Sajid Ullah Khan

Multimodal medical image fusion (MMIF) is essential for improving diagnostic accuracy by integrating useful details from various imaging modalities. However, current fusion methods have many challenges, such as modality‐specific drawbacks, restricted generalizability, high processing costs, and limited explainability. This paper introduces a novel framework called Federated and Explainable Multimodal Medical Image Fusion (FEMMIF) designed to address several challenges in medical image fusion through a hybrid approach. FEMMIF employs a modality‐agnostic dual‐branch encoder based on MobileNetV3 to extract both anatomical and functional features. These features are integrated using a cross‐attention technique and then processed through a Feature Importance Learning (FIL) module, which dynamically assigns weights to the contributions of each modality. The combined image is subsequently decoded with a decoder that utilizes residual refinement. Evaluations on various multimodal datasets, including MRI, PET, CT, and SPECT, show that FEMMIF consistently outperforms leading MMIF methodologies, achieving a structural similarity index (SSIM) greater than 0.91, improved entropy metrics, and faster inference speeds. The model demonstrates strong generalization across different modalities, reduced sensitivity to misalignment, and produces interpretable outputs suitable for clinical validation. The federated training approach also maintains privacy while achieving convergence comparable to centralized techniques. Overall, the FEMMIF framework demonstrates strong experimental performance but requires further prospective and multicenter validation before clinical deployment.

Objective: The thyroid gland plays a vital role in human body functions, including metabolism, and any dysfunction of the organ may exacerbate health risks. Early recovery from these conditions depends on the accurate diagnosis of accurate type of thyroid gland disorder. This study aims to develop an automated system for assisting physicians in clinical diagnosis of thyroid gland disorders. Method: The transfer learning capability of the deep neural network model DenseNet-201 was leveraged, and the model was tailored by altering its fully connected layer and the classification layer for the classification of thyroid gland conditions into seven categories, namely cold nodule, hot nodule, multi-nodular goiter, nodular goiter, thyroiditis, toxic diffuse goiter and normal. Results: The class-wise and overall performance of the method was evaluated by computing quantitative metrics like accuracy, specificity, precision, sensitivity, F 1-score, and area under the curve (AUC). The degree of similarity between the true labels of thyroid disorders annotated by experts and those predicted by the method was gauged using the kappa coefficient. For the five-fold cross-validated experimental results, we obtained an accuracy of 91.48 ± 2.79%, specificity of 98.58 ± 0.47%, precision of 91.57 ± 2.76%, sensitivity of 91.48 ± 2.77%, F 1-score of 91.38 ± 2.82%, and AUC of 0.988 ± 0.006. Additionally, the degree of similarity in the diagnostic capability of the proposed method and medical experts was measured by computing the kappa coefficient as 0.9148. Conclusion: The experimental results of the proposed method were compared with contemporary methods and illustrate relatively better performance in terms of accuracy, sensitivity, precision, and F 1-score. The value of the kappa coefficient, 0.9148, also depicts that the proposed method has the potential for applicability in clinical diagnosis to assist physicians in assessing the accurate type of thyroid disorders.

Autonomous underwater vehicles (AUVs) play a crucial role in oceanographic research, monitoring the environment, and exploring resources in the ocean. Nevertheless, the operational efficiency of these devices is frequently constrained by the limited battery capacity and the requirement for charging while connected to a power source. Wireless power transfer (WPT) offers a non-contact alternative to conventional wet-mate electrical connectors, with inductive coupling receiving particular attention because of its relatively high efficiency, safety, and suitability for underwater charging over short transfer gaps. However, it is limited by the transfer distance, coil misalignment, coupler design constraints, and eddy-current losses caused by conductive seawater. This review assesses the current state of inductive coupling technology with respect to AUVs. It discusses recent advancements in energy sources for AUVs, current technologies for underwater WPT, magnetic coupler design, challenges related to misalignment and their impact on WPT systems, as well as various strategies to mitigate these challenges. The review also covers the analysis and mitigation of eddy current loss and highlights the technical and engineering challenges encountered in power delivery. In general, inductive WPT has significant potential for making AUV charging performance more successful. Experimental studies have indicated efficiencies exceeding 90% in certain controlled situations; however, the literature also demonstrates that performance frequently deteriorates under misalignment and inadequate underwater conditions, which remain major challenges to practical implementation.

Healthcare, education, technological advancement, and farming are the key challenges facing developing countries, with agriculture unquestionably playing an important role in economic growth. Ensuring adequate food production is essential for citizens’ survival, as it is anticipated that efforts in this area would result in increased food productivity. A key approach to enhancing field productivity involves meticulous care of its components, starting with the production of crops. Wheat leaf rust poses a severe threat, particularly to young seedlings, constituting a significant fungal disease that can cause a 25% reduction in wheat productivity. To overcome these issues, this research work proposes a novel image fusion approach called Multi-Scale Discrete Wavelet Transform (MS-DWT). The method uses distinct fusion strategies to extract meaningful details from source images. After that, a Lab Color Space (LCS), followed by a color thresholding method, is employed for the detection and lesion localization of rust in the source images. Furthermore, the proposed model measures the area affected by rust in wheat crops, providing farmers with vital information during the post-medication (anti-rust spray) operation. The experimental findings demonstrate superior performance and achieve a classification accuracy of 98.85%, and the maximum testing accuracy was 99.17% on our generated dataset.

Seamless and uninterrupted connectivity within high-mobility environments (e.g. aerospace and high-speed terrestrial transportation) drives a critical demand for Space-Air-Ground Integrated Networks (SAGIN). In such dynamic environments, conventional handovers lead to frequent loss of connection, higher latency, and diminished quality-of-service (QoS). A new high-mobility handover procedure for SAGIN based on the concept of Integrated Sensing and Communication (ISAC) is introduced. We suggest a new method using real-time environmental perception to improve the decision-making of hand over, reducing the frequency of handover, and balancing the spatial, air and ground network resource allocation. It combines dynamic priority scheduling and a cross-layer communication protocol into a general handover procedure for seamless mobility between network layers. MATLAB-based simulations show that, when compared to existing ground-based and satellite-assisted handover methods, the overhead of LEO-enabled handovers often results in a substantial reduction in latency while simultaneously providing enhanced network throughput with minimum packet loss. The results show that our approach is appropriate to preserve an agreement on the quality received in highly dynamic environments, opening perspectives with novel scientific questions for SAGIN and high-mobility networks.

Diabetic retinopathy (DR) is a primary reason for visual impairment and blindness in individuals with diabetes worldwide. Timely detection of DR is essential to prevent vision loss in diabetics. However, noise and limited model transparency often compromise the accuracy of diagnosing retinal fundus images. Noise and interpretability are the two main challenges occurring in imaging datasets, overshadowing concerns such as class imbalance or device variability. These distortions are present in all datasets and devices, reducing the clarity of diagnostic signals at the pixel level and often obscuring early lesions within background noise. Addressing these challenges, this research introduces an innovative model called Explainable MINet‐ViT, which combines advanced noise reduction techniques with explainable deep learning for more reliable identification of DR. The model incorporates a multi‐level denoising network (MINet), modified by a noise‐specific pre‐processing module using a Variance‐Stabilizing Transform (VST) and deep residual feature mapping. A hybrid deep learning architecture that combines Convolutional Neural Networks (CNNs) with Vision Transformers (ViTs) is employed to extract both local and global spatial information. We apply explainability strategies, such as Grad‐CAM and SHAP, to ensure clinical interpretability by identifying the crucial retinal regions that influence model predictions. Quantitative and qualitative results show improved performance, robustness, and clinical applicability, achieving an accuracy of 97.6%, a sensitivity of 0.96, a specificity of 0.97, a Kappa of 0.92, and an AUC of 96.7%. Analyses of standard datasets reveal that our proposed model outperforms prior models in accuracy, noise robustness, and interpretability, rendering it exceptionally suitable for real‐world clinical applications.

Recent progress in the domain of wireless communication systems has brought to light emerging trends that underscore the evolution of modern communication paradigms. Nonetheless, the performance of such systems remains constrained by persistent technical limitations, including packet loss, bandwidth scarcity, and suboptimal spectral efficiency, all of which necessitate rigorous analytical investigations. To address these constraints, the present study leverages an adaptive modulation technique tailored to enhance communication efficiency. This technique is systematically evaluated considering various parameters, including the signal-to-interference-plus-noise ratio (SINR), link distance, and the Standard Propagation Model (SPM), across canonical wireless channel environments, such as Additive White Gaussian Noise (AWGN), Rayleigh, and Rician fading models. The primary objective is to minimize the Bit Error Rate (BER), maximize system throughput, and ensure efficient utilization of available bandwidth by strategically optimizing the SPM parameters for long-range cellular communication links. Simulation outcomes substantiate that the proposed SPM-integrated framework achieves a target BER = 10⁻² and exhibits enhanced spectral efficiency for 64-QAM modulation across a 6 km cell radius, thereby outperforming the conventional non-SPM approach. Notably, the proposed technique achieves SINR improvements of 40%, 43%, and 41% under AWGN, Rayleigh, and Rician conditions, respectively. To ensure optimal configuration, an advanced optimization algorithm is employed to dynamically select the most effective SPM parameters, enabling robust performance across varying channel conditions. Moreover, empirical analysis confirms that the SPM-enhanced system architecture facilitates the stable deployment of 64-QAM across all tested channel models, consistently yielding superior SINR and reduced BER relative to baseline techniques that lack SPM integration.