Protein toxicity prediction is crucial for drug discovery, safety assessment, and toxicological research. This study introduces $\mathrm{Deep\_{T}PPred}$, a novel hybrid deep learning (DL) model that integrates Convolutional Neural Networks (CNN) and Recurrent Neural Networks (RNN) for accurate protein toxicity prediction. The model effectively combines diverse protein sequence descriptors to capture complex sequence relationships by leveraging a feature fusion technique. The methodology involved advanced feature extraction, rigorous training, and performance evaluation using benchmark datasets. $\mathrm{Deep\_{T}PPred}$ demonstrates state-of-the-art performance with an accuracy of 0.9983, specificity of 0.9988, sensitivity of 0.9975, and Kappa and MCC values of 0.9963. These results underscore the proposed model's robustness, reliability, and generalization capability, surpassing existing models across all metrics. The study highlights the potential of hybrid DL and feature fusion techniques to significantly enhance protein toxicity prediction, providing valuable insights and tools for bioinformatics pipelines and applications.

Druggable proteins are defined as proteins that can interact with drugs to modulate certain biological activity. The identification of druggable proteins holds significant clinical importance, directly impacting the development of targeted therapies for diseases like cancer and metabolic disorders. Identifying druggable proteins involves various methods, including computational prediction models, mass spectrometry (MS), and biochemical assays, but achieving high accuracy remains a challenge. This study proposes DrugEL, an ensemble learning model that uses Bayesian inference to integrate predictions from multiple algorithms: Random Forest (RF), K-Nearest Neighbor (KNN), LightGBM (LGBM), and Decision Tree (DT) with seven feature extraction methods (LSA, AAC, PAAC, GAAC, NMBroto, AAIndex, and KNN). The results show that DrugEL outperforms existing models in terms of accuracy (0.9758), MCC (0.9515), AUC (0.9758), sensitivity (0.9742) and specificity (0.9774), particularly excelling with the LSA method.

Human skin disease, the most infectious dermatological ailment globally, is initially diagnosed by sight. Some clinical screening and dermoscopic analysis of skin biopsies and scrapings for accurate classification are medically compulsory. Classification of skin diseases using medical images is more challenging because of the complex formation and variant colors of the disease and data security concerns. Both the Convolution Neural Network (CNN) for classification and a federated learning approach for data privacy preservation show significant performance in the realm of medical imaging fields. In this paper, a custom image dataset was prepared with four classes of skin disease, a CNN model was suggested and compared with several benchmark CNN algorithms, and an experiment was carried out to ensure data privacy using a federated learning approach. An image augmentation strategy was followed to enlarge the dataset and make the model more general. The proposed model achieved a precision of 86%, 43%, and 60%, and a recall of 67%, 60%, and 60% for acne, eczema, and psoriasis. In the federated learning approach, after distributing the dataset among 1000, 1500, 2000, and 2500 clients, the model showed an average accuracy of 81.21%, 86.57%, 91.15%, and 94.15%. The CNN-based skin disease classification merged with the federated learning approach is a breathtaking concept to classify human skin diseases while ensuring data security.

Capsule endoscopy (CE), introduced as a modality for non-invasive examination of entire gastrointestinal tract, demands for an efficient computer-aided decision making system to relieve the physician from the responsibility of screening around 60,000 video frames per patient. An automatic and robust segmentation algorithm can aid the automation of CE screening and decision making procedure. In this paper, we propose a new segmentation algorithm based on GrowCut and apply the algorithm for CE images containing bleeding. To substitute the manual seed input in traditional GrowCut segmentation, the proposed Automated GrowCut (AGC) algorithm initially segments the images using clustering. The cluster centroids, subsequently labeled as bleeding, non-bleeding and background by a trained SVM classifier, serve as seeds for the GrowCut segmentation. A comprehensive evaluation and comparison with respect to ground truth exhibits that the proposed method can achieve a Dice Similarity Coefficient of 0.81, comparable to the interactive GrowCut, requiring only 13.96% of the computation time of interactive GrowCut. The comparison with two other state-of-the-art unsupervised segmentation methods, unsupervised GrowCut and Fuzzy c-means, further justifies the suitability of the proposed method for automated segmentation and annotation of CE images.

Wireless capsule endoscopy (WCE) plays a significant role in the non-invasive small intestine screening for obscure gastrointestinal bleeding detection. However, the task of reviewing 60,000 frames to detect the bleeding encumbers the clinician, leading to visual fatigue and false diagnosis. In this paper, we propose a color feature based bleeding detection system with feature selection using a modified ant colony optimization (MACO) algorithm. We have utilized the feature selection capability of MACO algorithm to find the optimum feature subset over the color space of RGB and HSV, which provided a classifier that outperforms the classifier formed from RGB and HSV features individually. Comprehensive experimental results reveal that the proposed MACO algorithm can detect the optimal feature subset with performance comparable to exhaustive search in case of individual classifier from RGB and HSV requiring 2% of the computational time compared to exhaustive search. The comparative study of feature selection showed that MACO can provide the most relevant features and improve the performance in terms of accuracy, sensitivity and computational time.

Joint optimization of the physical layer security with end-to-end delay management is studied in the uniquely constrained context of wireless body area networks (WBANs). A game-theoretic framework is proposed wherein body-worn sensor devices interact in the presence of wiretappers and under fading channel conditions to find the most secure multi-hop path to the hub, while adhering to the end-to-end delay requirements imposed by the application. We model the problem as the search for a Nash network topology where no unilateral deviation in strategy by any single sensor node improves the secrecy of its transmissions, and provide a distributed algorithm guaranteed to converge to a Pareto-dominant Nash solution. The framework is evaluated through numerical simulations in conditions approximating actual deployment of WBANs for moving and stationary scenarios. Results validate the merits of the proposed framework to improve the security of transmissions compared with the star topology and IEEE 802.15.6 two-hop topology extension with a best-channel algorithm, at the expense of an admissible increase in the end-to-end delay.

A game-theoretic approach is proposed to investigate the problem of relay selection and power control with quality of service constraints in multiple-access wireless body area networks (WBANs). Each sensor node seeks a strategy that ensures the optimal energy efficiency and, at the same time, provides a guaranteed upper bound on the end-to-end packet delay and jitter. The existence of Nash equilibrium for the proposed non-cooperative game is proved, the Nash power control solution is analytically calculated, and a distributed algorithm is provided that converges to a Nash relay selection solution. The game theoretic analysis is then employed in an IEEE 802.15.6-based WBAN to gauge the validity and effectiveness of the proposed framework. Performance behaviors in terms of energy efficiency and end-to-end delay and jitter are examined for various scenarios. Results demonstrate the merits of the proposed framework, particularly for moving WBANs under severe fading conditions.

A robust optimization model is considered for nonzero-sum discounted stochastic games with incomplete information in order to formally formulate and analyze the intrusion detection problem in wireless sensor networks (WSNs). Security requirements of WSNs are taken into account to characterize the game parameters and model the player objectives. To generalize the problem, the game data are assumed not to be fully known to the players, who take a robust optimization approach to address this data uncertainty. For assessing the validity and effectiveness of the framework, illustrative instances of the developed game model are generated. Equilibrium analysis reveals how the conflicting objectives of the intruder and intrusion detection system compel them to adopt different conservative stances toward data uncertainty. It is also shown, by numerical results, that the robust approach in the presence of uncertainty reduces the sensitivity of the solution with respect to data perturbations, and thus improves design stability.

The electrocardiogram (ECG) is a new and promising modality for biometric recognition. This signal is typically collected within welfare monitoring environments along with other vital signals. As opposed to static biometric modalities like the iris or the fingerprints, ECG is time-dependent thereby presenting the opportunity of continuously authenticating subjects in such environments. However, ECG is affected by both physical and psychological activities, which have to be taken into consideration for successful deployment of this biometric. This paper demonstrates the effects of psychology on the performance of ECG biometric systems and proposes a novel methodology for automatic template updating in order to mitigate the risks associated with poor biometric matching.

Biometric secret-key binding inherently requires signal processing and error correction schemes due to noisy measurement readings. Two previously proposed strategies, Quantization Index Modulation (QIM) and Wyner-Ziv (WZ) coding, are studied in the context of bio metric key binding. We characterize the tradeoff between key rate leakage and key rate-reconstruction distortion, showing that while WZ coding has a better rate-leakage tradeoff than QIM, the latter has a better rate-reconstruction tradeoff. A new strategy is proposed to combine the merits of these schemes. Known as distortion-enhanced Wyner-Ziv coding (DE-WZ), this scheme is demonstrated to exhibit improved flexibility based on numerical results for a uniform source model and scalar quantization.

Wireless sensor networks (WSN) can be an attractive solution for a plethora of communication applications, such as unattended event monitoring and tracking. One of the looming challenges that threaten the successful deployment of these sensor networks is source-location privacy, especially when a network is deployed to monitor sensitive objects. In order to enhance source location privacy in sensor networks, we propose the use of an opportunistic mesh networking scheme and examine four different approaches. Each approach has different selection criteria for the next relay node. In opportunistic mesh networks, each sensor transmits the packet over a dynamic path to the destination. Every packet from the source can therefore follow a different path toward the destination, making it difficult for an adversary to backtrack hop-by-hop to the origin of the sensor communication.

The effect of an arbitrary indoor infrastructure environment on the performance of a wireless multihop network is investigated. To this end, an accurate channel modeling tool based on 3D ray tracing is used first to evaluate the signal strength in different areas of the environment. Then, a discrete event simulator is applied to examine the performance of the network with two classes of routing protocols: traditional vs. opportunistic. It is shown that for an indoor environment, opportunistic routing performs better based on the obtained results, with respect to throughput, delay and delivery ratio.

Abstract This work investigates the feasibility of ECG‐based identity management in mobile health monitoring applications. A body area network that operates in conjunction with ECG biometric recognition is explored for mobile monitoring of patients, rescuers, pilots, soldiers, or field agents in general. Among the major challenges of this technology is the stability of the signals over the monitoring duration. Time dependency is responsible for ECG destabilization, which becomes a significant issue for reliable monitoring. We propose a novel framework that addresses this inadequacy, by updating a gallery template when feature matching is compromised. In addition, strategies for tackling privacy issues in medical data management are proposed. A protocol level solution is discussed, to deal with the ethical issues of this technology. An automatic way of aggregating and managing personal information is presented, designated to operate on the basis of anonymity. The experimental performance measured over long‐ECG recordings demonstrates promising results. Copyright © 2010 John Wiley & Sons, Ltd.

Biometric encryption (BE) has recently been identified as a promising paradigm to deliver security and privacy, with unique technical merits and encouraging social implications. An integral component in BE is a key binding method, which is the process of securely combining a signal, containing sensitive information to be protected (i.e., the key), with another signal derived from physiological features (i.e., the biometric). A challenge to this approach is the high degree of noise and variability present in physiological signals. As such, fuzzy methods are needed to enable proper operations, with adequate performance results in terms of false acceptance rate and false rejection rate. In this work, the focus will be on a class of fuzzy key binding methods based on dirty paper coding known as quantization index modulation. While the methods presented are applicable to a wide range of biometric modalities, the face biometric is selected for illustrative purposes, in evaluating the QIM-based solutions for BE systems. Performance evaluation of the investigated methods is reported using data from the CMU PIE face database.

Conventional OFDM systems notoriously suffer from a high peak-to-average power ratio (PAPR), which makes hardware implementation problematic. We propose a general framework for reducing PAPR using a combination of time-domain interpolation and resampling. In essence, this combination allows for both pulse shaping as well as phase changes to be performed. It is demonstrated that the framework offers not only improved flexibility, but also reduced computational complexity compared to existing methods. The former benefit is due to the availability of various user-controllable parameters in the framework, while the latter is a result of performing PAPR reduction directly in the time domain. We also show that the accompanying channel estimation and equalization procedures can be achieved in a practical manner, with only modest modifications. The simulation results illustrate that, even though the complexity requirements are substantially lower, the PAPR reduction capability of this framework is comparable to that of existing PAPR reduction methods.

In previous work, it has been found that using a data burst with size varying according to the encountered channel stationarity can improve efficiency and QoS for burst- by-burst wireless communication systems. In this paper, adaptive modulation is incorporated as a further means for improving spectral efficiency. The modulation scheme is switched according to the operating channel conditions, with an adaption metric based on the MSE. When the channel conditions are favorable, a high-throughput modulation mode is used for efficiency. Conversely, a robust lower-throughput mode is selected under adverse channel conditions, for quality. Obtained results show the utility of adaptive modulation in conjunction with a variable-size burst for optimizing spectral efficiency, while maintaining a prescribed level of QoS.

Burst-by-burst equalization strategies assume that data in a burst are received under quasi-stationary channel conditions. Hence, using a fixed-size burst results in conservative transmission efficiency, since worst case scenarios must be accounted for. We investigate the potential of a variable-size burst for efficiency in time-varying environments. First, a channel model explicitly describing environment changes is developed. Analysis of variance is then proposed for tracking the channel stationarity. Based on knowledge of the channel stationarity, the burst size is adapted most advantageously. Obtained results show the feasibility of a variable-size burst to improve efficiency compared to a fixed-size counterpart.