Search

Wireless multimedia sensor networks (WMSNs) face challenges such as high energy consumption, latency, and ensuring secure data transmission, especially for real-time multimedia applications. To address these issues, this study proposes the enhanced adaptive transmission control algorithm (EATCA), which integrates a bidirectional long short-term memory (LSTM) model and chaos-based lightweight cryptography. EATCA optimizes energy consumption, reduces latency, and improves data security through the use of adaptive clustering and predictive transmission techniques. The algorithm uses meta-heuristic optimization for cluster head selection and fuzzy logic for route prioritization to ensure efficient data transmission. A chaos-based encryption scheme ensures data confidentiality with minimal computational overhead. The simulation results show that EATCA achieves 43.53 % lower energy consumption compared to traditional low energy adaptive cluster hierarchy (LEACH) protocols, a 96 % packet delivery ratio (PDR), and a 42 ms latency, making it suitable for time-sensitive multimedia applications. The system also achieves an 85 % compression ratio while maintaining security overhead as low as 7 %. These results make EATCA a reliable and energy-efficient solution for optimizing WMSNs and pave the way for scalable and secure network implementations.

The active theme for exploration is currently the potential for visual gaze tracking and visual human–machine interactions in contemporary user devices. Eye views were used as an input tool to extract human behavioral knowledge and to achieve immersive user interactions in augmented and amplified reality systems. However, implementations of gaze on consumer devices operating under controlled circumstances face tough precision and reliability challenges. For evolving automotive electronic systems such as driver surveillance systems and new touch screens, accurate and successful gaze estimates are significant. Such systems have to work efficiently and at minimum cost in demanding unregulated, low-power conditions. Deep learning is omnipresent for most computer vision tasks, but it is still common to estimate the eye gaze. The realistic account of all challenges and opportunities of eye gaze tracking requires a complete study of available deep learning methods. In this chapter, huge potential efforts are made to study the background of eye gaze tracking with the detail of how deep learning made a contribution to computer vision-based tracking. In the end, this chapter also highlights a generic system model and comparison of various algorithms for deep learning-driven eye gaze direction diagnosis. [ABSTRACT FROM AUTHOR]

Internet of things (IoT) has necessitated the development of indoor photovoltaics to enable a web of self-powered wireless sensors/nodes. We analysed a CsPbI3 wide band gap perovskite for indoor photovoltaic application. An Indoor photovoltaic (IPV) device based on CsPbI3 showed a theoretical efficiency of 51.5% at a band gap of 1.8 eV under indoor light-emitting diode (LED) illumination. This high efficiency is due to better capture of the narrow emission spectrum of LED by a high band gap CsPbI3 absorber. Under low luminance of indoor light sources, there is a low density of photogenerated carriers, which increases the ratio of trapped electrons to photogenerated electrons. The low photogenerated carrier density lowered bulk recombination, and the high trapped electrons to photogenerated electrons ratio increases IPV sensitivity toward interfacial recombination. Finally, the device optimization strategies of the IPV device, distinct from outdoor illumination devices are highlighted. [ABSTRACT FROM AUTHOR]

Herein, a theoretical investigation is conducted on grain‐size inhomogeneity's impact and grain boundaries' (GBs') electrical nature in thin‐film solar cells. Using the Matthiessen rule, grain‐size‐dependent mobility is derived in polycrystalline material. The obtained grain‐size‐dependent mobility values are fed into the Poisson solver to calculate device performance. The severity of grain sizes in the lower region determines how grain size affects the photovoltaic performance by grain‐size‐dependent efficiency simulation. Low grain sizes become critical, especially for low‐thickness absorbers. The second aspect of the study assesses potential variation at GBs to reveal the impact of the electrical properties of GBs. Evidence shows that the acceptor defects at GB are benign for device performance, causing upward band bending at the GB and acting as electron barriers. Device performance is adversely affected by donor defects at GBs due to downward band bending. As summarized in the findings, the polycrystallinity‐induced cause–effect relationships of grains are likely to interest solar cell researchers. [ABSTRACT FROM AUTHOR]

Brain tumors are still diagnosed and classified based on the results of histopathological examinations of biopsy samples. The existing method requires extra effort from the user, takes too long, and can lead to blunders. These limitations underline the need of employing a fully automated deep learning system for the multi‐classification of brain tumors. In order to facilitate early detection, this study employs a convolutional neural network (CNN) to multi‐classify brain tumors. In this research, we present three distinct CNN models for use in three separate categorization tasks. The first CNN model can correctly categorize brain tumors 99.74% of the time. The second CNN model is 96.27% accurate in differentiating between normal, glioma, meningioma, pituitary, and metastatic brain tumors. The third CNN model successfully distinguishes between Grades II, III, and IV brain tumors 99.18% of the time. The Hybrid Particle Swarm Grey Wolf Optimization (HPSGWO) technique is used to quickly and accurately determine optimal values for all of CNN models most important hyperparameters. An HPSGWO algorithm is used to fine‐tune all the necessary hyperparameters for optimal classification performance. The results are compared with standard existing CNN models across a range of performance measures. The proposed models are trained using publicly available large clinical datasets. To verify their initial multi‐classification of brain tumors, clinicians and radiologists might use the proposed CNN models. [ABSTRACT FROM AUTHOR]