Search

Despite the current small object detection task has achieved significant improvements, it still suffers from some problems. For example, it is a challenge to extract small object features because of little information in the scene of small objects, which may lose the original feature information of small object, resulting in poor detection results. To address this problem, this paper proposes a two-stage calculation Transformer (TCT) based small object detection network. Firstly, a two-stage calculation Transformer is embedded in the backbone feature extraction network for feature enhancement. Based on the traditional Transformer values computation, multiple 1D dilated convolutional layer branches with different feature fusions are utilized to implement global self-attention for the purpose of improving the feature representation and information interaction. Secondly, this paper proposes an effective residual connection module to improve the low-efficiency convolution and activation of the current CSPLayer, which helps to advance the information flow and learn more rich contextual details. Finally, this paper proposes a feature fusion and refinement module for fusing multi-scale features and improving the target feature representation capability. Quantitative and qualitative experiments on PASCAL VOC2007+2012 dataset, COCO2017 dataset and TinyPerson dataset show that the proposed algorithm has better ability of target feature extraction and higher detection accuracy for small target detection, compared with YOLOX.

Topochemical transformation has emerged as a promising method for fabricating two-dimensional (2D) materials with precise control over their composition and morphology. However, the large-scale synthesis of ultrathin 2D materials with controllable thickness remains a tremendous challenge. Herein, we adopt an efficient topochemical synthesis strategy, employing a confined reaction space to fabricate ultrathin 2D Sn4P3 nanosheets in large-scale. By carefully adjusting the rolling number during the processing of Sn/Al foils, we have successfully fabricated Sn4P3 nanosheets with varied layer thicknesses, achieving a remarkable minimum thickness of two layers (~ 2.2 nm). Remarkably, the bilayer Sn4P3 nanosheets display an exceptional initial capacity of 1088 mAh·g−1, nearing the theoretical value of 1230 mAh·g−1. Furthermore, we reveal their high-rate property as well as outstanding cyclic stability, maintaining capacity without fading more than 3000 cycles. By precisely controlling the layer thickness and ensuring nanoscale uniformity, we enhance the lithium cycling performance of Sn4P3, marking a significant advancement in developing high-performance energy storage systems. [ABSTRACT FROM AUTHOR]

The tolerance requirement of lithium metal batteries in harsh environments presents great challenges to electrode materials and electrolytes because temperature plays a significant effect in electrochemical processes. In this study, a new artificial layer on a copper current collector that boosts the de‐solvation kinetics and provides electrostatic shielding effects is presented to enhance the electrochemical performance of lithium metal batteries. This new artificial layer is constructed with poly(diallyl dimethyl ammonium chloride) (PDDA) and exfoliated montmorillonite (MMT) nanosheets, which combine the advantages of both inorganic clay and organic polymer. Within this protective hybrid layer, the PDDA cations increase the interlayer spacing of MMT, broadening the diffusion pathways of Li+ and accelerating their fast diffusion. Moreover, the PDDA‐MMT protective layer facilitates Li+ de‐solvation at the interface of the Li anode and electrolyte, enabling the rapid and reversible plating/stripping of lithium metal. As a result, the as‐prepared PDDA‐MMT@Cu anode exhibits excellent stability, and good rate performance is achieved in a commercial electrolyte in the temperature range of −20–60 °C. By combining enhanced diffusion kinetics and electrostatic shielding as an artificial protective layer, this clay‐polymer composite provides a synergistic interaction and offers new inspiration for the development of lithium metal batteries. [ABSTRACT FROM AUTHOR]

In this big data era, the multi-tenant cloud platform plays an important role. However, as one of the major technologies adopted in the cloud platforms, the virtualization technology is not secure enough. The Hypervisor is a key layer in the virtualization software stack that manages vital tasks between guest virtual machines and the bare metal, such as resource allocation, sharing and isolation. Meanwhile, the Hypervisor suffers from vulnerabilities along with its large attack surface, which makes attacks on the Hypervisor threaten the cloud and applications above. Therefore, constructing protection mechanisms for the untrusted Hypervisor in the cloud is necessary. Survey from the perspective of the protection mechanism's construction: analyze the feasibility and challenges, and do cla-ssifications; present the related work in this field in terms of integrity detection mechanisms, defense mechanisms, and isolation mechanisms; give research trends and provide a valuable reference for future researches on the vir-tualization security and on building a reliable multi-tenant cloud platform.

The photo rechargeable device (PRD) has been continuously drawing attention because it combines energy conversion and storage in one device. As for the photoelectrode of PRD, the construction of heterojunction is of crucial importance to enhance the photo performance. In this work, a two-electrode photo rechargeable capacitor based on the p–n heterojunction of ZnO/ZIF-67 is fabricated. ZIF-67 not only serves as the energy storage material but also forms the p–n heterojunction together with ZnO. A fast volatilization method was adopted for the in situ growth of ZIF-67 on ZnO nanorods to ensure sufficient mass loading and fewer interface defects. The results show a photovoltage of 0.36 V (0.2 V higher than single ZnO), a specific capacitance of 759.0 mF/g, and an overall energy conversion efficiency of 0.49%. The enhanced photovoltage is attributed to the p–n heterojunction. Moreover, a practical button cell was also fabricated, with 91% Coulombic efficiency remaining after 3000 cycles in the dark. [ABSTRACT FROM AUTHOR]

Electrochemically active metal anodes, such as lithium, sodium, potassium, and zinc, have attracted great research interests in the advanced rechargeable batteries owing to their superior theoretical energy densities. Unfortunately, the metal anodes suffer from the huge volume changes with loss of active materials during the plating and stripping processes, resulting in fast capacity decay. Moreover, the random growth of dendrites on the metal anodes will penetrate the separator, causing severe safety issues. Engineering metal anodes by introducing the 2D materials are widely investigated to alleviate these issues. Benefitting from the ultrathin structure feature and unique electrical properties, 2D materials are regarded as one of the best host of metal anodes. Besides, the tunable active sites on basal plane enable 2D materials to achieve favorable interaction with metal anodes. Moreover, some 2D materials exhibit good mechanical strength and flexibility, serving as building block for the artificial solid electrolyte interphase. In this review, we mainly disclosed the correlations between the intrinsic properties of 2D materials and their functions in guiding uniform nucleation, controlling the growth of metals, and accommodating the volume change. Also, the challenges of 2D materials in metal anodes are well discussed. Finally, the future directions to develop high‐performance metal anodes by taking advantage of these unique features of 2D materials are proposed. [ABSTRACT FROM AUTHOR]