Internet of Things (IoT) technology has recently been applied in healthcare systems as an Internet of Medical Things (IoMT) to collect sensor information for the diagnosis and prognosis of heart disease. The main objective of the proposed research is to classify data and predict heart disease using medical data and medical images. The proposed model is a medical data classification and prediction model that operates in two stages. If the result from the first stage is efficient in predicting heart disease, there is no need for stage two. In the first stage, data gathered from medical sensors affixed to the patient’s body were classified; then, in stage two, echocardiogram image classification was performed for heart disease prediction. A hybrid linear discriminant analysis with the modified ant lion optimization (HLDA-MALO) technique was used for sensor data classification, while a hybrid Faster R-CNN with SE-ResNet-101 modelwass used for echocardiogram image classification. Both classification methods were carried out, and the classification findings were consolidated and validated to predict heart disease. The HLDA-MALO method obtained 96.85% accuracy in detecting normal sensor data, and 98.31% accuracy in detecting abnormal sensor data. The proposed hybrid Faster R-CNN with SE-ResNeXt-101 transfer learning model performed better in classifying echocardiogram images, with 98.06% precision, 98.95% recall, 96.32% specificity, a 99.02% F-score, and maximum accuracy of 99.15%.

Diabetes is a growing global health concern, affecting millions and leading to severe complications if not properly managed. The primary challenge in diabetes management is maintaining blood glucose levels (BGLs) within a safe range to prevent complications such as renal failure, cardiovascular disease, and neuropathy. Traditional methods, such as finger-prick testing, often result in low patient adherence due to discomfort, invasiveness, and inconvenience. Consequently, there is an increasing need for non-invasive techniques that provide accurate BGL measurements. Photoplethysmography (PPG), a photosensitive method that detects blood volume variations, has shown promise for non-invasive glucose monitoring. Deep neural networks (DNNs) applied to PPG signals can predict BGLs with high accuracy. However, training DNN models requires large and diverse datasets, which are typically distributed across multiple healthcare institutions. Privacy concerns and regulatory restrictions further limit data sharing, making conventional centralized machine learning (ML) approaches less effective. To address these challenges, this study proposes a federated learning (FL)-based solution that enables multiple healthcare organizations to collaboratively train a global model without sharing raw patient data, thereby enhancing model performance while ensuring data privacy and security. In the data preprocessing stage, continuous wavelet transform (CWT) is applied to smooth PPG signals and remove baseline drift. Adaptive cycle-based segmentation (ACBS) is then used for signal segmentation, followed by particle swarm optimization (PSO) for feature selection, optimizing classification accuracy. The proposed system was evaluated on diverse datasets, including VitalDB and MUST, under various conditions with data collected during surgery and anesthesia. The model achieved a root mean square error (RMSE) of 19.1 mg/dL, demonstrating superior predictive accuracy. Clarke error grid analysis (CEGA) confirmed the model’s clinical reliability, with 99.31% of predictions falling within clinically acceptable limits. The FL-based approach outperformed conventional deep learning models, making it a promising method for non-invasive, privacy-preserving glucose monitoring.

Breast cancer (BrCa) is the most common disease in women worldwide. Classifying the BrCa image is extremely important for finding BrCa at an earlier stage and monitoring BrCa during treatment. The computer‐aided detection methods have been used to interpret BrCa and improve the detection of BrCa during the screening and treatment stages. However, if a new BrCa image is generated for the treatment, it will not classify correctly. The main objective of this research is to classify the BrCa images for newly generated images. The model performs preprocessing, segmentation, feature extraction, and classification. In preprocessing, a hybrid median filtering (HMF) is used to eliminate the noise in the images. The contrast of the images is enhanced using quadrant dynamic histogram equalization (QDHE). Then, ROI segmentation is performed using the USE‐Net deep learning model. The CaffeNet model is used for feature extraction on the segmented images, and finally, classification is made using the improved random forest (IRF) with extreme gradient boosting (XGB). The model obtained 97.87% accuracy, 98.45% sensitivity, 95.24% specificity, 98.96% precision, and 98.70% f1‐score for ultrasound images. The model gives 98.31% accuracy, 99.29% sensitivity, 90.20% specificity, 98.82% precision, and 99.05% f1‐score for mammogram images.

Due to the limited light penetration in underwater environments, sonar equipment plays a crucial role in various commercial and military operations. However, underwater images often suffer from degradation due to scattering and absorption phenomena, resulting in poor visibility of submerged objects. To address this challenge, image enhancement techniques are essential for enhancing the appearance and visibility of underwater objects. This research proposes a novel approach called HLAST-ACNet, which combines the advantages of a hybrid Local Acuity Swin Transformer and an Adapted Coat-Net for Underwater Object Detection (UOD). The HLASwin-T-ACoat-Net leverages Contrast Limited Adaptive Histogram Equalization (CLAHE) to increase the quality of images. Additionally, it incorporates a path aggregation network to integrate deep and shallow feature maps and utilizes online complicated example mining to improve training efficiency. Furthermore, the algorithm improves Region of Interest (ROI) pooling by introducing ROI alignment, which mitigates quantization errors and enhances object detection accuracy. Compared to existing algorithms, the algorithms based on HLASTACNet demonstrate significant improvements in the URPC2018 and OUC datasets, achieving precision rates of 91.25% and 92.36%, respectively. The research model has a higher computational complexity than four existing methods, as evidenced by its GFLOPs, per-image processing time with a speed of 20ms, and the FPS measures for average processed frames per second reaching 2.28s. The research model effectively addressed the challenges and false detection with varying sizes of objects in complicated underwater environments.