Enhanced Classification of Imbalanced Medical Datasets using Hybrid Data-Level, Cost-Sensitive and Ensemble Methods DOI Open Access
Ayushi Gupta, Shikha Gupta

International Research Journal of Multidisciplinary Technovation, Journal Year: 2024, Volume and Issue: unknown, P. 58 - 76

Published: April 22, 2024

Addressing the class imbalance in classification problems is particularly challenging, especially context of medical datasets where misclassifying minority samples can have significant repercussions. This study dedicated to mitigating by employing a hybrid approach that combines data-level, cost-sensitive, and ensemble methods. Through an assessment performance, measured AUC-ROC values, Sensitivity, F1-Score, G-Mean 20 data-level four cost-sensitive models on seventeen - 12 small five large, hybridized model, SMOTE-RF-CS-LR has been devised. model integrates Synthetic Minority Oversampling Technique (SMOTE), classifier Random Forest (RF), Cost-Sensitive Logistic Regression (CS-LR). Upon testing diverse imbalanced ratios, it demonstrated remarkable achieving outstanding performance values majority datasets. Further examination model's training duration time complexity revealed its efficiency, taking less than second train each dataset. Consequently, proposed not only proves be time-efficient but also exhibits robust capabilities handling imbalance, yielding results

Language: Английский

Deep Learning in Automatic Diabetic Retinopathy Detection and Grading Systems: A Comprehensive Survey and Comparison of Methods DOI Creative Commons
Israa Y. AbuShawish, Sudipta Modak, Esam Abdel‐Raheem

et al.

IEEE Access, Journal Year: 2024, Volume and Issue: 12, P. 84785 - 84802

Published: Jan. 1, 2024

Language: Английский

Citations

10
SDRG-Net: Integrating Multi-Level Color Transformation Encryption and ICNN-IRDO Feature Analysis for Robust Diabetic Retinopathy Diagnosis DOI Creative Commons

Venkata Kotam Raju Poranki,

B. Srinivasarao

e-Prime - Advances in Electrical Engineering Electronics and Energy, Journal Year: 2025, Volume and Issue: unknown, P. 100895 - 100895

Published: Jan. 1, 2025

Language: Английский

Citations

0
Grad-CAM based explanations for multiocular disease detection using Xception net DOI

M Raveenthini,

R Lavanya,

Raúl Benítez

et al.

Image and Vision Computing, Journal Year: 2025, Volume and Issue: unknown, P. 105419 - 105419

Published: Jan. 1, 2025

Language: Английский

Citations

0
Transfer learning approach to modeling multichannel gate-all-around nanosheet field-effect transistors under work function fluctuations DOI
Sagarika Dash, Yiming Li

Engineering Applications of Artificial Intelligence, Journal Year: 2025, Volume and Issue: 148, P. 110322 - 110322

Published: Feb. 28, 2025

Language: Английский

Citations

0
A precise image-based retinal blood vessel segmentation method using TAOD-CFNet DOI
Yixin Yang, Lixiang Sun, Zhiwen Tang

et al.

Biomedical Signal Processing and Control, Journal Year: 2025, Volume and Issue: 107, P. 107815 - 107815

Published: March 18, 2025

Language: Английский

Citations

0
Predicting Visual Acuity after Retinal Vein Occlusion Anti-VEGF Treatment: Development and Validation of an Interpretable Machine Learning Model DOI

Chunlan Liang,

Lian Liu,

Tong Zhao

et al.

Journal of Medical Systems, Journal Year: 2025, Volume and Issue: 49(1)

Published: April 29, 2025

Language: Английский

Citations

0
Co-CrackSegment: A New Corporative Deep Learning Framework for Pixel-Level Semantic Segmentation of Concrete Cracks DOI Open Access
Nizar Faisal Alkayem, Ali Mayya, Xin Zhang

et al.

Published: Aug. 27, 2024

In the era of massive construction, damaged and aging infrastructure are becoming more common. Defects, such as cracking, spalling, etc., main types structural damage that widely occur. Hence, ensuring safe operation existing through health monitoring has emerged an important challenge facing engineers. recent years, intelligent approaches, data driven machine deep learning crack detection, gradually dominate over traditional methods. Among them, semantic segmentation using models is a process characterization accurate location portrait cracks pixel level classification. Most available studies rely on single model knowledge to perform this task. However, it well-known might suffer from low variance ability generalize in case alteration. By leveraging ensemble philosophy, novel corporative concrete method called Co-CrackSegment proposed. Firstly, five models, namely U-net, SegNet, DeepCrack19, DeepLabV3-ResNet50, DeepLabV3-ResNet101 trained serve core for Co-CrackSegment. To build Co-CrackSegment, new iterative approach based best evaluation metrics, dice score, IoU, accuracy, precision, recall metrics developed. Results show exhibits prominent performance compared weighted average by means considered statistical metrics.

Language: Английский

Citations

2
Co-CrackSegment: A New Collaborative Deep Learning Framework for Pixel-Level Semantic Segmentation of Concrete Cracks DOI Creative Commons
Nizar Faisal Alkayem, Ali Mayya, Lei Shen

et al.

Mathematics, Journal Year: 2024, Volume and Issue: 12(19), P. 3105 - 3105

Published: Oct. 4, 2024

In an era of massive construction, damaged and aging infrastructure are becoming more common. Defects, such as cracking, spalling, etc., main types structural damage that widely occur. Hence, ensuring the safe operation existing through health monitoring has emerged important challenge facing engineers. recent years, intelligent approaches, data-driven machines deep learning crack detection have gradually dominated over traditional methods. Among them, semantic segmentation using models is a process characterization accurate locations portraits cracks pixel-level classification. Most available studies rely on single-model knowledge to perform this task. However, it well-known single model might suffer from low variance ability generalize in case data alteration. By leveraging ensemble philosophy, novel collaborative concrete method called Co-CrackSegment proposed. Firstly, five models, namely U-net, SegNet, DeepCrack19, DeepLabV3-ResNet50, DeepLabV3-ResNet101 trained serve core for Co-CrackSegment. To build Co-CrackSegment, new iterative approach based best evaluation metrics, Dice score, IoU, pixel accuracy, precision, recall metrics developed. Results show exhibits prominent performance compared with weighted average by means considered statistical metrics.

Language: Английский

Citations

2
A Systematic Review on Deep Learning Techniques for Diabetic Retinopathy Segmentation and Detection Using Ocular Imaging Modalities DOI
Richa Vij, Sakshi Arora

Wireless Personal Communications, Journal Year: 2024, Volume and Issue: 134(2), P. 1153 - 1229

Published: Jan. 1, 2024

Language: Английский

Citations

1
A Novel Classification Approach for Retinal Disease Using Improved Gannet Optimization‐Based Capsule DenseNet DOI

S. Venkatesan,

M. Kempanna,

J Nagaraja

et al.

International Journal of Imaging Systems and Technology, Journal Year: 2024, Volume and Issue: 34(5)

Published: Aug. 22, 2024

ABSTRACT An unusual condition of the eye called diabetic retinopathy affects human retina and is brought on by blood's constant rise in insulin levels. Loss vision result. Diabetic can be improved receiving an early diagnosis to prevent further damage. A cost‐effective method accumulating medical treatments through appropriate DR screening. In this work, deep learning framework introduced for accurate classification retinal diseases. The proposed processes fundus images obtained from databases, addressing noise artifacts median filter (ImMF). It leverages UNet++ model precise segmentation disease‐affected regions. enhances feature extraction cross‐stage connections, improving results. segmented are then fed as input gannet optimization‐based capsule DenseNet (IG‐CDNet) disease classification. hybrid (CDNet) classifies optimized using optimization algorithm boost accuracy. Finally, accuracy dice score values achieved 0.9917 0.9652 APTOS‐2019 dataset.

Language: Английский

Citations

0