Optimized feature selection for enhanced accuracy in knee osteoarthritis detection and severity classification with machine learning

Biomedical Signal Processing and Control, Год журнала: 2024, Номер 97, С. 106670 - 106670

Опубликована: Авг. 10, 2024

Язык: Английский

---

A systematic review of the blockchain application in healthcare research domain: toward a unified conceptual model

Medical & Biological Engineering & Computing, Год журнала: 2025, Номер unknown

Опубликована: Янв. 10, 2025

Abstract Recently, research on blockchain applications in the healthcare domain has attracted increasing attention due to its strong potential. However, existing literature reveals limited studies defining use cases of clinical research, categorizing and comparing available studies. Therefore, this study aims explore significant potential through a comprehensive systematic review (SLR). To thoroughly investigate all aspects subject, we analyzed primary based questions (RQs) developed unified conceptual model using step-based creation . Studies from 2015 2023 were reviewed, 34 comprehensively by PICO template. In our findings, privacy emerged as most frequently cited requirement research. The mentioned for are ensuring data immutability security A issue identified beyond common limitations capacity scalability is lack standards compliance with legal frameworks like GDPR HIPAA. After these efforts, model, which, best knowledge, first support software developers researchers developing blockchain-based platforms efficiently. Graphical abstract

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

---

ADVANCED GENETIC ALGORITHM (GA)-INDEPENDENT COMPONENT ANALYSIS (ICA) ENSEMBLE MODEL FOR PREDICTING TRAPPED HUMANS THROUGH HYBRID DIMENSIONALITY REDUCTION

Scientific African, Год журнала: 2025, Номер unknown, С. e02564 - e02564

Опубликована: Янв. 1, 2025

Язык: Английский

---

Glaucoma Detection from Retinal Fundus Images using Graph Convolution Based Multi-Task Model

e-Prime - Advances in Electrical Engineering Electronics and Energy, Год журнала: 2025, Номер unknown, С. 100931 - 100931

Опубликована: Фев. 1, 2025

Язык: Английский

---

Optimizing deep learning models for glaucoma screening with vision transformers for resource efficiency and the pie augmentation method

PLoS ONE, Год журнала: 2025, Номер 20(3), С. e0314111 - e0314111

Опубликована: Март 21, 2025

Glaucoma is the leading cause of irreversible vision impairment, emphasizing critical need for early detection. Typically, AI-based glaucoma screening relies on fundus imaging. To tackle resource and time challenges in with convolutional neural network (CNN), we chose Data-efficient image Transformers (DeiT), a transformer, known its reduced computational demands, preprocessing decreased by factor 10. Our approach utilized meticulously annotated GlauCUTU-DATA dataset, curated ophthalmologists through consensus, encompassing both unanimous agreement (3/3) majority (2/3) data. However, DeiT’s performance was initially lower than CNN. Therefore, introduced “pie method," an augmentation method aligned ISNT rule. Along employing polar transformation to improved cup region visibility alignment transformer’s input elevated levels. The classification results demonstrated improvements comparable Using 3/3 data, excluding superior nasal regions, especially suspects, sensitivity increased 40.18% from 47.06% 88.24%. average area under curve (AUC) ± standard deviation (SD) glaucoma, no were 92.63 4.39%, 92.35 92.32 1.45%, respectively. With 2/3 temporal diagnosing 11.36% 47.73% 59.09%. AUC SD 68.22 4.45%, 68.23 73.09 3.05%, For datasets, values 84.53%, 84.54%, 91.05%, respectively, which CNN model that achieved 84.70%, 84.69%, 93.19%, Moreover, incorporation attention maps DeiT facilitated precise localization clinically significant areas, such as disc rim notching, thereby enhancing overall effectiveness screening.

Язык: Английский

---

Glaucoma diagnosis using Gabor and entropy coded Sine Cosine integration in adaptive partial swarm optimization-based FAWT

Biomedical Signal Processing and Control, Год журнала: 2025, Номер 107, С. 107832 - 107832

Опубликована: Март 26, 2025

Язык: Английский

---

Applying a Hybrid Gray Wolf‐Enhanced Whale Optimization Algorithm to the Capacitated Vehicle Routing Problem

Journal of Advanced Transportation, Год журнала: 2025, Номер 2025(1)

Опубликована: Янв. 1, 2025

The study presents a novel hybrid gray wolf and whale optimization algorithm (hGWOAM) for the capacitated vehicle routing problem (CVRP). By integrating enhanced (EWOA) optimizer (GWO) with tournament selection, opposition‐based learning, mutation techniques, hGWOAM enhances efficiency under capacity constraints. Computational evaluations demonstrate its superior performance, achieving lower percentage deviations (%dev) compared to existing algorithms across multiple case studies real‐world applications. In Case Study 1, achieved mean deviation than EWOA (0.89%), GWO (0.74%), SCA (0.59%), DA (1.63%), ALO (2.26%), MHPSO (1.85%), PSO (1.96%), DPGA (2.85%), SGA (4.14%). 2, outperformed (12.05%), (2.53%), (21.07%), (17.58%). application, it best %dev, surpassing (6.64%), (6.34%), (9.01%), (12.24%). These findings highlight hGWOAM’s potential optimizing logistics, reducing operational costs, minimizing environmental impact while also paving way future advancements in metaheuristic optimization.

Язык: Английский

---

HybridGWOSPEA2ABC: a novel feature selection algorithm for gene expression data analysis and cancer classification

Computer Methods in Biomechanics & Biomedical Engineering, Год журнала: 2025, Номер unknown, С. 1 - 22

Опубликована: Апрель 26, 2025

DNA micro-array technology has a remarkable impact on biological research, particularly in categorizing and diagnosing cancer studying gene features functions. With the availability of extensive collections cancer-related data, there been an increased focus developing optimized Machine Learning (ML) techniques for classification through pattern analysis identification specific genes type categorization. The relevant selection treating poses significant challenge, which requires efficient feature methods. This study introduces novel hybrid algorithm, selection, integrating Grey Wolf Optimizer (GWO), Strength Pareto Evolutionary Algorithm 2 (SPEA2), Artificial Bee Colony (ABC). combination uses intelligence evolutionary computation to enhance solution diversity, convergence efficiency, exploration exploitation capabilities high-dimensional expression data. algorithm was compared with five bio-inspired algorithms using different classifiers various datasets validate its effectiveness selection. HybridGWOSPEA2ABC demonstrated superior performance identifying biomarkers conventional algorithms. Comparison benchmark shown approach's enhanced capability addressing challenges data advancing problem classification. hybridization enhances by maintaining efficiently converging optimal solutions, improving search space. provides better understanding promotes effective methodologies disease detection

Язык: Английский

---

Modified quantum dilated convolutional neural network for cancer prediction using gene expression data

Computer Methods in Biomechanics & Biomedical Engineering, Год журнала: 2025, Номер unknown, С. 1 - 13

Опубликована: Май 20, 2025

This paper proposes a modified Quantum Dilated Convolutional neural network (QDCNN) to detect cancer using gene expression data. Primarily, the input data is taken from specified dataset. Then, transformation done Adaptive Box-Cox and feature fusion by Deep Neural Network (DNN) with Kulczynski. The refined features are then fed into QDCNN, which effectively predicts cancer. QDCNN attains an accuracy of 90.6%, True Positive Rate (TPR) 89.0%, False Negative (FNR) 0.109, Matthews correlation coefficient (MCC) 89.9% when PANCAN

Язык: Английский

---