The networking of virtual reality applications will play an important role in the emerging global Internet of Things (IoT) framework and it is expected to provide the foundation of the expected 5G tactile Internet ecosystem. However, considerable challenges are ahead in terms of technological constraints and infrastructure costs. The raw data rate (5 Gbps–60 Gbps) required achieving an online immersive experience that is indistinguishable from real life vastly exceeds the capabilities of future broadband networks. Therefore, simply providing high bandwidth is insufficient in compensating for this difference, because the demands for scale and supply vary widely. This requires exploring holistic solutions that exceed the traditional network domain, and integrating virtual reality (VR) data capture, encoding, network, and user navigation. Emerging services are extremely inefficient in terms of mass use and data management, which significantly reduces the user experience, due to their heuristic design choices. Other key aspects must be considered, such as wireless operation, ultra-low latency, client/network access, system deployment, edge computing/cache, and end-to-end reliability. A vast number of high-quality works have been published in this area and they will be highlighted in this survey. In addition to a thorough summary of recent progress, we also present an outlook of future developments in the quality of immersive experience networks and unified data set measurement in VR video transmission, focusing on the expansion of VR applications, security issues, and business issues, which have not yet been addressed, and the technical challenges that have not yet been completely solved. We hope that this paper will help researchers and developers to gain a better understanding of the state of research and development in VR.

With the advent of the information age, VR video streaming services have emerged in large numbers in scenarios such as immersive entertainment, smart education, and the Internet of Vehicles. People are also demanding an increasing number of virtual-reality (VR) services, and service providers must ensure a good user experience. Therefore, the quality of the VR user’s experience is receiving increasing attention from academia and industry. The review in this paper focuses on a comprehensive summary of the current state of quality-of-experience (QoE) technologies applied to VR video streaming. First, we review the main influencing factors of QoE and VR video streaming. Second, the user QoE for VR evaluation is discussed. Third, the modeling of QoE for VR video streaming, the QoE-oriented VR optimization problem, and enabling techniques of machine learning for VR video streaming improvement are summarized. Lastly, we present current challenges and possible future research directions.

Maritime small object detection is critical for UAV-based sea surveillance but remains challenging due to the small size of targets and interference from sea reflections and waves. This paper proposes SeaLSOD-YOLO, a lightweight detection algorithm based on YOLOv11, designed to improve small object detection accuracy while maintaining real-time performance. The method incorporates four key modules: Shallow Multi-scale Output Reconstruction, which fuses shallow and mid-level features to preserve fine-grained details; SPPF-FD, which combines spatial pyramid pooling with frequency-domain adaptive convolution to enhance sensitivity to high-frequency textures and suppress sea-surface interference; attention-based feature fusion, which emphasizes small object features through channel and spatial attention; and dynamic multi-scale sampling, which optimizes feature representation across different scales. Experiments on the SeaDroneSee dataset demonstrate that, compared with YOLOv11s, the proposed method improves precision from 75.6% to 81.9%, recall from 62.6% to 73.5%, and mAP@0.5 from 67.9% to 77.0%. The mAP@0.5:0.95 also increases from 41.1% to 44.9%. The model achieves an inference speed of 256 FPS. Although the parameter size increases from 18.2 MB to 30.8 MB, the method maintains a favorable balance between detection accuracy and computational efficiency. Comparative evaluation further shows superior performance in detecting small maritime objects such as buoys and lifeboats. These results indicate that SeaLSOD-YOLO effectively balances accuracy, efficiency, and real-time capability in complex maritime environments. Future work will focus on further optimization of attention mechanisms and upsampling strategies to enhance the detection of extremely small targets.

With the rapid development of virtual reality (VR) video networked applications, the use of network caching mechanisms to guarantee the quality of VR services has been proven to be a very effective method. Most of the existing methods on cache placement prediction only consider the one-sided information of user viewpoints and do not consider the video characteristic information of virtual reality, because the asymmetry of the two types of information causes the accuracy of current predictions to gradually decrease, which affects the cache hit rate and leads to VR performance metrics that cannot be guaranteed. In this paper, we analyze the demanding requirements of VR for low latency and high bandwidth in a multi-access point (multi-AP) scenario environment, and further improve the cache hit rate of user requests by increasing network throughput. First, the throughput of VR users after associating APs is analyzed using a Markov model. Second, a nonlinear mixed integer programming problem is constructed with the goal of maximizing the overall throughput of the network system. Finally, combining the characteristics of the VR video content itself and the popularity of the requested video content, the symmetry of the information is guaranteed by considering the ratio between the video characteristic information and the user feature information to determine the weights. The experimental results demonstrate that the proposed algorithm achieves the improvement of cache hit rate and the improvement of network throughput while ensuring the quality of service.

In industrial vision, the lack of defect samples is one of the key constraints of depth vision quality inspection. This paper mainly studies defect detection under a small training set, trying to reduce the dependence of the model on defect samples by using normal samples. Therefore, we propose a Knowledge-Embedding Relational Network. We propose a Knowledge-Embedding Relational Network (KRN): firstly, unsupervised clustering and convolution features are used to model the knowledge of normal samples; at the same time, based on CNN feature extraction assisted by image segmentation, the conv feature is obtained from the backbone network; then, we build the relationship between knowledge and prediction samples through covariance, embed the knowledge, further mine the correlation using gram operation, normalize the power of the high-order features obtained by covariance, and finally send them to the prediction network. Our KRN has three attractive characteristics: (I) Knowledge Modeling uses the unsupervised clustering algorithm to statistically model the standard samples so as to reduce the dependence of the model on defect data. (II) Covariance-based Knowledge Embedding and the Gram Operation capture the second-order statistics of knowledge features and predicted image features to deeply mine the robust correlation. (III) Power Normalizing suppresses the burstiness of covariance module learning and the complexity of the feature space. KRN outperformed several advanced baselines in small training sets on the DAGM 2007, KSDD, and Steel datasets.