Early detection of brain tumors in MRI images is vital for improving treatment results. However, deep learning models face challenges like limited dataset diversity, class imbalance, and insufficient interpretability. Most studies rely on small, single-source datasets and do not combine different feature extraction techniques for better classification. To address these challenges, we propose a robust and explainable stacking ensemble model for multiclass brain tumor classification. To address these challenges, we propose a stacking ensemble model that combines EfficientNetB0, MobileNetV2, GoogleNet, and Multi-level CapsuleNet, using CatBoost as the meta-learner for improved feature aggregation and classification accuracy. This ensemble approach captures complex tumor characteristics while enhancing robustness and interpretability. The proposed model integrates EfficientNetB0, MobileNetV2, GoogleNet, and a Multi-level CapsuleNet within a stacking framework, utilizing CatBoost as the meta-learner to improve feature aggregation and classification accuracy. We created two large MRI datasets by merging data from four sources: BraTS, Msoud, Br35H, and SARTAJ. To tackle class imbalance, we applied Borderline-SMOTE and data augmentation. We also utilized feature extraction methods, along with PCA and Gray Wolf Optimization (GWO). Our model was validated through confidence interval analysis and statistical tests, demonstrating superior performance. Error analysis revealed misclassification trends, and we assessed computational efficiency regarding inference speed and resource usage. The proposed ensemble achieved 97.81% F1 score and 98.75% PR AUC on M1, and 98.32% F1 score with 99.34% PR AUC on M2. Moreover, the model consistently surpassed state-of-the-art CNNs, Vision Transformers, and other ensemble methods in classifying brain tumors across individual four datasets. Finally, we developed a web-based diagnostic tool that enables clinicians to interact with the proposed model and visualize decision-critical regions in MRI scans using Explainable Artificial Intelligence (XAI). This study connects high-performing AI models with real clinical applications, providing a reliable, scalable, and efficient diagnostic solution for brain tumor classification.

Early and accurate detection of breast cancer is essential for reducing mortality rates and improving clinical outcomes. However, deep learning (DL) models used in healthcare face significant challenges, including concerns about data privacy, domain-specific overfitting, and limited interpretability. To address these issues, we propose BreastSwinFedNetX, a federated learning (FL)-enabled ensemble system that combines four hierarchical variants of the Swin Transformer (Tiny, Small, Base, and Large) with a Random Forest (RF) meta-learner. By utilizing FL, our approach ensures collaborative model training across decentralized and institution-specific datasets while preserving data locality and preventing raw patient data exposure. The model exhibits strong generalization and performs exceptionally well across five benchmark datasets—BreakHis, BUSI, INbreast, CBIS-DDSM, and a Combined dataset—achieving an F1 score of 99.34% on BreakHis, a PR AUC of 98.89% on INbreast, and a Matthews Correlation Coefficient (MCC) of 99.61% on the Combined dataset. To enhance transparency and clinical adoption, we incorporate explainable AI (XAI) through Grad-CAM, which highlights class-discriminative features. Additionally, we deploy the model in a real-time web application that supports uncertainty-aware predictions and clinician interaction and ensures compliance with GDPR and HIPAA through secure federated deployment. Extensive ablation studies and paired statistical analyses further confirm the significance and robustness of each architectural component. By integrating transformer-based architectures, secure collaborative training, and explainable outputs, BreastSwinFedNetX provides a scalable and trustworthy AI solution for real-world breast cancer diagnostics.

Tomatoes are essential fruits in numerous nations for their vast demand. It is very important to maintain the freshness of tomatoes. One of the primary challenges in the recent culinary landscape is accurately identifying healthy tomatoes while effectively eliminating damaged or rejected ones. Existing approaches employ various strategies for categorizing tomato fruit, but they often suffer from inaccuracies, slow detection, and suboptimal performance. Thus, motivated by this gap, in this paper, we propose a novel machine learning (ML) framework, ViT-SENet-Tom, which is a hybrid vision transformer (ViT) model with squeeze and excitation (SENet) block network for fast, accurate, and efficient tomato fruit classification. The framework works on three tomato classes, respectively, the ripe, unripe, and reject. In developing the proposed model, we utilized advanced and newly designed layers and functions. This integration created a more complex and sophisticated neural network, significantly enhancing efficiency and contributing to the model’s novelty. Our chosen dataset was small initially, but we implemented augmentation techniques to increase its size. This approach made our system more reliable, efficient, and effective. The hybrid ViT-SENet framework employs encoders and self-attention networks with squeeze and excitation channel functions to allow precise, robust, fast, and efficient tomato classification. In simulation, the framework achieves a training accuracy of 99.87% and validation accuracy of 93.87%, indicating the precise classification of tomatoes. Besides, this work tests accuracy using fivefold cross-validation. The highest accuracy seen at fold-5 is 99.90%. These testing results demonstrate the efficacy of the proposed framework in real-deployment scenarios. The implementation has the potential to provide enhanced and more sustainable food security and safety in future.

Glioblastoma is the most common adult brain tumor, significantly impacts disability and mortality. Early and accurate diagnosis of glioma subtypes is essential, but manual categorization is challenging due to their complexity, prompting the need for automated solutions. We developed an innovative mixed convolution‐transformer model to classify glioma subtypes, including astrocytoma, glioblastoma, oligodendroglioma, and normal brain tissue, using whole slide images. The novelty of this model lies in its remarkable efficiency and precise results. Multiple advanced and complex layers are incorporated during its development to enhance its performance, ensuring that it delivers fast and accurate classification results for glioma. This proposed model obtains an overall training accuracy of 97.41%, peaking at 98.12% for validation and 97.35% for testing. Next, our model architecture is independently evaluated by comparing its training performances on the CIFAR‐10 and CIFAR‐100 datasets with the vision transformer and compact convolutional transformer models. Results across various datasets demonstrate that the model consistently outperforms existing models. This performance underscores the effectiveness of our proposed approach in classifying glioma subtypes accurately and efficiently, highlighting its potential impact on healthcare and disability. This system enhances the classification of glioma subtypes and facilitates swift identification, ensuring appropriate and timely treatment.

Sleep is a crucial period of rest necessary for optimal cognitive function, psychological well-being, and execution of everyday tasks. In the field of sleep healthcare, the primary objective is to identify and classify the various sleep states. Implementing sleep state detection in a system is problematic and essential for accurate diagnosis. Our study used an integrated framework to recognize sleep states. The dataset contained approximately eight lakh data points sorted into two groups: onset and wake-up. We successfully deployed a cutting-edge Convolutional Recurrent Transformer (CRT) model for sleep state detection. The training accuracy of our detection model was measured at 97.83%, a constant validation accuracy of 97.07%, and a testing measurement accuracy of 97.23%, were maintained. These scores indicate the model's proficiency in precisely recognizing the sleep states. Our system's detection capabilities demonstrate the ability to identify different sleep states, enhance the accuracy of diagnoses and increase healthcare outcomes in this specialized field.