14, 2024

Язык: Английский

---

Glaucoma detection with explainable AI using convolutional neural networks based feature extraction and machine learning classifiers

IET Image Processing, Год журнала: 2024, Номер 18(13), С. 3827 - 3853

Опубликована: Авг. 19, 2024

Abstract Glaucoma is an eye disease that damages the optic nerve as a result of vision loss, it leading cause blindness worldwide. Due to time‐consuming, inaccurate, and manual nature traditional methods, automation in glaucoma detection important. This paper proposes explainable artificial intelligence (XAI) based model for automatic using pre‐trained convolutional neural networks (PCNNs) machine learning classifiers (MLCs). PCNNs are used feature extractors obtain deep features can capture important visual patterns characteristics from fundus images. Using extracted MLCs then classify healthy An empirical selection CNN MLC parameters has been made performance evaluation. In this work, total 1,865 1,590 images different datasets were used. The results on ACRIMA dataset show accuracy, precision, recall 98.03%, 97.61%, 99%, respectively. Explainable aims create increase user's trust model's decision‐making process transparent interpretable manner. assessment image misclassification carried out facilitate future investigations.

Язык: Английский

---

A comparative study on the effectiveness of interactive versus non-interactive optic disc photograph training modules for glaucoma diagnosis among ophthalmology residents

Cogent Education, Год журнала: 2025, Номер 12(1)

Опубликована: Янв. 2, 2025

Early glaucoma detection through accurate optic disc interpretation is essential but challenging for ophthalmology residents. This study evaluated the effectiveness of interactive (ITM) versus non-interactive (NITM) web-based training modules in improving skills diagnosis among Ninety-six residents from five centers Thailand were randomized into ITM or NITM groups. Both groups completed pre- and post-tests containing 30 standardized photographs used self-study with 100 images obtained CLARUS™ 500 over two months. The group received immediate feedback on their answers, while only viewed correct answers without interaction. demonstrated significant improvement scores after (P < 0.001), no difference between = 0.231). Third-year showed greater score compared to first-year 0.009). Satisfaction comparable 0.416). findings suggest that both improve residents' ability evaluate glaucomatous discs, though statistically was found approaches.

Язык: Английский

---

Advancing multiple sclerosis diagnosis through an innovative hybrid AI framework incorporating Multi-view ResNet and quantum RIME-inspired metaheuristics

Ain Shams Engineering Journal, Год журнала: 2025, Номер 16(2), С. 103241 - 103241

Опубликована: Янв. 13, 2025

Язык: Английский

---

Optimising deep learning models for ophthalmological disorder classification

Scientific Reports, Год журнала: 2025, Номер 15(1)

Опубликована: Янв. 24, 2025

Abstract Fundus imaging, a technique for recording retinal structural components and anomalies, is essential observing identifying ophthalmological diseases. Disorders such as hypertension, glaucoma, diabetic retinopathy are indicated by alterations in the optic disc, blood vessels, fovea, macula. Patients frequently deal with various conditions either one or both eyes. In this article, we have used different deep learning models categorisation of disorders into multiple classes labels utilising transfer learning-based convolutional neural network (CNN) methods. The Ocular Disease Intelligent Recognition (ODIR) database experiments, it contains fundus images patient’s left right We compared performance two optimisers, Stochastic Gradient Descent (SGD) Adam, separately. best result was achieved using MobileNet model Adam optimiser, yielding testing accuracy 89.64%.

Язык: Английский

---

Classification of Diabetic Retinopathy Severity Using Deep Learning Techniques on Retinal Images

Cybernetics & Systems, Год журнала: 2024, Номер unknown, С. 1 - 25

Опубликована: Июль 13, 2024

Язык: Английский

---