Brain tumor classification is essential for clinical diagnosis and treatment planning. Deep learning models have shown great promise in this task, but they are often challenged by the complex and diverse nature of brain tumors. To address this challenge, we propose a novel deep residual and region-based convolutional neural network (CNN) architecture, called Res-BRNet, for brain tumor classification using magnetic resonance imaging (MRI) scans. Res-BRNet employs a systematic combination of regional and boundary-based operations within modified spatial and residual blocks. The spatial blocks extract homogeneity, heterogeneity, and boundary-related features of brain tumors, while the residual blocks significantly capture local and global texture variations. We evaluated the performance of Res-BRNet on a challenging dataset collected from Kaggle repositories, Br35H, and figshare, containing various tumor categories, including meningioma, glioma, pituitary, and healthy images. Res-BRNet outperformed standard CNN models, achieving excellent accuracy (98.22%), sensitivity (0.9811), F1-score (0.9841), and precision (0.9822). Our results suggest that Res-BRNet is a promising tool for brain tumor classification, with the potential to improve the accuracy and efficiency of clinical diagnosis and treatment planning.

The swift advancement of cyber-physical systems (CPSs) across sectors such as healthcare, transportation, critical infrastructure, and energy enhances the crucial requirement for robust cybersecurity measures to protect these systems from cyberattacks. The cyber-physical method is a hybrid of cyber and physical components, and a safety breach in the element is central to catastrophic consequences. Cyberattack recognition and mitigation techniques in CPSs include using numerous models like intrusion detection systems (IDSs), access control mechanisms, encryption, and firewalls. Cyberattack detection employing deep learning (DL) contains training neural networks to identify patterns indicative of malicious actions within system logs or network traffic, allowing positive classification and mitigation of cyber-attacks. By leveraging the integral ability of DL methods to learn complex representations, this technique enhances the accuracy and efficiency of detecting diverse and growing cyber-attacks. Thus, the study proposes an automated Cyberattack Detection using Binary Metaheuristics with Deep Learning (ACAD-BMDL) method in a CPS environment. The ACAD-BMDL method mainly focuses on enhancing security in the CPS environment via the cyberattack detection process. The ACAD-BMDL method uses Z-score normalization to scale the input dataset. In addition, the binary grey wolf optimizer (BGWO) model is utilized to choose an optimal feature subset. Moreover, the Enhanced Elman Spike Neural Network (EESNN) model detects cyber-attacks. Furthermore, the Archimedes Optimization Algorithm (AOA) model is employed to select the optimum hyperparameter for the EESNN model. The empirical analysis of the ACAD-BMDL technique is performed on a benchmark dataset. The experimental validation of the ACAD-BMDL technique portrayed a superior accuracy value of 99.12% and 99.36% under NSLKDD2015 and CICIDS2017 datasets in the CPS environment.

Cancer is a life-threatening disease resulting from a genetic disorder and a range of metabolic anomalies. In particular, lung and colon cancer (LCC) are among the major causes of death and disease in humans. The histopathological diagnoses are critical in detecting this kind of cancer. This diagnostic testing is a substantial part of the patient’s treatment. Thus, the recognition and classification of LCC are among the cutting-edge research regions, particularly in the biological healthcare and medical fields. Earlier disease diagnosis can significantly reduce the risk of fatality. Machine learning (ML) and deep learning (DL) models are used to hasten these cancer analyses, allowing researcher workers to analyze a considerable proportion of patients in a limited time and at a low price. This manuscript proposes the Predictive Analytics of Complex Healthcare Systems Using the DL-based Disease Diagnosis Model (PACHS-DLBDDM) method. The proposed PACHS-DLBDDM method majorly concentrates on the detection and classification of LCC. At the primary stage, the PACHS-DLBDDM methodology utilizes Gabor Filtering (GF) to preprocess the input imageries. Next, the PACHS-DLBDDM methodology employs the Faster SqueezeNet to generate feature vectors. In addition, the convolutional neural network with long short-term memory (CNN-LSTM) approach is used to classify LCC. To optimize the hyperparameter values of the CNN-LSTM approach, the Chaotic Tunicate Swarm Algorithm (CTSA) approach was implemented to improve the accuracy of classifier results. The simulation values of the PACHS-DLBDDM approach are examined on a medical image dataset. The performance validation of the PACHS-DLBDDM model portrays the superior accuracy value of 99.54% over other DL models.

Intrusion Detection Systems are essential for securing both IoT and enterprise networks, yet models that appear nearly perfect on in domain tests often collapse under cross domain evaluation. This generalization gap have limited the operational reliability and slowed the adoption. We present FORT-IDS, a Federated, Optimized, Robust and Trustworthy multi phase framework that unifies cross domain assessment, adversarial hardening and explainable artificial intelligence within a single adaptive pipeline. FORT-IDS follows a five stage flow. First, it reduces drift by aligning heterogeneous feature spaces through lightweight mapping and normalization. Second, it applies SHAP and LIME to reveal unstable features whose attributions vary across sites or time. Third, it targets those features with adversarial augmentation and focused retraining to reshape brittle decision boundaries. Fourth, it is aggregating the clients update via adaptive attention weighted federated learning so that higher quality contributions exert greater influence while privacy is preserved. Fifth, it employs continual replay to retain corrective updates, preventing forgetting across rounds. We evaluate on UNSW-NB15 with 93,000 samples and 45 features and on DDoS Botnet IoT with 1.9 million samples and 30 features. Class imbalance is mitigated with SMOTE applied only to the training partitions. Baselines achieve \(F_1\) scores: Logistic Regression 0.80; Random Forest 0.91; MLP 0.95. Cross-dataset transfer remains asymmetric in training on DDoS and testing on UNSW yields \(F_1 = 0.39\), while the training on UNSW and testing on DDoS have reached \(F_1 = 0.99\). Advanced models improve in-dataset results: Graph Neural Networks reach 1.00 on DDoS and 0.91 on UNSW, surpassing CNN at 0.85 and LSTM at 0.82. SMOTE balances labels within the source training data only, so cross-dataset asymmetry can persist. By turning explanations into robustness actions and preserving them through federation and replay, FORT-IDS narrows the gap between lab performance and dependable deployment in dynamic IIoT and enterprise environments.

Background: Kidney stones, common urological diseases worldwide, are formed from hard urine minerals in the kidneys. Early detection is essential to prevent kidney damage and manage recurring stones. CT imaging has made significant progress in providing detailed information for disease diagnosis. Aim: This study aimed to enhance kidney stone detection through advanced imaging and machine learning techniques. Objective: The Guided Bilateral Feature Detector was proposed to identify and extract features for kidney stone detection in CT images. Unlike traditional filters like Gaussian and Bilateral filters, the Guided Bilateral Filter Detector prevented halo artifacts and preserved image edges by employing a guide weight. The extracted features were combined with the SVM algorithm to accurately detect kidney stones in CT images. Methods: The proposed detector used the Guided Bilateral Filter to reduce the halo artifacts in the images and enhance the feature detection process. The detector operated in four stages to extract important features from CT images, and a 128-feature point generator provided a more detailed representation in aiding kidney stone detection and classification. The proposed detector combined with the Support Vector Machine algorithm to improve reliability and reduce computational requirements. Results: Experimental results showed that the proposed Guided Bilateral Feature Detector with SVM outperformed existing models, including SIFT+SVM, SURF+SVM, PCA+KNN, EANet, Inception v3, VGG16, and Resnet50. The key performance metrics achieved included an accuracy of 98.56%, precision of 98.9%, recall of 99.2%, and an F1 score of 99%. Conclusion: The findings indicate that the Guided Bilateral Feature Detector with SVM significantly enhances the accuracy and reliability of kidney stone detection, providing valuable implications for clinical practice and future research in medical imaging.