Medicinal plants are essential to healthcare since ancient times and are integral to developing drugs and other medical treatments. More than 25% of medicines in developed countries are produced from medicinal plants, while in developing countries, approximately 80% of individuals receive primary healthcare from these plants. Traditionally, these plants are identified manually by experts, which is tedious, time-consuming, subjective and dependent on the availability of experts. Furthermore, a wrong detection can result in serious health issues or death. This signifies the need for a more reliable approach to identifying medicinal plants, which is accurate and practical. Several automated methods were proposed previously, utilizing deep learning and traditional machine learning (TML) techniques, but they require singular leaf images and failed to achieve sufficient accuracy when demonstrated in a different setting. Capturing singular leaf images for each plant is also time-consuming and laborious. This paper presents a robust, accurate and practical system to identify medicinal plants from smartphone-captured plant images in the site of plants. The proposed system utilized a cascaded architecture to extract features using a pre-trained ResNet50 model, which were optimized using Particle Swarm Optimization (PSO) to classify the plants using a Support Vector Machine (SVM). The proposed ResNet50-PSO-SVM network classified seven medicinal plants with 99.60% accuracy, outperforming the state-of-the-art (99%). The system was demonstrated for three different smartphones, classifying an image in 0.15 seconds with 97.79% accuracy on average. The system’s high accuracy, rapid identification time and robustness ensured its practical use.

Diabetic foot ulcer (DFU) classification increasingly relies on learning frameworks that exploit medical image representations without centralizing raw data. This study presents FedQDFU, a feature-based federated quantum learning benchmark evaluating a compact variational quantum classifier (VQC) on deep features from VGG16, ResNet50, and DenseNet121, using frozen ImageNet-pretrained backbones as fixed feature extractors over 1020 samples from a single public DFU dataset (696 unique feature vectors after exact-duplicate auditing). Because patient identifiers are unavailable, duplicate-group-aware splitting prevents exact duplicates from crossing the development-test boundary but cannot guarantee patient-level independence. A 6-qubit VQC (43 parameters) is compared with a 129-parameter multilayer perceptron (MLP) under centralized learning, FedAvg, and FedProx across IID and Dirichlet non-IID client distributions ($\alpha \in \{10,1,0.5,0.1\}$), with federated-compatible random projection, five-seed evaluation of the primary conditions, and bootstrap confidence intervals on a locked test set. At the primary non-IID setting ($\alpha =0.5$), the VQC attains a lower mean Matthews correlation coefficient (MCC) than the MLP on all three backbones (e.g., ResNet50: $0.782\pm 0.011$ vs. $0.650\pm 0.041$ under FedAvg); the paired bootstrap difference is significant only for ResNet50 at the uncorrected pairwise 95% level ($\Delta \mathrm{MCC}=-0.115$, 95% CI $[-0.218,-0.016]$), an effect a Bonferroni correction across backbones is estimated to remove, so this result should be read as suggestive rather than confirmed. No significant centralized-to-FedAvg VQC degradation is detected for any backbone, and a ResNet50 parameter-matched experiment shows the MLP-VQC gap persists even when the MLP is reduced to 41 parameters. Under severe non-IID heterogeneity ($\alpha =0.1$), the seed-42 partition contains one entirely single-class client (mean pairwise Jensen–Shannon divergence 0.5653 vs. 0.1787 at $\alpha =0.5$); federated aggregation of this degenerate update yields a fixed-threshold collapse in which all test samples are assigned the positive class (MCC = 0, sensitivity = 1, specificity = 0). This outcome is realization-dependent rather than deterministic: only one of five paired partition/training seeds reproduces the complete collapse.

ObjectivesBrain tumor classification using magnetic resonance imaging (MRI) is routinely performed to guide treatment deci-sions, improve patient outcomes, and optimize healthcare resources. However, conventional manual examination is time-consuming, tedious, and subject to inter-observer variability.MethodsSeveral approaches based on deep learning (DL) and traditional machine learning (ML) have been proposed to auto-mate diagnosis, but they often failed to achieve adequate accuracy. In this study, we proposed a hybrid method that incorporates deep learning-based feature engineering with a traditional machine learning classifier.ResultsThe proposed Stacked Network Model (SNM) achieved outstanding precision and high classification accuracy, reach-ing 99.81% using the ResNet50-BFO-LR pipeline, with an average inference time of 0.11 seconds per image. These results outperform recent state-of-the-art approaches.ConclusionsThe findings highlight the potential of combining optimized deep features with classical classifiers to provide fast and accurate brain tumor diagnosis.

Stroke is one of the leading causes of death and long-term disability worldwide, and effective prevention depends on fast, reliable, and privacy-preserving risk assessment. This study proposes a federated IoT-enabled framework that combines feature-optimized machine learning (ML) with real-time patient monitoring to predict and detect brain stroke risk. The system operates in two stages: (i) a stroke prediction module that builds an ML model for risk assessment and (ii) an IoT-based framework that continuously monitors patients and triggers timely alerts. The ML pipeline starts from a clinical–physiological dataset containing 17 initial attributes and applies a feature optimization strategy based on feature importance, selection, and reduction to identify the most informative predictors of stroke. To support multi-center deployment while protecting patient confidentiality, the ML pipeline is embedded within a standard Federated Averaging (FedAvg) architecture, where multiple home or hospital IoT gateways collaboratively train a shared global model without exchanging raw patient data. In each communication round, clients perform local training and the server aggregates client model parameters to update the global model. The resulting federated global model matches the performance of the centralized baseline, achieving 99.44% test accuracy while preserving data locality. Integrated with IoT devices, the system can detect pre-stroke syndromes in real time and automatically notify family members or emergency medical services, making it suitable for both home and hospital environments and offering a practical path toward early intervention and improved stroke outcomes.