Modelling and predicting the effect of social distancing and travel restrictions on COVID-19 spreading

To date, the only effective means to respond to the spreading of the COVID-19 pandemic are non-pharmaceutical interventions (NPIs), which entail policies to reduce social activity and mobility restrictions. Quantifying their effect is difficult, but it is key to reducing their social and economic consequences. Here, we introduce a meta-population model based on temporal networks, calibrated on the COVID-19 outbreak data in Italy and applied to evaluate the outcomes of these two types of NPIs. Our approach combines the advantages of granular spatial modelling of meta-population models with the ability to realistically describe social contacts via activity-driven networks. We focus on disentangling the impact of these two different types of NPIs: those aiming at reducing individuals’ social activity, for instance through lockdowns, and those that enforce mobility restrictions. We provide a valuable framework to assess the effectiveness of different NPIs, varying with respect to their timing and severity. Results suggest that the effects of mobility restrictions largely depend on the possibility of implementing timely NPIs in the early phases of the outbreak, whereas activity reduction policies should be prioritized afterwards.     

Engineering New Microvascular Networks On-Chip: Ingredients,Assembly, and Best Practices

Tissue engineered grafts show great potential as regenerative implants for diseased or injured tissues within the human body. However, these grafts suffer from poor nutrient perfusion and waste transport, thus decreasing their viability post-transplantation. Graft vascularization is therefore a major area of focus within tissue engineering because biologically relevant conduits for nutrient and oxygen perfusion can improve viability post-implantation. Many researchers used microphysiological systems as testing platforms for potential grafts owing to an ability to integrate vascular networks as well as biological characteristics such as fluid perfusion, 3D architecture, compartmentalization of tissue-specific materials, and biophysical and biochemical cues. Although many methods of vascularizing these systems exist, microvascular self-assembly has great potential for bench-to-clinic translation as it relies on naturally occurring physiological events. In this review, the past decade of literature is highlighted, and the most important and tunable components yielding a self-assembled vascular network on chip are critically discussed: endothelial cell source, tissue-specific supporting cells, biomaterial scaffolds, biochemical cues, and biophysical forces. This paper discusses the bioengineered systems of angiogenesis, vasculogenesis, and lymphangiogenesis and includes a brief overview of multicellular systems. It concludes with future avenues of research to guide the next generation of vascularized microfluidic models.     

Small molecule ternary solar cell with two synergistic electron acceptors for enhanced photovoltaic performance

In recent years, tremendous progresses have been achieved for solution processed organic solar cells (OSCs). The strategy of adding a third component to fabricate ternary solar cells has emerged as an effective method to enhance the power conversion efficiency (PCE) of devices. Furthermore, small molecules feature as lower viscosity and excellent repeatability which facilitate the effective morphology control during fabrication process for enhanced photovoltaic performance. Herein, we report a series of ternary solar cells based on a liquid crystal molecule BTR and two electron acceptors of PC₇₁BM and Y6. These molecules show complementary absorption to broaden spectra coverage and form energy levels cascade for efficient charge transfer. Meanwhile, thanks to the improved molecular packing and formed efficient charge transport network in the ternary blend film, the optimal ternary device possesses the improved charge dynamics and suppressed charge recombination. Thus, ternary solar cells deliver the highest PCE of 11.82% with simultaneously enhanced parameters of J_SC, V_OC and FF. This finding further illustrates the important roles of synergistic effect of fullerenes and non-fullerene acceptors in fabricating highly efficient ternary solar cells.     

Discriminative graph convolution networks for hyperspectral image classification

Recently, the proposal of graph convolutional networks (GCN) has successfully implemented into hyperspectral image data representation and analysis. In spite of the great success, there are still several major challenges in hyperspectral image classification, including within-class diversity, and between-class similarity, which generally degenerate hyperspectral image classification performance. To address the problems, we propose a discriminative graph convolution networks (DGCN) for hyperspectral image classification. This method introduces the concepts of within-class scatter and between-class scatter, which respectively reflect the global geometric structure and discriminative information of the input space. The experimental results on the hyperspectral data sets show that the proposed method has good classification performance.     

基于深度学习的视频跟踪研究进展综述

近年来深度学习迅猛发展，颠覆了语音识别、图像分类、文本理解等领域的算法设计思路。深度学习因其具备强大的特征提取能力，在图像识别领域的成绩尤为突出。然而深度学习与视频监控领域的结合并不多，由于深度模型具有多层网络结构，算法复杂度大，训练和更新模型时比较耗时，很难满足实时性要求。回顾了深度学习的发展史，介绍了最近10年来国内外深度学习主要模型，论述了基于深度学习的目标跟踪算法，指出了各算法的优缺点，最后对当前该领域存在的问题和发展前景进行了总结和展望。     

结合细粒度特征与深度卷积网络的手绘图检索

目的 传统的手绘图像检索方法主要集中在检索相同类别的图像，忽略了手绘图像的细粒度特征。对此，提出了一种新的结合细粒度特征与深度卷积网络的手绘图像检索方法，既注重通过深度跨域实现整体匹配，也实现细粒度细节匹配。方法 首先构建多通道混合卷积神经网络，对手绘图像和自然图像分别进行不同的处理；其次通过在网络中加入注意力模型来获取细粒度特征；最后将粗细特征融合，进行相似性度量，得到检索结果。结果 在不同的数据库上进行实验，与传统的尺度不变特征（SIFT）、方向梯度直方图（HOG）和深度手绘模型Deep SaN（sketch-a-net）、Deep 3DS（sketch）、Deep TSN（triplet sketch net）等5种基准方法进行比较，选取了Top-1和Top-10，在鞋子数据集上，本文方法Top-1正确率提升了12%，在椅子数据集上，本文方法Top-1正确率提升了11%，Top-10提升了3%，与传统的手绘检索方法相比，本文方法得到了更高的准确率。在实验中，本文方法通过手绘图像能在第1幅检索出绝大多数的目标图像，达到了实例级别手绘检索的目的。结论 提出了一种新的手绘图像检索方法，为手绘图像和自然图像的跨域检索提供了一种新思路，进行实例级别的手绘检索，与原有的方法相比，检索精度得到明显提升，证明了本文方法的可行性。     

基于深度残差学习的自动驾驶道路场景理解

随着道路场景理解技术的快速发展，自主驾驶领域取得了长足的进步。在相关任务中，包括道路分割、分类和车辆检测的实时性和准确性是安全性的一个关键问题。为此，提出了一个具有编/解码器网络结构的基于深度残差学习的方法。一方面，编码器网络结构使用不同层次的残差网络来提取高维中的抽象特征，这些特征在接下来的三个任务中共享使用；另一方面，解码器网络结构采用一种子任务的并行计算机制，即道路分割、车辆检测和道路分类任务同时执行。此外，全卷积神经网络用于对提取的图像特征进行上采样以解决道路分割问题。最终，实验结果表明在保证高精度的前提下处理帧率可达到15 fps以上。     

Distributed learning consensus control based on neural networks for heterogeneous nonlinear multiagent systems

This paper considers a novel distributed iterative learning consensus control algorithm based on neural networks for the control of heterogeneous nonlinear multiagent systems. The system's unknown nonlinear function is approximated by suitable neural networks; the approximation error is countered by a robust term in the control. Two types of control algorithms, both of which utilize distributed learning laws, are provided to achieve consensus. In the provided control algorithms, the desired reference is considered to be an unknown factor and then estimated using the associated learning laws. The consensus convergence is proven by the composite energy function method. A numerical simulation is ultimately presented to demonstrate the efficacy of the proposed control schemes.     

Improved non-maximum suppression for object detection using harmony search algorithm

Non-maximum suppression (NMS) plays a key role in many modern object detectors. It is responsible to remove detection boxes that cover the same object. NMS greedily selects the detection box with maximum score; other detection boxes are suppressed when the degree of overlap between these detection boxes and the selected box exceeds a predefined threshold. Such a strategy easily retain some false positives, and it limits the ability of NMS to perceive nearby objects in cluttered scenes. This paper proposes an effective method combining harmony search algorithm and NMS to alleviate this problem. This method regards the task of NMS as a combination optimization problem. It seeks final detection boxes under the guidance of an objective function. NMS is applied to each harmony to remove imprecise detection boxes, and the remaining boxes are used to calculate the fitness value. The remaining detection boxes in a harmony with highest fitness value are chosen as the final detection results. The standard Pattern Analysis, Statistical Modeling and Computational Learning Visual Object Classes dataset and the Microsoft Common Objects in Context dataset are used in all of the experiments. The proposed method is applied to two popular detection networks, namely Faster Region-based Convolutional Neural Networks and Region-based Fully Convolutional Networks. The experimental results show that the proposed method improves the average precision of these two detection networks. Moreover, the location performance and average recall of these two detectors are also improved.