A hybrid explainable model based on advanced machine learning and deep learning models for classifying brain tumors using MRI images

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: Jan. 10, 2025

Brain tumors present a significant global health challenge, and their early detection accurate classification are crucial for effective treatment strategies. This study presents novel approach combining lightweight parallel depthwise separable convolutional neural network (PDSCNN) hybrid ridge regression extreme learning machine (RRELM) accurately classifying four types of brain (glioma, meningioma, no tumor, pituitary) based on MRI images. The proposed enhances the visibility clarity tumor features in images by employing contrast-limited adaptive histogram equalization (CLAHE). A PDSCNN is then employed to extract relevant tumor-specific patterns while minimizing computational complexity. RRELM model proposed, enhancing traditional ELM improved performance. framework compared with various state-of-the-art models terms accuracy, parameters, layer sizes. achieved remarkable average precision, recall, accuracy values 99.35%, 99.30%, 99.22%, respectively, through five-fold cross-validation. PDSCNN-RRELM outperformed pseudoinverse (PELM) exhibited superior introduction led enhancements performance parameters sizes those models. Additionally, interpretability was demonstrated using Shapley Additive Explanations (SHAP), providing insights into decision-making process increasing confidence real-world diagnosis.

Language: Английский

---

PMFSNet: Polarized multi-scale feature self-attention network for lightweight medical image segmentation

Computer Methods and Programs in Biomedicine, Journal Year: 2025, Volume and Issue: unknown, P. 108611 - 108611

Published: Jan. 1, 2025

Language: Английский

---

NM-USNet: A novel generative model for parathyroid glands detection in nuclear medicine

Biomedical Signal Processing and Control, Journal Year: 2025, Volume and Issue: 104, P. 107493 - 107493

Published: Jan. 18, 2025

Language: Английский

---

Heatmap-guided balanced multi-task learning approach for glistening characterization in OCT images

Biomedical Signal Processing and Control, Journal Year: 2025, Volume and Issue: 104, P. 107527 - 107527

Published: Jan. 26, 2025

Language: Английский

---

Takagi–Sugeno–Kang Fuzzy Neural Network for Nonlinear Chaotic Systems and Its Utilization in Secure Medical Image Encryption

Mathematics, Journal Year: 2025, Volume and Issue: 13(6), P. 923 - 923

Published: March 11, 2025

This study introduces a novel control framework based on the Takagi–Sugeno–Kang wavelet fuzzy neural network, integrating brain imitated network and cerebellar network. The proposed controller demonstrates high robustness, making it an excellent candidate for handling intricate nonlinear dynamics, effectively mapping input–output relationships efficiently learning from data. To enhance its performance, controller’s parameters are fine-tuned using Lyapunov stability theory. Compared to existing approaches, model exhibits superior capabilities achieves outstanding performance metrics. Furthermore, applies this synchronization technique secure transmission of medical images. By encrypting image into chaotic trajectory before transmission, system ensures data security. On receiving end, original is successfully reconstructed synchronization. Experimental results confirm effectiveness reliability model, as well encryption decryption process. Specifically, average_RMSE cerebral (TFWBCC) method 2.004 times smaller than articulation (CMAC) method, 1.923 RCMAC 1.8829 TSKCMAC 1.8153 emotional (BELC) method.

Language: Английский

---

Modal Feature Supplementation Enhances Brain Tumor Segmentation

International Journal of Imaging Systems and Technology, Journal Year: 2025, Volume and Issue: 35(3)

Published: April 3, 2025

ABSTRACT For patients with brain tumors, effectively utilizing the complementary information between multimodal medical images is crucial for accurate lesion segmentation. However, features across different modalities remains a challenging task. To address these challenges, we propose modal feature supplement network (MFSNet), which extracts modality simultaneously using both main and an auxiliary network. During this process, supplements of network, enabling tumor We also design enhancement module (MFEM), cross‐layer fusion (CFFM), edge (EFSM). MFEM enhances performance by fusing from networks. CFFM additional contextual adjacent encoding layers at scales, are then passed into corresponding decoding layers. This aids in preserving more details during upsampling. EFSM improves deformable convolution to extract boundary features, used final output layer. evaluated MFSNet on BraTS2018 BraTS2021 datasets. The Dice scores whole tumor, core, enhancing regions were 90.86%, 90.59%, 84.72%, 92.28%, 92.47%, 86.07%, respectively. validates accuracy segmentation, demonstrating its superiority over other networks similar type.

Language: Английский

---

Synthetic Data Generation via Generative Adversarial Networks in Healthcare: A Systematic Review of Image- and Signal-Based Studies

IEEE Open Journal of Engineering in Medicine and Biology, Journal Year: 2024, Volume and Issue: 6, P. 183 - 192

Published: Nov. 28, 2024

Generative Adversarial Networks (GANs) have emerged as a powerful tool in artificial intelligence, particularly for unsupervised learning. This systematic review analyzes GAN applications healthcare, focusing on image and signal-based studies across various clinical domains. Following Preferred Reporting Items Systematic reviews Meta-Analyses (PRISMA) guidelines, we reviewed 72 relevant journal articles. Our findings reveal that magnetic resonance imaging (MRI) electrocardiogram (ECG) signal acquisition techniques were most utilized, with brain (22%), cardiology (18%), cancer (15%), ophthalmology (12%), lung (10%) being the researched areas. We discuss key architectures, including cGAN (31%) CycleGAN along datasets, evaluation metrics, performance outcomes. The highlights promising data augmentation, anonymization, multi-task learning results. identify current limitations, such lack of standardized metrics direct comparisons, propose future directions, development no-reference immersive simulation scenarios, enhanced interpretability.

Language: Английский

---