ContourTL-Net: Contour-Based Transfer Learning Algorithm for Early-Stage Brain Tumor Detection

International Journal of Biomedical Imaging, Год журнала: 2024, Номер 2024, С. 1 - 20

Опубликована: Апрель 29, 2024

Brain tumors are critical neurological ailments caused by uncontrolled cell growth in the brain or skull, often leading to death. An increasing patient longevity rate requires prompt detection; however, complexities of tissue make early diagnosis challenging. Hence, automated tools necessary aid healthcare professionals. This study is particularly aimed at improving efficacy computerized tumor detection a clinical setting through deep learning model. novel thresholding-based MRI image segmentation approach with transfer model based on contour (ContourTL-Net) suggested facilitate malignancies an initial phase. The utilizes contour-based analysis, which for object detection, precise segmentation, and capturing subtle variations morphology. employs VGG-16 architecture priorly trained “ImageNet” collection feature extraction categorization. designed utilize its ten nontrainable three trainable convolutional layers dropout layers. proposed ContourTL-Net evaluated two benchmark datasets four ways, among unseen case considered as aspect. Validating data crucial determine model’s generalization capability, domain adaptation, robustness, real-world applicability. Here, presented outcomes demonstrate highly accurate classification data, achieving perfect sensitivity negative predictive value (NPV) 100%, 98.60% specificity, 99.12% precision, 99.56% F1 -score, 99.46% accuracy. Additionally, compared state-of-the-art methodologies further enhance effectiveness. solution outperforms existing solutions both seen potential significantly improve efficiency accuracy, earlier diagnoses improved outcomes.

Язык: Английский

---

Development of brain tumor radiogenomic classification using GAN-based augmentation of MRI slices in the newly released gazi brains dataset

BMC Medical Informatics and Decision Making, Год журнала: 2024, Номер 24(1)

Опубликована: Окт. 4, 2024

Язык: Английский

---

Innovative fusion of VGG16, MobileNet, EfficientNet, AlexNet, and ResNet50 for MRI-based brain tumor identification

Iran Journal of Computer Science, Год журнала: 2024, Номер unknown

Опубликована: Дек. 9, 2024

Язык: Английский

---

A privacy-preserved horizontal federated learning for malignant glioma tumour detection using distributed data-silos

PLoS ONE, Год журнала: 2025, Номер 20(2), С. e0316543 - e0316543

Опубликована: Фев. 11, 2025

Malignant glioma is the uncontrollable growth of cells in spinal cord and brain that look similar to normal glial cells. The most essential part nervous system cells, which support brain’s functioning prominently. However, with evolution glioma, tumours form invade healthy tissues brain, leading neurological impairment, seizures, hormonal dysregulation, venous thromboembolism. Medical tests, including medical resonance imaging (MRI), computed tomography (CT) scans, biopsy, electroencephalograms are used for early detection glioma. these tests expensive may cause irritation allergic reactions due ionizing radiation. deep learning models highly optimal disease prediction, however, challenge associated it requirement substantial memory storage amalgamate patient’s information at a centralized location. Additionally, also has patient data-privacy concerns anonymous generalization, regulatory compliance issues, data leakage challenges. Therefore, proposed work, distributed privacy-preserved horizontal federated learning-based malignant model been developed by employing 5 10 different clients’ architectures independent identically (IID) non-IID distributions. Initially, developing this model, collection MRI scans non-tumour done, further pre-processed performing balancing image resizing. configuration development pre-trained MobileNetV2 base have performed, then applied learning(FL) framework. configurations kept as 0.001, Adam, 32, 10, FedAVG, rate, optimizer, batch size, local epochs, global aggregation, rounds, respectively. provided prominent accuracy architecture 99.76% 99.71% IID distributions, These outcomes demonstrate optimized generalizes improved when compared state-of-the-art models.

Язык: Английский

---

Enhanced glioma tumor detection and segmentation using modified deep learning with edge fusion and frequency features

Scientific Reports, Год журнала: 2025, Номер 15(1)

Опубликована: Фев. 26, 2025

Computer-aided automatic brain tumor detection is crucial for timely diagnosis and treatment, especially in regions with limited access to medical expertise. However, existing methods often overlook edge pixel information during segmentation, leading reduced boundary accuracy, achieve high performance primarily on highly enhanced images, making them less effective enhancement-lagging clinical data. To address these gaps, this study proposes the Edge Incorporative Fusion (EIF) algorithm, which enhances edge-pixel contrast MRI combined Gabor Transform (GaT) spatial-frequency domain conversion improve accuracy. These innovations integrated into a Modified Deep Learning (MDL) architecture that reduces time by optimizing internal layers while maintaining superior classification performance. The EIF-MDL system developed using Python programming language implemented Jupyter Notebook flexibility reproducibility. was evaluated PLCO NU datasets, achieving sensitivity of 98.58%, specificity 99.09%, accuracy 99.1%, Dice similarity coefficient 98.96%, outperforming state-of-the-art methods. This robust system's ability excel across both images highlights its potential accurate glioma diagnosis, particularly resource-constrained healthcare environments.

Язык: Английский

---

Advanced Brain Tumor MR Image Classification Using a Combination Undecimated Wavelet Transform, EfficientNet-B0 and PCA via Multi-SVM Analysis

Iranian Journal of Science and Technology Transactions of Electrical Engineering, Год журнала: 2025, Номер unknown

Опубликована: Март 12, 2025

Язык: Английский

---

Federated and Centralized Machine Learning for Cell Segmentation: A Comparative Analysis

Electronics, Год журнала: 2025, Номер 14(7), С. 1254 - 1254

Опубликована: Март 22, 2025

The automatic segmentation of cell images plays a critical role in medicine and biology, as it enables faster more accurate analysis diagnosis. Traditional machine learning faces challenges since requires transferring sensitive data from laboratories to the cloud, with possible risks limitations due patients’ privacy, data-sharing regulations, or laboratory privacy guidelines. Federated addresses issues by introducing decentralized approach that removes need for laboratories’ sharing. task is divided among participating clients, each training global model situated on cloud its local dataset. This guarantees only transmitting updated weights cloud. In this study, centralized compared federated one, demonstrating they achieve similar performances. Stemming benchmarking available models, Cellpose, having shown better recall precision (F1=0.84) than U-Net (F1=0.50) StarDist (F1=0.12), was used baseline testbench implementation. results show both binary multi-class metrics remain high when employing solution (F1=0.86) (F12clients=0.86). These were also stable across an increasing number clients reduced samples (F14clients=0.87, F116clients=0.86), proving effectiveness central aggregation locally trained models.

Язык: Английский

---

A robust and personalized privacy-preserving approach for adaptive clustered federated distillation

Scientific Reports, Год журнала: 2025, Номер 15(1)

Опубликована: Апрель 23, 2025

Abstract Federated learning (FL) is a promising approach that addresses privacy, and scalability concerns in contrast to traditional centralized methods. Challenges such as personalization data heterogeneity issues remain critical. Clustered federated (CFL) has been proposed alleviate these by establishing specialized global models for sets of similar users. Although CFL enhances adaptability highly statistically heterogeneous environments, it may suffer from real-time distribution changes due limitations fixed cluster configurations. This study presents the robust model personalized distillation (RMPFD), privacy-enhanced framework. The RMPFD framework employs an adaptive hierarchical clustering strategy generate semi-global grouping clients with distributions, allowing them train independently. Meta-learning used each enhance local classification accuracy non-independent Identically distributed (non-IID) distributions. Experimental evaluations conducted on CIFAR- 10, 100, Fashion-MNIST Enron email datasets reveal reduces communication overhead approximately 15% 20%, compared Averaging (FedAvg) other baseline techniques. Moreover, improves convergence rates accuracy, leading improvement over 12% performance FL

Язык: Английский

---

Real-Time Detection of Meningiomas by Image Segmentation: A Very Deep Transfer Learning Convolutional Neural Network Approach

Tomography, Год журнала: 2025, Номер 11(5), С. 50 - 50

Опубликована: Апрель 24, 2025

Developing a treatment strategy that effectively prolongs the lives of people with brain tumors requires an accurate diagnosis condition. Therefore, improving preoperative classification meningiomas is priority. Machine learning (ML) has made great strides thanks to development convolutional neural networks (CNNs) and computer-aided tumor detection systems. The deep layers automatically extract important dependable information from input space, in contrast more traditional network layers. One recent promising advancement this field ML. Still, there dearth studies being carried out area. starting analysis magnetic resonance images, we have suggested research work tried-and-tested methodical for real-time meningioma by image segmentation using very transfer CNN model or DNN (VGG-16) CUDA. Since VGGNet greater level accuracy than other models like AlexNet, GoogleNet, etc., chosen employ it. VGG constructed small filters consists 13 3 fully connected Our takes sMRI FLAIR input. VGG's leverage minimal receptive field, i.e., × 3, smallest possible size still captures up/down left/right. Moreover, are also 1 convolution acting as linear transformation This followed ReLU unit. stride fixed at pixel keep spatial resolution preserved after convolution. All hidden our use ReLU. A dataset consisting 264 3D segments three different classes (meningioma, tuberculoma, normal) was employed. number epochs Sequential Model set 10. Keras used were Dense, Dropout, Flatten, Batch Normalization, According simulation findings, successfully classified all data used, 99.0% overall accuracy. performance metrics implemented confusion matrix indicate model's high classification. good outcomes demonstrate possibility method useful diagnostic tool, promoting better understanding, prognostic tool clinical outcomes, efficient planning tool. It demonstrated several computed previously good. Consequently, think approach way identify tumors.

Язык: Английский

---

Challenges, optimization strategies, and future horizons of advanced deep learning approaches for brain lesion segmentation

Methods, Год журнала: 2025, Номер unknown

Опубликована: Апрель 1, 2025

Язык: Английский

---