YOLSO: You Only Look Small Object

Small object detection is challenging and far from satisfactory. Most general object detectors suffer from two critical issues with small objects: (1) Feature extractor based on classification network cannot express the characteristics of small objects reasonably due to insufficient appearance information of targets and a large amount of background interference around them. (2) The detector requires a much higher location accuracy for small objects than for general objects. This paper proposes an effective and efficient small object detector YOLSO to address the above problems. For feature representation, we analyze the drawbacks in previous backbones and present a Half-Space Shortcut(HSSC) module to build a background-aware backbone. Furthermore, a coarse-to-fine Feature Pyramid Enhancement(FPE) module is introduced for layer-wise aggregation at a granular level to enhance the semantic discriminability. For loss function, we propose an exponential L1 loss to promote the convergence of regression, and a focal IOU loss to focus on prime samples with high classification confidence and high IOU. Both of them significantly improves the location accuracy of small objects. The proposed YOLSO sets state-of-the-art results on two typical small object datasets, MOCOD and VeDAI, at a speed of over 200 FPS. In the meantime, it also outperforms the baseline YOLOv3 by a wide margin on the common COCO dataset.     

The phosphorescence emission in undoped lead-halide Cs4PbBr6 single crystals at low temperature

Recently, all-inorganic cesium lead-halide perovskites have shown their promise for light emission applications, due to the excellent optical performance. Herein, we report that the initially nonphosphorescent undoped lead-halide Cs₄PbBr₆ single crystals (SCs) exhibit an ultralong phosphorescence emission under X-ray excitation at low temperatures. It is shown that the dramatic change has been taken place in radioluminescence spectra and the broad-band emission gradually appeared with the decrease of temperature. Below 210 K, the radioluminescence spectra can be deconvoluted into one narrow peak located at 530 nm and two broad peaks centered at 595 nm and 672 nm respectively. Subsequently, the time-dependent radioluminescence spectra in undoped lead-halide Cs₄PbBr₆ SCs were investigated. The ultralong phosphorescence emission can persist over 120 min at 70 K. We consider that ultralong phosphorescence originates from defect-related emission. To the best of our knowledge, our finding is the first time that undoped Cs₄PbBr₆ SCs exhibit the phosphorescence emission, which will offer a paradigm to motivate revolutionary applications on perovskite.     

Structural similarity measure between UML class diagrams based on UCG

Requirements Engineering - In software reuse, the reuse of UML class diagram produced in design phase has received more attention due to the important influence on the following developing process....     

基于RFID标签阵列的睡眠期间呼吸量连续监测系统

睡眠期间连续且准确的呼吸量监测有助于推断用户的睡眠阶段以及提供一些慢性疾病的线索。现有工作主要针对呼吸频率进行感知和监测,缺乏对呼吸量进行连续监测的手段。针对上述问题提出了一种基于商用无线射频识别(RFID)标签的无线感知用户睡眠期间呼吸量的系统——RF-SLEEP。RF-SLEEP通过阅读器连续收集附着在胸部表面的标签阵列返回的相位值及时间戳数据,计算出呼吸引起的胸部不同点的位移量,基于广义回归神经网络(GRNN)构建胸部不同点的位移量与呼吸量之间的关系模型,从而实现对用户睡眠期间呼吸量的评估。RF-SLEEP通过在用户肩膀处附着双参考标签,消除用户睡眠期间翻转身体对胸部位移计算造成的误差。实验结果表明,RFSLEEP对不同用户睡眠期间的呼吸量连续监测的平均精确度为92.49%。     

Preventive replacement policies with time of operations,mission durations,minimal repairs and maintenance triggering approaches

When a mission arrives at a random time and lasts for a duration, it becomes an interesting problem to plan replacement policies according to the health condition and repair history of the operating unit, as the reliability is required at mission time and no replacement can be done preventively during the mission duration. From this viewpoint, this paper proposes that effective replacement policies should be collaborative ones gathering data from time of operations, mission durations, minimal repairs and maintenance triggering approaches. We firstly discuss replacement policies with time of operations and random arrival times of mission durations, model the policies and find optimum replacement times and mission durations to minimize the expected replacement cost rates analytically. Secondly, replacement policies with minimal repairs and mission durations are discussed in a similar analytical way. Furthermore, the maintenance triggering approaches, i.e., replacement first and last, are also considered into respective replacement policies. Numerical examples are illustrated when the arrival time of the mission has a gamma distribution and the failure time of the unit has a Weibull distribution. In addition, simple case illustrations of maintaining the production system in glass factories are given based on the assumed data.     

光时域反射计的设计与实现

文章介绍了一种光时域反射计的设计和实现,详细介绍了各个硬件模块和使用FPGA实现数据等效采样和消噪处理的方法.模块配备LAN和RS232接口,方便集成以及分布式部署,利于推广应用,从而带来较好的经济、社会和管理效益.     

基于能量预测的传感器网络信任机制研究

信任机制最近已建议作为一个无线传感器网络(WSNs)有效的安全机制.文中提出了一种信任机制(EPTM),该机制不仅可以防止被入侵的节点或者恶意节点选举为簇头,而且还设计出一种新型副簇头节点来监察簇头以防止他们的恶意行为.特别介绍了一种基于能量预测的方法来检测拒绝服务攻击(DoS)的节点,选出值得信赖的簇.最后通过仿真验证了机制的可行性,结果表明:EPTM可以有效防御拒绝服务(DoS)攻击.     

一种改进的基于在线解耦的轻量级动态污点分析方法

在针对二进制程序的实际分析场景中,动态分析方法因具有更高的准确性而得到更为广泛的关注和运用.但是以动态污点分析为代表的细粒度分析方法通常会产生较高的性能和资源开销.为缓解这些问题,本文提出一种改进的轻量级在线解耦的动态污点分析方法,其主要思想是在线解耦程序执行和指令分析,同步构建分析代码并完成分析.该方法能够进一步降低对目标程序运行时的性能影响,并易将分析过程迁移至其它上下文环境中,从而可进一步减小对目标程序运行环境的干扰.本文通过采用多种实际程序进行实验,结果表明该方法能够比传统在线污点分析具有更高的分析性能,并具有较低的系统资源消耗,易于在实际环境中进行部署和应用.     

含蓄热式电采暖的综合能源系统日前优化调度策略研究

以含蓄热式电采暖的综合能源系统日前优化调度运行问题为研究对象,分析了蓄热式电采暖系统运行的各种特点,建立了含蓄热式电采暖的综合能源系统运行优化调度模型。将某典型蓄热电采暖用户场景作为算例进行验证,结果表明通过对蓄热式电采暖运行情况进行优化,可以有效降低其对电网的负荷冲击,减小电网峰谷差,提升电采暖设备的使用效率,大幅降低用户采暖的运行费用。该方法可为蓄热式电采暖用户参与和电网互动、促进可再生能源消纳、提升采暖系统经济运行水平、降低电网峰谷差以及促进蓄热式电采暖系统推广等提供参考依据。