ABSTRACT With increasing focus on sustainable agriculture and AI‐enabled solutions, this work proposes AloeVeraNet, a compact deep learning model designed for the efficient and real‐time detection of aloe vera leaf diseases on edge devices. The model employs depthwise and pointwise convolutions to achieve a significantly reduced parameter count (289 K) and model size (1.10 MB), enabling deployment in low‐resource environments. With 96.09% accuracy, AloeVeraNet sets a new benchmark in classifying aloe vera leaf conditions: healthy, rust‐infected and spot‐affected, outperforming MobileNetV2, EfficientNetV2‐S and VGG16. This sustainable, artificial intelligence (AI)‐based solution supports precision agriculture through optimised computation, energy efficiency and local disease monitoring, all without relying on cloud infrastructure, thereby contributing to environmentally responsible farming practices. This study demonstrates the value of integrating AI with sustainable edge computing in creating resilient and inclusive solutions for the agricultural sector.

Cardiac arrhythmia is one of the prime reasons for death globally. Early diagnosis of heart arrhythmia is crucial to provide timely medical treatment. Heart arrhythmias are diagnosed by analyzing the electrocardiogram (ECG) of patients. Manual analysis of ECG is time-consuming and challenging. Hence, effective automated detection of heart arrhythmias is important to produce reliable results. Different deep-learning techniques to detect heart arrhythmias such as Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), Transformer, and Hybrid CNN-LSTM were proposed. However, these techniques, when used individually, are not sufficient to effectively learn multiple features from the ECG signal. The fusion of CNN and LSTM overcomes the limitations of CNN in the existing studies as CNN-LSTM hybrids can extract spatiotemporal features. However, LSTMs suffer from long-range dependency issues due to which certain features may be ignored. Hence, to compensate for the drawbacks of the existing models, this paper proposes a more comprehensive feature fusion technique by merging CNN, LSTM, and Transformer models. The fusion of these models facilitates learning spatial, temporal, and long-range dependency features, hence, helping to capture different attributes of the ECG signal. These features are subsequently passed to a majority voting classifier equipped with three traditional base learners. The traditional learners are enriched with deep features instead of handcrafted features. Experiments are performed on the MIT-BIH arrhythmias database and the model performance is compared with that of the state-of-art models. Results reveal that the proposed model performs better than the existing models yielding an accuracy of 99.56%.

Real-time social multimedia level threat monitoring is becoming harder, due to higher and rapidly increasing data induction. Data induction through electric smart devices is greater compared to information processing capacity. Nowadays, data becomes humongous even coming from the single source. Therefore, when data emanates from all heterogeneous sources distributed over the globe makes data magnitude harder to process up to a needed scale. Big data and Deep learning have become standard in providing well-known solutions built-up using algorithms and techniques in resolving data matching issues. Now, with the involvement of sensors and automation in generating data obscures everything, predicting results to overcome a current era of ever enhancing demands and getting real-time visualization brings the need of feature like human behavior mode extraction to overcome any future threats. Big data analytics can bring the opportunity of predicting any misfortune even before they happen. Map reduce feature of big data supports massive data oriented process execution using distributed processing. Real-time human feature identification and detection can occur through sensors and internet sources. A behavioral prediction can further classify the information collected for introducing enhanced security extents. Real-time sensor devices are producing 24/7-hour data for further processing recording each event. IoT-based sensors can support in behavioral analysis model of a human. Real-time human behavioral monitoring based on image processing and IoT using big data analytics.

Abstract Convolutional Neural Networks (CNNs) have made a great impact on attaining state‐of‐the‐art results in image task classification. Weight initialization is one of the fundamental steps in formulating a CNN model. It determines the failure or success of the CNN model. In this paper, we conduct a research based on the mathematical background of different weight initialization strategies to determine the one with better performance. To have smooth training, we expect the activation of each layer of the CNN model follow the standard normal distribution with mean 0 and SD 1. It prevents gradients from vanishing and leads to more smooth training. However, it was obtained that even with the appropriate weight initialization technique, a regular Rectified Linear Unit (ReLU) activation function increases the activation mean value. In this paper, we address this issue by proposing weight initialization based (WIB)‐ReLU activation function. The proposed method resulted in more smooth training. Moreover, the experiments showed that WIB‐ReLU outperforms ReLU, Leaky ReLU, parametric ReLU, and exponential linear unit activation functions and results in up to 20% decrease in loss value and 5% increase in accuracy score on both Fashion‐MNIST and CIFAR‐10 databases.

Summary In the last decade, artificial intelligent systems based on neural networks have gradually become primary source for clinical decision support systems (CDSS) and are being used in diverse areas of medical diagnosis, classification, and prediction. An artificial neural network (ANN) consists of a large number of processing units which performs the computation in a parallel and distributed environment. They learn the pattern from the examples provided to it and then generalize based on the concepts they have learned while training. This paper presents a review of the current status of ANN and its variants as CDSS in various medical disciplines. The work focuses and describes the methods making use of simple ANN and use of real‐time approaches based on big data using ANN in cloud computing environment for various medical applications. Critical analysis of various methods based on smart approaches indicates that feed‐forward back propagation ANN performs sufficiently better in the domain of medicines with a high degree of accuracy.

Summary Deep learning based neural networks and their variants have gained popularity due to their inherent flexibility to handle unforeseen especially when a chaotic time series big data are required to be dealt with. There are innumerable applications that are beneficiary of vast interest in computational intelligent approaches that include but not limited to robotics, healthcare, transport, industrial, decision making, and gaming. This paper attempts to investigate the effectiveness of using a neural nonlinear autoregressive with exogenous inputs (NARX) controller in an emerging application field of balancing systems like inverted pendulum (IP) using big data. This paper's aim has been to control an IP cart system by designing a neural NARX controller, and the focus is primarily on real‐time processing in industrial applications grounded on big data ecosystems. In the proposed work, an IP system is mathematically modeled and first controlled utilizing a combination of classical proportional‐integral‐derivative (PID) controllers for cart and pendulum. Second, a chaotic time series input–output data are obtained and are used to train two NARX controllers for cart and pendulum, respectively. Both the controllers are designed as single‐input single‐output systems with one layer each at input and output with suitable number of hidden layers and neurons. Performance comparison of NARX system behavior with PID controller indicates that the NARX controllers successfully adapt to two different kinds of unknown inputs and effectively stabilize the plant. Simulation results confirm that NARX controllers follow the training parameters and exhibit superior performance and overall system stability than PID control. Experimental results demonstrate the effectiveness of the approach.

The industrial Internet of Things (IIoT) has the potential to transform several industries. However, reliable data collection, processing, and analytics cannot be performed if the underlying cloud storage is not trustworthy. The current IIoT architectures tend to adopt cloud computing as a backbone for their implementation and further deployment. Therefore, the trustworthiness of the cloud storage is of paramount importance for the design of efficient and reliable IIoT. This becomes very challenging if the data storage is outsourced to a third party, which may raise many problems such as noncompliance to service-level agreement, data theft, and privacy issues. Mobile agents have a number of characteristics including mobility, lightweight, autonomous, reactive, and intelligent making them ideal for deployment in many distributed applications. Therefore, in this article, we propose a multiagent-based approach to address the geolocation assurance problem of the outsourced data in the context of IIoT. As a result, we achieve efficient geolocation assurance with manageable costs without the need of a trusted third party.

New integrated technologies have changed various existing fields and converted into new and advanced data communication systems including, smart agriculture, smart homes, smart health, and smart transportation systems. Internet of Things (IoT) has evolved a new theme to vehicular networks field known as the Internet of Vehicles (IoV). This paper presents a comprehensive review and detailed background and motivation to evolve the heterogeneous vehicular networks. Paper also proposed new integrated models and key technologies related to network maintenance, a six-layered architecture model based on protocol stack and network elements, network model based on cloud services, big data analytical model based on data acquisition and analytics, security model based on detection and prevention systems. Paper also envisioned existing challenges and future directions to design the new integrated models.

The Internet of Things (IoT), with smart sensors, collects and generates big data streams for a wide range of applications. One of the important applications in this regard is video analytics which includes object detection. It has been considered as an important research area particularly after the development of deep neural networks. We demonstrate the applications, effectiveness, and efficiency of the convolutional neural network algorithms, i.e., Faster-RCNN and Mask-RCNN, to facilitate video analytics in the IoT domain, for overhead view multiple object detection and segmentation. We used the Faster-RCNN and Mask-RCNN models trained on the frontal view data set. To evaluate the performance of both algorithms, we used a newly recorded overhead view data set containing images of different objects having variation in field of view, background, illumination condition, poses, scales, sizes, angles, height, aspect ratio, and camera resolutions. Although the overhead view appearance of an object is significantly different as compared to a frontal view, even then the experimental results show the potential of the deep learning models by achieving the promising results. For Faster-RCNN, we achieved a true-positive rate (TPR) of 94% with a false-positive rate (FPR) of 0.4% for the overhead view images of persons, while for other objects the maximum obtained TPR is 92%. The Mask-RCNN model produced TPR of 93% with FPR of 0.5% for person images and maximum TPR of 92% for other objects. Furthermore, the detailed discussion is made on output results which highlights the challenges and possible future directions.

This article proposes a novel methodology to detect malicious uniform resource locators (URLs) using simulated expert (SE) and knowledge-base system (KBS). The proposed study not only efficiently detects known malicious URLs but also adapts countermeasure against the newly generated malicious URLs. Moreover, this article also explored which lexical features are contributing more in final decision using a factor analysis method, and thus help in avoiding the involvement of human experts. Furthermore, we apply the following state-of-the-art machine learning (ML) algorithms, i.e., naïve Bayes (NB), decision tree (DT), gradient boosted trees (GBT), generalized linear model (GLM), logistic regression (LR), deep learning (DL), and random rest (RF), and evaluate the performance of these algorithms on a large-scale real data set of data-driven Web applications. The experimental results clearly demonstrate the efficiency of NB in the proposed model as NB outperforms when compared to the rest of the aforementioned algorithms in terms of average minimum execution time (i.e., 3 s) and is able to accurately classify the 107 586 URLs with 0.2% error rate and 99.8% accuracy rate.

Recently, Industry 4.0 has attracted much attention. It has close relations with the Internet of Things (IoT). On the other hand, convolutional neural networks (CNNs) have shown promising performance in many foundational services of the IoT applications. For the IoT applications with high-speed data streams and the requirement of time-sensitive actions, fast processing is demanded on small-scale platforms or even on IoT devices themselves. Therefore, it is inappropriate to employ cumbersome CNNs in IoT applications, making the study of model compression necessary. In knowledge transfer, it is common to employ a deep, well-trained network, called teacher, to guide a shallow, untrained network, called student, to have better performance. Previous works have made many attempts to transfer single-scale knowledge from teacher to student, leading to degradation of generalization ability. In this article, we introduce multiscale representations to knowledge transfer, which facilitates the generalization ability of student. We divide student and teacher into several stages. Student learns from multiscale knowledge provided by teacher at the end of each stage. Extensive experiments demonstrate the effectiveness of our proposed method both on image classification and on single image super-resolution. The huge performance gap between student and teacher is significantly narrowed down by our proposed method, making student suitable for IoT applications.

A rapid development has been seen in the Vehicular ad hoc networks (VANETs) because of their applicability and significance in the fields of traffic management, road monitoring and safety, infotainment, and on-demand services. Route planning in vehicular networks based on efficient collection of real-time data can effectively mitigate traffic congestion problems in urban areas. Furthermore, real-time data is shared by using an effective sharing mechanism to avoid redundancy of the collected information. However, dynamic route replanning and effective sharing mechanisms based on real-time data are still challenging problems. Therefore, based on the aforementioned constraints, this paper describes a route discovery technique that uses real time data collected from various vehicles using the Internet of Things. The proposed scheme is based on the novel data dissemination technique for information sharing among the roadside units. RTRD is comprised of VANETs, vehicular traffic servers, and a 5G-based cellular system of public transportation. By considering the traffic congestion in urban areas, the optimal path is calculated to re-plan routes based on the k shortest path algorithm, and a load balancing technique is adopted to avoid further congestion.

Internet of Things (IoT) plays a key role in connecting the e-health system with the cyber world through new services and seamless interconnection between heterogeneous devices. Therefore, it becomes computationally inefficient to analyze and select features from such massive volume of data. Therefore, keeping in view the needs above, this paper presents a system architecture that selects features by using Artificial Bee Colony (ABC). Moreover, a Kalman filter is used in Hadoop ecosystem that is used for removal of noise. Furthermore, traditional MapReduce with ABC is used that enhance the processing efficiency. Moreover, a complete four-tier architecture is also proposed that efficiently aggregate the data, eliminate unnecessary data, and analyze the data by the proposed Hadoop-based ABC algorithm. To check the efficiency of the proposed algorithms exploited in the proposed system architecture, we have implemented our proposed system using Hadoop and MapReduce with the ABC algorithm. ABC algorithm is used to select features, whereas, MapReduce is supported by a parallel algorithm that efficiently processes a huge volume of data sets. The system is implemented using MapReduce tool at the top of the Hadoop parallel nodes with near real-time. Moreover, the proposed system is compared with Swarm approaches and is evaluated regarding efficiency, accuracy, and throughput by using ten different data sets. The results show that the proposed system is more scalable and efficient in selecting features.

Load balancing and energy conservation techniques are one of the important constraints in the design of in wireless sensor network (WSN). Usually, clustering technique helps the network in the minimum utilization of energy that results in enhancing network lifetime. Moreover, various nodes in the multihop network that are near to the base station drain their battery very quickly thus result in creating hot spot problem in a network. To overcome such constraints, this paper proposes a multi-layer clustering architecture for selection of forwarding node, rotation of cluster head, and inter and intra-cluster routing communication. The proposed scheme efficiently tackle the rotation of forwarder node by incorporating routing table (table list) at each node. Moreover, the rotation is performed by the consideration of two threshold levels of the residual energy of a node. Also, the exploitation of decision maker node, forwarder node, backup forwarder node, and non-forwarder node enhancing the routing strategy in a network. The performance of the proposed scheme is tested and evaluated by C programming language. The results show that the proposed scheme successful achieve better results than TLPER and EADUC in energy consumption per node, end-to-end communication, hop count in cluster formation.

Aphasia is a common adult language disorder acquired after a stroke, head injury, tumor, etc. Accurate diagnosis influences the prognosis of any speech and language disorder including aphasia. Therefore, in this paper we have proposed a semi-automated Aphasia diagnosis and classification framework employing feature extraction and pattern matching techniques of the digital signal processing (DSP). The proposed scheme evaluates the acoustic properties, time consumed, and speech characteristics for each language component i.e. naming, repetition, and comprehension. The naming and repetition tasks utilize DSP techniques. The proposed solution is highly scalable since it determines the diagnosis based on acoustic properties instead of the language characteristics. Thus, it eases extending into multiple languages. The mathematical relationships calculate the corresponding score for each component. The framework then determines the diagnosis according to the obtained scores. Since it occupies computational analysis of the speech signals, it reduces the subjectivity of the manual diagnosis process, meanwhile increasing the efficiency and accuracy by consistent diagnosis decisions. Finally, it distinguishes two sub types of Aphasia i.e. Anomic Aphasia and Wernicke's Aphasia. The results clearly revealed the efficiency improvement achieved by replacing the live auditory model with pre-recorded auditory model.