ABSTRACT We propose a Essential Secret Image Sharing scheme using Linear Homogeneous Recurrence Relation and polynomials for sharing a grayscale or color secret image in the semihonest model. In our scheme, the dealer generates essential and nonessential shares of a secret image. A combiner needs shares to reconstruct the secret image, where at least are essential shares. Unlike most state‐of‐the‐art schemes restricting to be equal to , our scheme also allows for . This merit makes reconstruction possible even if up to essential shares are unavailable. Additionally, compared to state‐of‐the‐art schemes, our scheme offers substantial reductions in share sizes—by factors formed from , , , and . Thus, with this reduced size of shares, leading to reduced share storage costs, our scheme has a broader range of applications, including those with limited budgets. Moreover, in cases where , the shares generation period in our scheme, during which an adversary can potentially steal the secret image from the dealer, is at least 42% shorter than that in the state‐of‐the‐art scheme supporting .

The increase in the use of autonomous vehicles has motivated a paradigm shift in the transportation domain as it redefines the boundaries of urban mobility by augmenting safety measures. This research paper explores an innovative approach to enhance 6G millimeter-wave (mmWave) beam prediction for vehicle-to-infrastructure (V2I) communications by using generative adversarial networks (GANs). By generating synthetic data samples effectively and balancing the real-world dataset, we improve the accuracy of beam prediction models significantly. Our proposed method of training random forests on synthetic data (RFGAN) to predict beam indices provides the solution for imbalanced class issues and significantly improves the predictive performance of mmWave beam selection, contributing to more reliable and efficient V2I communications. This work also performs comparative analysis with state-of-the-art models in top-K evaluation metrics, average power loss, and overhead savings related to adapting to the new approach.

Abstract We propose a novel secret image sharing scheme with ciphertext‐based share authentication (SISS‐CSA) for sharing grayscale and color secret images in the malicious model. In SISS‐CSA, the dealer and each participant, individually acting as a combiner, can identify each invalid share received from the malicious participant(s) before using it to reconstruct the secret image. This capability, which most comparable schemes lack, prevents reconstructing an incorrect secret image. In SISS‐CSA, the asymptotic time complexities of operations executed by the dealer in the shares generation phase and executed by each combiner in the secret image reconstruction phase are and , respectively. Here, is the number of grayscale values in the secret image, is the number of generated shares, and is the threshold number of shares required for reconstructing the secret image. These asymptotic time complexities and the size of additional information each combiner stores for identifying invalid share(s) are comparatively lesser than those in the state‐of‐the‐art schemes. Furthermore, we obtain a maximum of reduction in the size of additional information each combiner stores for share authentication using the ciphertext‐based share authentication compared to using the standard SHA‐256. To the best of our knowledge, none of the related share authentication approaches achieves this much reduction. We prove the properties of SISS‐CSA using theoretical analysis. We also provide experimental results validating the implications of theoretical analysis corresponding to asymptotic time complexities and the random nature of shares.

Abstract Due to inter‐class homogeneity and intra‐class variability, the classification of skin lesions in dermoscopy images has remained difficult. Although deep convolutional neural networks (DCNNs) have achieved satisfactory performance for binary skin cancer classification, multiclass skin lesion classification is still an open problem due to the limited training samples and class imbalance issues. To tackle these issues, in this article, we propose a multi‐CNN ensemble approach dubbed for multiclass skin lesion classification. The includes three individual CNN models, each helping in extracting different high‐level features from skin images and thereby yielding different prediction results. First, we design a lightweight CNN model to extract prominent features and train it from scratch, which primarily aims at avoiding the data scarcity problem. Then, we ensemble two different pre‐trained CNN models with the lightweight model to improve the performance and generalization capability. The proposed ensemble approach can effectively fuse the predictions of each individual CNN model using the averaging method. The approach is validated using a benchmark data set, HAM10000, which contains skin lesion images of seven different classes. The results demonstrate that the outperforms base CNN models and state‐of‐the‐art approaches without using any external data.

The advent of Internet Of Things has led to the problem of an explosive outburst of data. Because of this, there is a need of an advanced data acquisition and data reduction system. We present two approaches for data reduction by sensor selection, leveraging the power of Machine Learning. The first approach uses Principal Component Analysis and the second approach uses Reinforcement Learning to identify k significant sensors from the n total sensors that closely resemble the original data statistically. This approach is applied on the data available from sensors deployed in South Brazil, as well as on the data collected from a sensor field we set up in the Computer Engineering Department. The results from both the datasets show significant data reduction while maintaining the characteristics of the original data.

Abstract Response and recovery are the two most crucial aspects associated with post disaster management. Both these operations need real-time data and location at which the event occurs. These operations require event-based collection of data at critical times. The events may occur at any place and the data may be needed from either or all of the events for strategic planning and event handling. It is a challenging task to process the events whose locations are uncertain. In this view, there is a need to know the real-time location of events occurring in the surrounding. Moreover, there is a possibility of huge amount of signal processing among the devices deployed in the terrain. In this context, an event localization algorithm is proposed based on the Direction of Arrival estimation technique in the Internet of Things environment. It estimates the location of events by mapping the deployed devices using concentric circular array in the region. Further, the Cramer–Rao bound for the proposed algorithm is derived and compared with the existing schemes for efficacy. The algorithm is implemented on real test bed and presented with comparative evaluation to validate the work.

Disaster is an uncertain phenomenon that arises due to natural as well as man-made calamities. Disaster often causes a high degree of destruction, especially in a very densely populated region. To handle such a situation, efficient resource management strategies are required. Resource management is the most crucial phase of disaster management. Efficient and in-time allocation of resources is very important; otherwise, it may result in more fatalities. In this context, we propose the resource management algorithm, which deals with both over- and underdemand for resources. Resource management requires efficient resource allocation, and in case of overdemand for resources, it must be followed by resource scheduling. In this paper, we introduce a resource allocation technique which is based on multiple objectives having a different set of constraints. We also propose the resource scheduling algorithm based on various parameters. The proposed algorithm uses multiobjective theory for resource allocation which is followed by the implementation of priority-based scheduling technique, in the case of overdemand for resources. Our proposed methods are compared to the existing approaches in the literature. From the simulation results, it is clear that our methods perform optimum resource allocation and scheduling operations.

For postdisaster management, rescue and recovery operations are very critical. It is desired that the rescue and recovery operation should be handled through efficient resource management to minimize the postdisaster effects in terms of human loss and other types of damage. Resource management requires addressing various challenging issues like scheduling and monitoring of the resources which need real time information of various activities or events occurring anytime, anywhere, and anyplace. To satisfy such requirements, Internet of Things, an advanced upcoming technology, can be utilized for resource monitoring and scheduling. In this context, we propose resource scheduling algorithm for the postdisaster management. As mentioned above various tasks of rescue and recovery operation should be carried out with different priority and there should not be deadlock while availing the resources. In our approach, we estimate the waiting time using queuing theory for the availability of the resources for different activities that are to be performed at various locations. The simulation results of the proposed method are analyzed using different standard parameters like resource utilization and the waiting time for different activities. The proposed method is further visualized through real time annotation of resources and activities represented with the help of Google maps using android based application on the smartphone. The proposed algorithm is further compared in terms of computational complexity and fairness analysis for the effective utilization of the available resources.

Real-time information access and service delivery are very crucial for handling different events and providing the quick response for different applications. These tasks can be achieved through the network of devices which enables the device-to-device communication. Internet of Things is a realization of such kind of network. The advantage of Internet of Things is that it allows us to access device at anytime and from anywhere. It directs us to provide not only location aware services but also poses a need to have better network coverage for performing a specific task using the network of devices. Location information and proper deployment of devices are very important for effective utilization of Internet of Things for a given application. In this context, two fundamental tasks, i.e. efficient deployment of devices followed by the issue of location estimation of these devices need to be addressed on priority. For this reason, in this paper, we propose two algorithms; for efficient deployment of the devices and for location estimation of these devices, respectively. For deployment, we explore quasi-random strategy and for location estimation, we consider received signal strength indicator-based approach. The proposed methods are compared with existing techniques to prove its efficacy in Internet of Things environment.

Connecting all devices through Internet is now practical due to Internet of Things. IoT assures numerous applications in everyday life of common people, government bodies, business, and society as a whole. Collaboration among the devices in IoT to bring various applications in the real world is a challenging task. In this context, we introduce an application‐based two‐layer architectural framework for IoT which consists of sensing layer and IoT layer. For any real‐time application, sensing devices play an important role. Both these layers are required for accomplishing IoT‐based applications. The success of any IoT‐based application relies on efficient communication and utilization of the devices and data acquired by the devices at both layers. The grouping of these devices helps to achieve the same, which leads to formation of cluster of devices at various levels. The clustering helps not only in collaboration but also in prolonging overall network lifetime. In this paper, we propose two clustering algorithms based on heuristic and graph, respectively. The proposed clustering approaches are evaluated on IoT platform using standard parameters and compared with different approaches reported in literature.

In recent times, graph based spectral clustering algorithms have received immense attention in many areas like, data mining, object recognition, image analysis and processing. The commonly used similarity measure in the clustering algorithms is the Gaussian kernel function which uses sensitive scaling parameter and when applied to the segmentation of noise contaminated images leads to unsatisfactory performance because of neglecting the spatial pixel information. The present work introduces a novel framework for spectral clustering which embodied local spatial information and fuzzy based similarity measure to tackle the above mentioned issues. In our approach, firstly we filter the noise components from original image by using the spatial and gray–level information. The similarity matrix is then constructed by employing a similarity measure which takes into account the fuzzy c-partition matrix and vectors of the cluster centers obtained by fuzzy c-means clustering algorithm. In the last step, spectral clustering technique is realized on derived similarity matrix to obtain the desired segmentation result. Experimental results on segmentation of synthetic and Berkeley benchmark images with noise demonstrates the effectiveness and robustness of the proposed method, giving it an edge over the clustering based segmentation method reported in the literature.

In most of the standard Pan-sharpening methods, it is observed that color distortion appears in the final Pansharped image. In the area of image analysis in remote sensing applications, spectral information plays an important role and due to the color distortion, image analysis may get affected. In this paper, a fuzzy based hybrid multispectral image fusion method using wavelet transform is proposed which provides novel tradeoff solution between the spectral and spatial fidelity and preserves more detail spectral and spatial information. New fuzzy based hybrid image fusion rules are also proposed. Proposed method is applied on several registered Panchromatic and Multispectral images and simulation results are compared using standard image fusion parameters. The simulation results of proposed method are also compared with five different standard and recently proposed Pan sharpening methods. It has been observed from simulation results that the proposed algorithm preserves better spatial and spectral information and better visual quality as compared to earlier reported methods.

Neuro-fuzzy systems have been used in past years for robot navigation applications because of their ability to learn human expertise and to utilize this knowledge to develop autonomous navigation strategies. In this paper, neuro-fuzzy based systems are developed for behavior based control of a mobile robot for reactive navigation. The proposed systems transform sensors' input to yield wheel velocities. Novel algorithms are proposed for a) to find the range of the mobile robots from nearby obstacles and b) to generate training pairs for neural network, optimally. With a view to ascertain the efficacy of proposed system; developed neuro-fuzzy system's performance is compared to neural and fuzzy based approaches. Simulation results show effectiveness of proposed system in all kind of obstacle environments.

This paper proposes a novel region based image fusion scheme based on high boost filtering concept using discrete wavelet transform. In the recent literature, region based image fusion methods show better performance than pixel based image fusion method. Proposed method is a novel idea which uses high boost filtering concept to get an accurate segmentation using discrete wavelet transform. This concept is used to extract regions from input registered source images which are then compared with different fusion rules. The new MMS fusion rule is also proposed to fuse multimodality images. The different fusion rules are applied on various categories of input source images and resultant fused image is generated. Proposed method is applied on large number of registered images of various categories of multifocus and multimodality images and results are compared using standard reference based and nonreference based image fusion parameters. It has been observed from simulation results that our proposed algorithm is consistent and preserves more information compared to earlier reported pixel based and region based methods.

Data association and model selection are important factors for tracking multiple targets in a dense clutter environment without using a priori information about the target dynamic. We propose a neural-network-based tracking algorithm, incorporating a interacting multiple model and show that it is possible to track both maneuvering and nonmaneuvering targets simultaneously in the presence of dense clutter. Moreover, it can be used for real-time application. The proposed method overcomes the problem of data association by using the method of expectation maximization and Hopfield network to evaluate assignment weights. All validated observations are used to update the target state. In the proposed approach, a probability density function (pdf) of an observed data, given target state and observation association, is treated as a mixture pdf. This allows to combine the likelihood of an observation due to each model, and the association process is defined to incorporate an interacting multiple model, and consequently, it is possible to track any arbitrary trajectory

We propose an effective technique using a wavelet-based temporal decomposition algorithm to detect single-pixel targets with motion from frame to frame. We next integrate the proposed detection algorithm with an interacting multiple-model method and multiple filter bank approach to provide an effective solution for tracking multiple single-pixel nonmaneuvering and maneuvering targets. Through Monte Carlo simulations, we establish the efficiency and robustness of the proposed approach. Based on exhaustive empirical study, we demonstrate the effectiveness of our proposed approach in tracking multiple single-pixel targets in a sequence of infrared images with clutter and occlusion due to moving clouds in airborne applications.

Detection and tracking of approaching targets in an infrared (IR) image sequence is important for surveillance applications. In this paper an algorithm is proposed which provides a complete solution (track while scan) for detection and tracking for IRST system. The proposal method uses only motion as a cue to detect the target. Detection is followed by tracking. In a real scenario, the movement of a target is arbitrary and no a priori information is available. We propose a tracking method which tracks maneuvering and nonmaneuvering targets simultaneously using a filter bank. The switch-over amongst the filters is based on a single-step decision logic.

We propose an algorithm for detection and tracking of small and fast moving objects, like a ping pong ball or a cricket ball, in sports video sequences. For detection, the proposed method uses only motion as a cue; moreover it does not use any texture information. Our method is able to detect the object with very low contrast and negligible texture content. Along with detection, we also propose a tracking algorithm using the multiple filter bank approach. Thus we provide a complete solution. The tracking algorithm is able to track maneuvering as well as non-maneuvering movements of the object without using any a priori information about the target dynamics

Data association and model selection are important factors for tracking multiple targets in a dense clutter environment. We propose a sequential probabilistic multiple hypotheses tracking (PMHT) based algorithm using interacting multiple modelling (IMM), namely the IMM-PMHT algorithm. Inclusion of IMM enables any arbitrary trajectory to be tracked without any a priori information about the target dynamics. IMM allows us to incorporate different dynamic models for the targets and PMHT helps to avoid the uncertainty about the measurement origin. It operates in an iterative mode using an expectation-maximization (EM) algorithm. The proposed algorithm uses only measurement association as missing data, which simplifies E-step and M-step. It is computationally more efficient, and an important characteristic of our proposed algorithm is that it operates in a single batch model, i.e. sequential, and hence can be used for real time tracking.

Performance of any tracking algorithm depends upon the model selected to capture the target dynamics. In real world applications, no a priori knowledge about the target motion is available. Moreover, it could be a maneuvering target. The proposed method is able to track maneuvering or nonmaneuvering multiple point targets with large motion (/spl plusmn/20 pixels) using multiple filter bank in an IR image sequence in the presence of clutter and occlusion due to clouds. The use of multiple filters is not new, but the novel idea here is that it uses single-step decision logic to switch over between filters. Our approach does not use any a priori knowledge about maneuver parameters, nor does it exploit a parameterized nonlinear model for the target trajectories. This is in contrast to: (i) interacting multiple model (IMM) filtering which required the maneuver parameters, and (ii) extended Kalman filter (EKF) or unscented Kalman filter (UKF), both of which require a parameterized model for the trajectories. We compared our approach for target tracking with IMM filtering using EKF and UKF for nonlinear trajectory models. UKF uses the nonlinearity of the target model, where as a first order linearization is used in case of the EKF. RMS for the predicted position error (RMS-PPE) obtained using our proposed methodology is significantly less in case of highly maneuvering target.

Performance of any tracking algorithm depends upon the model selected to capture the target dynamics. In real world applications, no a priori knowledge about the target motion is available. Moreover, it could be a maneuvering target. The proposed method is able to track maneuvering or nonmaneuvering multiple point targets with large motion (/spl mnplus/20 pixels) using multiple filter bank in an IR image sequence in the presence of clutter and occlusion due to clouds. The use of multiple filters is not new, but the novel idea here is that it uses single-step decision logic to switch over between filters. Our approach does not use any a priori knowledge about maneuver parameters, nor does it exploit a parameterized nonlinear model for the target trajectories. This is in contrast to: (i) interacting multiple model (IMM) filtering which required the maneuver parameters, and (ii) extended Kalman filter (EKF) or unscented Kaiman filter (UKF), both of which require a parameterized model for the trajectories. We compared our approach for target tracking with IMM filtering using EKF and UKF for non-linear trajectory models. UKF uses the nonlinearity of the target model, where as a first order linearization is used in case of EKF. RMS for the predicted position error (RMS-PPE) obtained using our proposed methodology is significantly less in case of highly maneuvering target.