Bio-inspired feature selection for early diagnosis of Parkinson’s disease through optimization of deep 3D nested learning DOI Creative Commons

S. Priyadharshini,

K. Ramkumar,

V. Subramaniyaswamy

et al.

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: Oct. 8, 2024

Parkinson's disease (PD) is one of the most common neurodegenerative disorders that affect quality human life millions people throughout world. The probability getting affected by this increases with age, and it among elderly population. Early detection can help in initiating medications at an earlier stage. It significantly slow down progression disease, assisting patient to maintain a good for more extended period. Magnetic resonance imaging (MRI)-based brain area active research used diagnose PD early understand key biomarkers. prior investigations using MRI data mainly focus on volume, structural, morphological changes basal ganglia (BG) region diagnosing PD. Recently, researchers have emphasized significance studying other areas comprehensive understanding also analyze happening tissue. Thus, perform accurate diagnosis treatment planning identification PD, work focuses learning onset from images taken whole-brain novel 3D-convolutional neural network (3D-CNN) deep architecture. conventional 3D-Resent model, after various hyper-parameter tuning architectural changes, has achieved accuracy 90%. In work, 3D-CNN architecture was developed, several ablation studies, model yielded results improved 93.4%. Combining features 3D ResNet models Canonical Correlation Analysis (CCA) resulted 95% accuracy. For further enhancements performance, feature fusion optimization employed, utilizing techniques. Whale based biologically inspired approach selected basis convergence diagram. performance compared methods given 97%. This represents critical advancement improving techniques emphasizing importance nested bio-inspired selection.

Language: Английский

Advanced ellipse overlap computation based on segment area of circles DOI Creative Commons
Minhye Kim,

Yongkuk Kim,

Giphil Cho

et al.

Alexandria Engineering Journal, Journal Year: 2025, Volume and Issue: 119, P. 425 - 436

Published: Feb. 7, 2025

Language: Английский

Citations

0
Explainable AI supported hybrid deep learnig method for layer 2 intrusion detection DOI
Ilhan Firat Kilinçer

Egyptian Informatics Journal, Journal Year: 2025, Volume and Issue: 30, P. 100669 - 100669

Published: March 23, 2025

Language: Английский

Citations

0
FCN-PD: An Advanced Deep Learning Framework for Parkinson’s Disease Diagnosis Using MRI Data DOI Creative Commons

Manal Alrawis,

Farah Mohammad, Saad Al-Ahmadi

et al.

Diagnostics, Journal Year: 2025, Volume and Issue: 15(8), P. 992 - 992

Published: April 14, 2025

Background/Objectives: Parkinson's disease (PD) is a progressive neurodegenerative disorder characterized by motor dysfunction, cognitive decline, and diminished quality of life. Early accurate diagnosis essential for effective management. However, traditional diagnostic approaches, which rely on clinical observations subjective assessments, often lead to delays inaccuracies. This research aims address these limitations proposing FCN-PD, an advanced deep learning framework PD using MRI data. Methods: The FCN-PD incorporates hybrid feature extraction phase that combines EfficientNet capture local spatial details attention mechanisms extract global contextual information. These features are then processed Fully Connected Network (FCN) final classification. architecture enables the model effectively represent hierarchical handle high-dimensional data while mitigating issues such as overfitting redundancy. Results: performance was evaluated three publicly available datasets. On PPMI dataset, it achieved accuracy 97.2%, outperforming CNN-based models 5.3%. OASIS 95.6% accuracy, MIRIAD reached 96.8% accuracy. results establish superior alternative existing methods. Conclusions: demonstrates significant improvements in efficiency Its robust captures both features, making promising tool integration early detection, ultimately contributing better patient outcomes.

Language: Английский

Citations

0
A multimodal multistream multilevel fusion network for finger joint angle estimation with hybrid sEMG and FMG sensing DOI Creative Commons
Zhouping Chen, Mohamed Amin Gouda,

Longcheng Ji

et al.

Alexandria Engineering Journal, Journal Year: 2024, Volume and Issue: 110, P. 9 - 23

Published: Oct. 5, 2024

Language: Английский

Citations

1
Bio-inspired feature selection for early diagnosis of Parkinson’s disease through optimization of deep 3D nested learning DOI Creative Commons

S. Priyadharshini,

K. Ramkumar,

V. Subramaniyaswamy

et al.

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: Oct. 8, 2024

Parkinson's disease (PD) is one of the most common neurodegenerative disorders that affect quality human life millions people throughout world. The probability getting affected by this increases with age, and it among elderly population. Early detection can help in initiating medications at an earlier stage. It significantly slow down progression disease, assisting patient to maintain a good for more extended period. Magnetic resonance imaging (MRI)-based brain area active research used diagnose PD early understand key biomarkers. prior investigations using MRI data mainly focus on volume, structural, morphological changes basal ganglia (BG) region diagnosing PD. Recently, researchers have emphasized significance studying other areas comprehensive understanding also analyze happening tissue. Thus, perform accurate diagnosis treatment planning identification PD, work focuses learning onset from images taken whole-brain novel 3D-convolutional neural network (3D-CNN) deep architecture. conventional 3D-Resent model, after various hyper-parameter tuning architectural changes, has achieved accuracy 90%. In work, 3D-CNN architecture was developed, several ablation studies, model yielded results improved 93.4%. Combining features 3D ResNet models Canonical Correlation Analysis (CCA) resulted 95% accuracy. For further enhancements performance, feature fusion optimization employed, utilizing techniques. Whale based biologically inspired approach selected basis convergence diagram. performance compared methods given 97%. This represents critical advancement improving techniques emphasizing importance nested bio-inspired selection.

Language: Английский

Citations

0