Oceans cover more than 70% of the Earth’s surface. For various reasons, almost 95% of these areas remain unexplored. Underwater wireless communication (UWC) has widespread applications, including real-time aquatic data collection, naval surveillance, natural disaster prevention, archaeological expeditions, oil and gas exploration, shipwreck exploration, maritime security, and the monitoring of aquatic species and water contamination. The promising concept of the Internet of Underwater Things (IoUT) is having a great influence in several areas, for example, in small research facilities and average-sized harbors, as well as in huge unexplored areas of ocean. The IoUT has emerged as an innovative technology with the potential to develop a smart ocean. The IoUT framework integrates different underwater communication techniques such as optical, magnetic induction, and acoustic signals. It is capable of revolutionizing industrial projects, scientific research, and business. The key enabler technology for the IoUT is the underwater wireless sensor network (UWSN); however, at present, this is characterized by limitations in reliability, long propagation delays, high energy consumption, a dynamic topology, and limited bandwidth. This study examines the literature to identify potential challenges and risks, as well as mitigating solutions, associated with the IoUT. Our findings reveal that the key contributing elements to the challenges facing the IoUT are underwater communications, energy storage, latency, mobility, a lack of standardization, transmission media, transmission range, and energy constraints. Furthermore, we discuss several IoUT applications while highlighting potential future research directions.

The enhanced proliferation of connected entities needs a deployment of innovative technologies for the next generation wireless networks. One of the critical concerns, however, is the spectrum scarcity, due to the unprecedented broadcast penetration rate nowadays. Based on this, visible light communication (VLC) has recently emerged as a viable solution to secure high-speed communications. VLC, a high data rate communication technology, has proven its stature as a promising complementary to its radio frequency (RF) counterpart. VLC is a cost-effective, energy-efficient, and secure technology that exploits the current infrastructure, specifically within indoor and underwater environments. Yet, despite their appealing capabilities, VLC systems face several limitations which constraint their potentials such as LED’s limited bandwidth, dimming, flickering, line-of-sight (LOS) requirement, impact of harsh weather conditions, noise, interference, shadowing, transceiver alignment, signal decoding complexity, and mobility issue. Consequently, non-orthogonal multiple access (NOMA) has been considered an effective technique to circumvent these shortcomings. The NOMA scheme has emerged as a revolutionary paradigm to address the shortcomings of VLC systems. The potentials of NOMA are to increase the number of users, system’s capacity, massive connectivity, and enhance the spectrum and energy efficiency in future communication scenarios. Motivated by this, the presented study offers an overview of NOMA-based VLC systems. This article provides a broad scope of existing research activities of NOMA-based VLC systems. This article aims to provide firsthand knowledge of the prominence of NOMA and VLC and surveys several NOMA-enabled VLC systems. We briefly highlight the potential and capabilities of NOMA-based VLC systems. In addition, we outline the integration of such systems with several emerging technologies such as intelligent reflecting surfaces (IRS), orthogonal frequency division multiplexing (OFDM), multiple-input and multiple-output (MIMO) and unmanned aerial vehicles (UAVs). Furthermore, we focus on NOMA-based hybrid RF/VLC networks and discuss the role of machine learning (ML) tools and physical layer security (PLS) in this domain. In addition, this study also highlights diverse and significant technical hindrances prevailing in NOMA-based VLC systems. We highlight future research directions, along with provided insights that are envisioned to be helpful towards the effective practical deployment of such systems. In a nutshell, this review highlights the existing and ongoing research activities for NOMA-based VLC systems, which will provide sufficient guidelines for research communities working in this domain and it will pave the way for successful deployment of these systems.

Early clinical diagnosis and treatment of disease rely heavily on measuring the many various types of medical information that are scattered throughout the body. Continuous and accurate monitoring of the human body is required in order to identify abnormal medical signals and to locate the factors that contribute to their occurrence in a timely manner. In order to fulfill this requirement, a variety of battery-free and self-powered methods of information collecting have been developed. For the purpose of a health monitoring system, this paper presents smart wearable sensors that are based on triboelectric nanogenerators (TENG) and piezoelectric nanogenerators (PENG), as well as hybrid nanogenerators that combine piezoelectric and triboelectric nanogenerators (PTNG). Following the presentation of the PENG and TENG principles, a summary and discussion of the most current developments in self-powered medical information sensors with a variety of purposes, structural designs, and electric performances follows. Wearable sensors that generate their own electricity are crucial not only for the proper development of children and patients with unique conditions, but for the purpose of maintaining checks on the wellbeing of the elderly and those who have recently recovered from illness, and for administering any necessary medical care. This work sought to do two things at once: provide perspectives for health monitoring, and open up new avenues for the analysis of long-distance biological movement status.

This paper proposes an image denoising training framework based on Wasserstein Generative Adversarial Networks (WGAN) and applies it to cell image denoising. Cell image denoising is a challenging task which has high requirement on the recovery of feature details. Current popular convolutional neural network (CNN) based denoising methods encounter a blurriness issue that denoised images are blurry on texture details, which is fatal for the cell image denoising. In this paper, to solve the blurriness issue, we first theoretically analyze the cause of the blurriness issue. Subsequently, an image denoising training framework with WGAN based adversarial learning is proposed. This training framework solves the blurriness issue by guiding the denoising network to find the distribution space of real clean images rather than the distribution space of blurry images and introducing feature information. Experimental results show that this training framework can effectively solve the blurriness issue and achieve better denoising performance than the state-of-the-art denoising methods. Meanwhile, the application of this training framework on cell image denoising also achieves satisfactory performance. Recovered cell images of this training framework are clear on feature details.

Facial landmark detection is a key component of the face recognition pipeline as well as facial attribute analysis and face verification. Recently convolutional neural network-based face alignment methods have achieved significant improvement, but occlusion is still a major source of a hurdle to achieve good accuracy. In this paper, we introduce the attentioned distillation module in our previous work Occlusion-adaptive Deep Network (ODN) model, to improve performance. In this model, the occlusion probability of each position in high-level features are inferred by a distillation module. It can be learnt automatically in the process of estimating the relationship between facial appearance and facial shape. The occlusion probability serves as the adaptive weight on high-level features to reduce the impact of occlusion and obtain clean feature representation. Nevertheless, the clean feature representation cannot represent the holistic face due to the missing semantic features. To obtain exhaustive and complete feature representation, it is vital that we leverage a low-rank learning module to recover lost features. Considering that facial geometric characteristics are conducive to the low-rank module to recover lost features, the role of the geometry-aware module is, to excavate geometric relationships between different facial components. The role of attentioned distillation module is, to get rich feature representation and model occlusion. To improve feature representation, we used channel-wise attention and spatial attention. Experimental results show that our method performs better than existing methods.