A novel and efficient statistical and soft-computing intelligence integrated feature selection technique for human chronic diseases prediction

Multimedia Tools and Applications, Journal Year: 2025, Volume and Issue: unknown

Published: April 9, 2025

Language: Английский

---

Federated Learning for Secure and Privacy-Preserving Medical Diagnostics

Advances in computational intelligence and robotics book series, Journal Year: 2025, Volume and Issue: unknown, P. 155 - 186

Published: March 7, 2025

Federated learning is an emerging powerful approach for training ML models across decentralized data sources, such as in healthcare, without actually transferring the sensitive to a central server. Although FL naturally reduces risks of privacy since are confined local devices, it still can be highly vulnerable sophisticated attacks model inversion, poisoning, and inference attacks. Advanced privacy-preserving techniques that nowadays studied addressing security concerns this paper include differential privacy, homomorphic encryption, secure aggregation. These, therefore, open room adopted solution some privacy-sensitive domains. This conducts thorough review discussed outlines customized framework healthcare applications allow patient safeguarded detrimental effects on performance accuracy. Conclusion:

Language: Английский

---

An improved breast cancer classification with hybrid chaotic sand cat and Remora Optimization feature selection algorithm

PLoS ONE, Journal Year: 2024, Volume and Issue: 19(4), P. e0300622 - e0300622

Published: April 11, 2024

Breast cancer is one of the most often diagnosed cancers in women, and identifying breast histological images an essential challenge automated pathology analysis. According to research, global BrC around 12% all cases. Furthermore, 25% women suffer from BrC. Consequently, prediction depends critically on quick precise processing imaging data. The primary reason deep learning models are used detection that they can produce findings more quickly accurately than current machine learning-based techniques. Using a BreakHis dataset, we demonstrated this work viability automatically classifying first stage pre-processing, which employs Adaptive Switching Modified Decision Based Unsymmetrical Trimmed Median Filter (ASMDBUTMF) remove high-density noise. After image has been pre-processed, it segmented using Thresholding Level set approach. Next, propose hybrid chaotic sand cat optimization technique, together with Remora Optimization Algorithm (ROA) for feature selection. suggested strategy facilitates acquisition functionality attributes, hence simplifying procedure. Additionally, aids resolving problems pertaining optimization. Following selection, best characteristics proceed categorization A DL classifier called Conditional Variation Autoencoder discriminate between cancerous benign tumors while categorizing them. classification accuracy 99.4%, Precision 99.2%, Recall 99.1%, F- score 99%, Specificity 99.14%, FDR 0.54, FNR 0.001, FPR 0.002, MCC 0.98 NPV 0.99 were obtained proposed compared other research results our desirable.

Language: Английский

---

Predicting invasion in early-stage ground-glass opacity pulmonary adenocarcinoma: a radiomics-based machine learning approach

BMC Medical Imaging, Journal Year: 2024, Volume and Issue: 24(1)

Published: Sept. 13, 2024

Language: Английский

---

FS-WOA-stacking: A novel ensemble model for early diagnosis of breast cancer

Biomedical Signal Processing and Control, Journal Year: 2024, Volume and Issue: 95, P. 106374 - 106374

Published: April 26, 2024

Language: Английский

---

A novel fusion of genetic grey wolf optimization and kernel extreme learning machines for precise diabetic eye disease classification

PLoS ONE, Journal Year: 2024, Volume and Issue: 19(5), P. e0303094 - e0303094

Published: May 20, 2024

In response to the growing number of diabetes cases worldwide, Our study addresses escalating issue diabetic eye disease (DED), a significant contributor vision loss globally, through pioneering approach. We propose novel integration Genetic Grey Wolf Optimization (G-GWO) algorithm with Fully Convolutional Encoder-Decoder Network (FCEDN), further enhanced by Kernel Extreme Learning Machine (KELM) for refined image segmentation and classification. This innovative combination leverages genetic grey wolf optimization boost FCEDN’s efficiency, enabling precise detection DED stages differentiation among types. Tested across diverse datasets, including IDRiD, DR-HAGIS, ODIR, our model showcased superior performance, achieving classification accuracies between 98.5% 98.8%, surpassing existing methods. advancement sets new standard in offers potential automating fundus analysis, reducing reliance on manual examination, improving patient care efficiency. findings are crucial enhancing diagnostic accuracy outcomes management.

Language: Английский

---

Hybrid bio-inspired computing in medical image data analysis: A review

Intelligent Decision Technologies, Journal Year: 2024, Volume and Issue: unknown, P. 1 - 18

Published: Aug. 30, 2024

Inspired by the fundamentals of biological evolution, bio-inspired algorithms are becoming increasingly popular for developing robust optimization techniques. These metaheuristic algorithms, unlike gradient descent methods, computationally more efficient and excel in handling higher order multi-dimensional non-linear. OBJECTIVES: To understand hybrid Bio-inspired domain Medical Imaging its challenges feature selection METHOD: The primary research was conducted using three major indexing database Scopus, Web Science Google Scholar. RESULT: included 198 articles, after removing 103 duplicates, 95 articles remained as per criteria. Finally 41 were selected study. CONCLUSION: We recommend that further area based field diagnostic imaging clustering. Additionally, there is a need to investigate use Deep Learning models integrating include strengths each enhances overall model.

Language: Английский

---

Revolutionizing breast cancer diagnosis with a comprehensive approach using digital mammogram-based feature extraction and selection for early-stage identification

Biomedical Signal Processing and Control, Journal Year: 2024, Volume and Issue: 94, P. 106268 - 106268

Published: April 16, 2024

Language: Английский

---

Optimizing Breast Cancer Diagnosis: Harnessing the Power of Nature-Inspired Metaheuristics for Feature Selection with Soft Voting Classifiers

International Journal of Cognitive Computing in Engineering, Journal Year: 2024, Volume and Issue: unknown

Published: Sept. 1, 2024

Language: Английский

---

BELUGA WHALE LION OPTIMIZATION IN DEEP-NETS FOR HUMAN AGE ESTIMATION USING HAND X-RAY

Biomedical Engineering Applications Basis and Communications, Journal Year: 2024, Volume and Issue: 36(04)

Published: July 10, 2024

Bones undergo significant changes in size and shape with the growth of child, bone age estimation is crucial for determining growth, genetic endocrine disorders children. Hand X-ray images are extensively utilized diagnosing The variation chronological indicates presence disorders, problems, abnormalities. Traditionally, estimated manually by inspecting images, which extremely time-consuming prone to error. Further, accuracy estimate depends on experience medical practitioner, thus it suffers from intra- inter-observer variability. Hence, overcome these issues, essential devise automatic methods that can high a short duration. In this work, using Deep Residual Network (DRN), whose learnable factors adjusted devised Beluga Whale Lion Optimization (BWLO) algorithm. BWLO_DRN examined its superiority considering metrics, like accuracy, True Positive Rate (TPR), Negative (TNR), corresponding values 89.8%, 86.8%, 90% found be achieved experimental results.

Language: Английский

---