基于推理注意力机制的二阶段网络图像修复

现有的图像修复方法在处理大面积缺失或高度纹理化的图像时，通常会产生扭曲的结构或与周围区域不一致的模糊纹理，无法重建合理的图像结构。为此，提出了一种基于推理注意力机制的二阶段网络图像修复方法。首先通过边缘生成网络生成合理的幻觉边缘信息，然后在图像补全网络完成图像的重建工作。为了进一步生成视觉效果更逼真的图像，提高图像修复的精确度，在图像补全网络采用推理注意力机制，有效控制了生成特征的不一致性，从而生成更有效的信息。所提方法在多个数据集上进行了实验验证，结果表明该图像修复方法的结构相似性指数达到了88.9%，峰值信噪比达到了25.56 dB，与现有的图像修复方法相比，该方法具有更高的图像修复精确度，生成的图像更逼真。     

Dilated and soft attention-guided convolutional neural network for breast cancer histology images classification

Breast cancer is one of the most common types of cancer in women, and histopathological imaging is considered the gold standard for its diagnosis. However, the great complexity of histopathological images and the considerable workload make this work extremely time-consuming, and the results may be affected by the subjectivity of the pathologist. Therefore, the development of an accurate, automated method for analysis of histopathological images is critical to this field. In this article, we propose a deep learning method guided by the attention mechanism for fast and effective classification of haematoxylin and eosin-stained breast biopsy images. First, this method takes advantage of DenseNet and uses the feature map's information. Second, we introduce dilated convolution to produce a larger receptive field. Finally, spatial attention and channel attention are used to guide the extraction of the most useful visual features. With the use of fivefold cross-validation, the best model obtained an accuracy of 96.47% on the BACH2018 dataset. We also evaluated our method on other datasets, and the experimental results demonstrated that our model has reliable performance. This study indicates that our histopathological image classifier with a soft attention-guided deep learning model for breast cancer shows significantly better results than the latest methods. It has great potential as an effective tool for automatic evaluation of digital histopathological microscopic images for computer-aided diagnosis.     

爆炸排淤填石法中淤泥的本构模型

以爆炸排淤填石法为背景 ,对不同应变率阶段淤泥的本构模型进行了分析 ,利用ANSYS/LS-DYNA动态有限元分析程序对所选择的模型进行验证和确认。计算和分析结果表明 :在形成爆炸空腔的高应变率阶段 ,淤泥表现为理想不可压缩流体的性质 ;在小药量小抵抗线条件下 ,甚至在淤泥自重作用下的较低应变率变形阶段 ,其黏性效应也可以忽略不计。     

Structure-sensitive semantic matching for aggregate question answering over knowledge base

Aggregate question answering essentially returns answers for given questions by obtaining query graphs with unique dependencies between values and corresponding objects. Word order dependency, as the key to uniquely identify dependency of the query graph, reflects the dependencies between the words in the question. However, due to the semantic gap caused by the expression difference between questions encoded with word vectors and query graphs represented with logical formal elements, it is not trivial to match the correct query graph for the question. Most existing approaches design more expressive query graphs for complex questions and rank them just by directly calculating their similarities, ignoring the semantic gap between them. In this paper, we propose a novel Structure-sensitive Semantic Matching(SSM) approach that learns aligned representations of dependencies in questions and query graphs to eliminate their gap. First, we propose a cross-structure matching module to bridge the gap between two modalities(i.e., textual question and query graph). Then, we propose an entropy-based gated AQG filter to remove the structural noise caused by the uncertainty of dependencies. Finally, we present a two-channel query graph representation that fuses the semantics of abstract structure and grounding content of the query graph explicitly. Experimental results show that SSM could learn aligned representations of questions and query graphs to eliminate the gaps between their dependencies, and improves up to 12% (F1 score) on aggregation questions of two benchmark datasets.     

Electrochemical Quantification of Neurotransmitters in Single Live Cell Vesicles Shows Exocytosis is Predominantly Partial

Exocytosis plays an essential role in the communication between cells in the nervous system. Understanding the regulation of neurotransmitter release during exocytosis and the amount of neurotransmitter content that is stored in vesicles is of importance, as it provides fundamental insights to understand how the brain works and how neurons elicit a certain behavior. In this minireview, we summarize recent progress in amperometric measurements for monitoring exocytosis in single cells and electrochemical cytometry measurements of vesicular neurotransmitter content in individual vesicles. Important steps have increased our understanding of the different mechanisms of exocytosis. Increasing evidence is firmly establishing that partial release is the primary mechanism of release in multiple cell types.     

4D Printing of Shape Memory Materials for Textiles: Mechanism,Mathematical Modeling,and Challenges

Shape memory materials (SMMs) in 3D printing (3DP) technology garnered much attention due to their ability to respond to external stimuli, which direct this technology toward an emerging area of research, “4D printing (4DP) technology.” In contrast to classical 3D printed objects, the fourth dimension, time, allows printed objects to undergo significant changes in shape, size, or color when subjected to external stimuli. Highly precise and calibrated 4D materials, which can perform together to achieve robust 4D objects, are in great demand in various fields such as military applications, space suits, robotic systems, apparel, healthcare, sports, etc. This review, for the first time, to the best of the authors’ knowledge, focuses on recent advances in SMMs (e.g., polymers, metals, etc.) based wearable smart textiles and fashion goods. This review integrates the basic overview of 3DP technology, fabrication methods, the transition of 3DP to 4DP, the chemistry behind the fundamental working principles of 4D printed objects, materials selection for smart textiles and fashion goods. The central part summarizes the effect of major external stimuli on 4D textile materials followed by the major applications. Lastly, prospects and challenges are discussed, so that future researchers can continue the progress of this technology.     

High-temperature oxidation behavior of monolithic Zr3[Al(Si)]4C6 and its composite incorporating 30vol% ZrB2-SiC: Providing a basis for broadening the service temperature

To provide a basis for the high-temperature oxidation of ultra-high temperature ceramics (UHTCs), the oxidation behavior of Zr₃[Al(Si)]₄C₆ and a novel Zr₃[Al(Si)]₄C₆-ZrB₂-SiC composite at 1500 °C were investigated for the first time. From the calculation results, the oxidation kinetics of the two specimens follow the oxidation dynamic parabolic law. Zr₃[Al(Si)]₄C₆ exhibited a thinner oxide scale and lower oxidation rate than those of the composite under the same conditions. The oxide scale of Zr₃[Al(Si)]₄C₆ exhibited a two-layer structure, while that of the composite exhibited a three-layer structure. Owing to the volatilization of B₂O₃ and the active oxidation of SiC, a porous oxide layer formed in the oxide scale of the composite, resulting in the degradation of its oxidation performance. Furthermore, the cracks and defects in the oxide scale of the composite indicate that the reliability of the oxide scale was poor. The results support the service temperature of the obtained ceramics.     

A Structure−Activity Relationship Study of Novel Hydroxamic Acid Inhibitors around the S1 Subsite of Human Aminopeptidase N

Aminopeptidase N (APN/CD13) is a zinc-dependent ubiquitous transmembrane ectoenzyme that is widely present in different types of cells. APN is one of the most extensively studied metalloaminopeptidases as an anti-cancer target due to its significant role in the regulation of metastasis and angiogenesis. Previously, we identified a potent and selective APN inhibitor, N-(2-(Hydroxyamino)-2-oxo-1-(3′,4′,5′-trifluoro-[1,1′-biphenyl]-4-yl)ethyl)-4-(methylsulfonamido)benzamide ( 3 ). Herein, we report the further modifications performed to explore SAR around the S1 subsite of APN and to improve the physicochemical properties. A series of hydroxamic acid analogues were synthesised, and the pharmacological activities were evaluated in vitro. N-(1-(3′-Fluoro-[1,1′-biphenyl]-4-yl)-2-(hydroxyamino)-2-oxoethyl)-4-(methylsulfonamido)benzamide ( 6 f ) was found to display an extremely potent inhibitory activity in the sub-nanomolar range.     

基于阶段注意力机制的电力线提取算法

电力系统维护是电力系统稳定运行的重要保障，应用智能算法的无人机电力巡检则为电力系统维护提供便捷。电力线提取是自主电力巡检以及保障飞行器低空飞行安全的关键技术，结合深度学习理论进行电力线提取是电力巡检的重要突破点。本文将深度学习方法用于电力线提取任务，结合电力线图像特点嵌入改进的图像输入策略和注意力模块，提出一种基于阶段注意力机制的电力线提取模型（SA-Unet）。本文提出的SA-Unet模型编码阶段采用阶段输入融合策略（Stage input fusion strategy， SIFS），充分利用图像的多尺度信息减少空间位置信息丢失。解码阶段通过嵌入阶段注意力模块（Stage attention module，SAM）聚焦电力线特征，从大量信息中快速筛选出高价值信息。实验结果表明，该方法在复杂背景的多场景中具有良好的性能。