基于轻量级MobileNet-SSD模型的人流量检测

利用深度神经网络模型识别行人目标并检测具有十分高的价值。现实高密度行人检测场景中由于硬件基础以及网络性能消耗等问题的影响,常常需要选取处理速度高、硬件条件要求低的网络且同时兼顾视频监控的连续特性,因此选取了轻量级MobileNet-SSD网络来高效处理人头目标并引入帧间差分的方式,来有效对人头的椭圆特征目标进行跟踪;结合相关数理方法实现行人跨线计数的高性能人流量检测解决方案。在不同数据集上同现今一流的检测模型作比较,该方法皆表现出优良的检测性能。

Regularization by Multiple Dual Frames for Compressed Sensing Magnetic Resonance Imaging With Convergence Analysis

Plug-and-play priors are popular for solving illposed imaging inverse problems. Recent efforts indicate that the convergence guarantee of the imaging algorithms using plug-andplay priors relies on the assumption of bounded denoisers. However, the bounded properties of existing plugged Gaussian denoisers have not been proven explicitly. To bridge this gap, we detail a novel provable bounded denoiser termed as BMDual,which combines a trainable denoiser using dual tight frames and the well-known block-matching and 3D filtering(BM3D)denoiser. We incorporate multiple dual frames utilized by BMDual into a novel regularization model induced by a solver. The proposed regularization model is utilized for compressed sensing magnetic resonance imaging(CSMRI). We theoretically show the bound of the BMDual denoiser, the bounded gradient of the CSMRI data-fidelity function, and further demonstrate that the proposed CSMRI algorithm converges. Experimental results also demonstrate that the proposed algorithm has a good convergence behavior, and show the effectiveness of the proposed algorithm.

Cascaded ELM-Based Joint Frame Synchronization and Channel Estimation over Rician Fading Channel with Hardware Imperfections

Due to the interdependency of frame synchronization(FS)and channel estimation(CE),joint FS and CE(JFSCE)schemes are proposed to enhance their functionalities and therefore boost the overall performance of wireless communication systems.Although traditional JFSCE schemes alleviate the influence between FS and CE,they show deficiencies in dealing with hardware imperfection(HI)and deterministic line-of-sight(LOS)path.To tackle this challenge,we proposed a cascaded ELM-based JFSCE to alleviate the influence of HI in the scenario of the Rician fading channel.Specifically,the conventional JFSCE method is first employed to extract the initial features,and thus forms the non-Neural Network(NN)solutions for FS and CE,respectively.Then,the ELMbased networks,named FS-NET and CE-NET,are cascaded to capture the NN solutions of FS and CE.Simulation and analysis results show that,compared with the conventional JFSCE methods,the proposed cascaded ELM-based JFSCE significantly reduces the error probability of FS and the normalized mean square error(NMSE)of CE,even against the impacts of parameter variations.

Dynamic interactive bitwise meta-holography with ultra-high computational and display frame rates

Interactive holography offers unmatched levels of immersion and user engagement in the field of future display.Despite of the substantial progress has been made in dynamic meta-holography,the realization of real-time,highly smooth interactive holography remains a significant challenge due to the computational and display frame rate limitations.In this study,we introduced a dynamic interactive bitwise meta-holography with ultra-high computational and display frame rates.To our knowledge,this is the first reported practical dynamic interactive metasurface holographic system.We spa-tially divided the metasurface device into multiple distinct channels,each projecting a reconstructed sub-pattern.The switching states of these channels were mapped to bitwise operations on a set of bit values,which avoids complex holo-gram computations,enabling an ultra-high computational frame rate.Our approach achieves a computational frame rate of 800 kHz and a display frame rate of 23 kHz on a low-power Raspberry Pi computational platform.According to this methodology,we demonstrated an interactive dynamic holographic Tetris game system that allows interactive gameplay,color display,and on-the-fly hologram creation.Our technology presents an inspiration for advanced dynamic meta-holography,which is promising for a broad range of applications including advanced human-computer interaction,real-time 3D visualization,and next-generation virtual and augmented reality systems.

Resilience-incorporated seismic risk assessment of precast concrete frames with“dry”connections

A resilience-incorporated risk assessment framework is proposed and demonstrated in this study to manifest the advantageous seismic resilience of precast concrete frame(PCF)structures with“dry”connections in terms of their low damage and rapid recovery.The framework integrates various uncertainties in the seismic hazard,fragility,capacity,demand,loss functions,and post-earthquake recovery.In this study,the PCF structures are distinguished from ordinary reinforced concrete frame(RCF)structures by characterizing multiple limit states for the PCF based on its unique damage mechanisms.Accordingly,probabilistic story-wise pushover analyses are performed to yield story-wise capacities for the predefined limit states.In the seismic resilience analysis,a step-wise recovery model is proposed to idealize the functionality recovery process,with separate considerations of the repair and non-repair events.The recovery model leverages the economic loss and downtime to delineate the stochastic post-earthquake recovery curves for the resilience loss estimation.As such,contingencies in the probabilistic post-earthquake repairs are incorporated and the empirical judgments on the recovery parameters are largely circumvented.The proposed framework is demonstrated through a comparative study between two“dry”connected PCFs and one RCF designed as alternative structural systems for a prototype building.The results from the risk quantification indicate that the PCFs show reduced loss hazards and lower expected losses relative to the RCF.Particularly,the PCF equipped with energy dissipation devices at the“dry”connections largely reduces the expected economic loss,downtime,and resilience loss by 29%,56%,and 60%,respectively,compared to the RCF.

Experimental and numerical study regarding H-steel all-bolted connection steel frame with composite wall boards

H-steel all-bolted connection steel frame structures with heat preservation and decoration composite wall boards were investigated and the seismic performances of three scaled specimens were studied.The failure modes,hysteresis curves,bearing capacity,ductility,energy dissipation capacity,stiffness degradation and strain distribution were discussed.The calculation method of structural theoretical internal force was presented.The results showed that the overall structural seismic performance was better,and the structural ductility met the demands of elastic-plastic inter-story drift angle for seismic design.The H-steel weak-axis connection structure obtained better energy dissipation capacity,and its bearing capacity and stiffness were slightly different from the strong-axis connection.The heat preservation and decoration performance of composite wallboard and the all-bolted connection of the steel frame realized prefabrication during the whole construction period.The plastic hinge of the steel beam can be moved outwards because of the L-angles,which effectively avoids stress concentration in joint areas and expands the plastic hinge range.The errors between the theoretical structural capacity calculated by the plastic analysis method and the test results were within 2.44%.In addition,structural failure mechanisms and bearing capacities were verified by the finite element(FE)analysis,and the effects of the main parameters on the structures were investigated.The FE verification results were the same as in the test.The research results provide theoretical support and technical guidance for the application of thermal insulation and decorative composite wall panels in H-shaped steel all-bolted steel frames.

Evaluation of the seismic behavior of reinforced concrete structures with flat slab-column gravity frame and shear walls through nonlinear analysis methods

This paper presents an investigation of the seismic behavior of reinforced concrete(RC)structures in which shear walls are the main lateral load-resisting elements and the participation of flat slab floor systems is not considered in the seismic design procedure.In this regard,the behavior of six prototype structures(with different heights and plan layouts)is investigated through nonlinear static and time history analyses,implemented in the OpenSees platform.The results of the analyses are presented in terms of the behavior of the slab-column connections and their mode of failure at different loading stages.Moreover,the global response of the buildings is discussed in terms of some parameters,such as lateral overstrength due to the gravity flat slab-column frames.According to the nonlinear static analyses,in structures in which the slab-column connections were designed only for gravity loads,the slab-column connections exhibited a punching mode of failure even in the early stages of loading.However,the punching failure was eliminated in structures in which a minimum transverse reinforcement recommended in ACI 318(2019)was provided in the slabs at joint regions.Furthermore,despite neglecting the contribution of gravity flat slab-column frames in the lateral load resistance of the structures,a relatively significant overstrength was imposed on the structures by the gravity frames.

Field Load Test Based SHM System Safety Standard Determination for Rigid Frame Bridge

The deteriorated continuous rigid frame bridge is strengthened by external prestressing. Static loading tests wereconducted before and after the bridge rehabilitation to verify the effectiveness of the rehabilitation process. Thestiffness of the repaired bridge is improved, and the maximum deflection of the load test is reduced from 37.9 to27.6 mm. A bridge health monitoring system is installed after the bridge is reinforced. To achieve an easy assessmentof the bridge’s safety status by directly using transferred data, a real-time safety warning system is createdbased on a five-level safety standard. The threshold for each safety level will be determined by theoretical calculationsand the outcomes of static loading tests. The highest risk threshold will be set at the ultimate limit statevalue. The remaining levels, namely middle risk, low risk, and very low risk, will be determined usingreduction coefficients of 0.95, 0.9, and 0.8, respectively.

Frame Length Dependency for Fundamental Frequency Extraction in Noisy Speech

The fundamental frequency plays a significant part in understanding and perceiving the pitch of a sound. The pitch is a fundamental attribute employed in numerous speech-related works. For fundamental frequency extraction, several algorithms have been developed which one to use relies on the signal’s characteristics and the surrounding noise. Thus, the algorithm’s noise resistance becomes more critical than ever for precise fundamental frequency estimation. Nonetheless, numerous state-of-the-art algorithms face struggles in achieving satisfying outcomes when confronted with speech recordings that are noisy with low signal-to-noise ratio (SNR) values. Also, most of the recent techniques utilize different frame lengths for pitch extraction. From this point of view, This research considers different frame lengths on male and female speech signals for fundamental frequency extraction. Also, analyze the frame length dependency on the speech signal analytically to understand which frame length is more suitable and effective for male and female speech signals specifically. For the validation of our idea, we have utilized the conventional autocorrelation function (ACF), and state-of-the-art method BaNa. This study puts out a potent idea that will work better for speech processing applications in noisy speech. From experimental results, the proposed idea represents which frame length is more appropriate for male and female speech signals in noisy environments.

Real-Time Mosaic Method of Aerial Video Based on Two-Stage Key Frame Selection Method

A two-stage automatic key frame selection method is proposed to enhance stitching speed and quality for UAV aerial videos. In the first stage, to reduce redundancy, the overlapping rate of the UAV aerial video sequence within the sampling period is calculated. Lagrange interpolation is used to fit the overlapping rate curve of the sequence. An empirical threshold for the overlapping rate is then applied to filter candidate key frames from the sequence. In the second stage, the principle of minimizing remapping spots is used to dynamically adjust and determine the final key frame close to the candidate key frames. Comparative experiments show that the proposed method significantly improves stitching speed and accuracy by more than 40%.

远程开放教育中移动学习的接受与学习体验探究——基于COI和FRAME综合模式的考察

随着智能手机的普及、移动互联网的发展和云平台的应用,移动学习已成为我国远程开放教育普遍采用的学习形式,移动学习的接受和学习体验成为亟待研究的课题。本研究借鉴西方移动学习理论,通过批判性综合探究社区理论(COI)和移动学习理性分析框架(FRAME),进行问卷调查。结果表明,远程开放教育移动学习的教学设计需要充分考虑设备维度的限制,加强内容碎片化和输入方式多元化设计,增强移动学习交流与互信,提升合作学习探究效果,加强导学课或线上学习素养课的实施。远程开放教育需要面对新的移动社会的变化,持续思考,更新、优化、整合教育系统,以进一步激发学习者在移动学习中发挥、利用、实现自己的潜能。

基于改进CenterNet的轻量级无锚框SAR图像多尺度舰船检测算法

针对复杂场景下多尺度舰船检测精度较差、效率较低、泛化性较弱等问题,本文构建了一个轻量级无锚框的合成孔径雷达图像舰船检测框架。为满足合成孔径雷达图像舰船实时检测的需求,提出了基于改进CenterNet的轻量级无锚框合成孔径雷达图像舰船检测方法,通过预测目标关键点的信息及检测框的相关属性,实现了合成孔径雷达图像舰船快速准确定位与检测。为解决合成孔径雷达图像样本稀缺的问题,采用了适用于合成孔径雷达舰船图像的数据增强方法以扩充训练样本,并引入了多尺度训练以增强模型泛化性能。实验结果表明:本文方法具有检测效率高、检测精度好、泛化性能强等优势,能实现复杂场景下多尺度舰船的实时高精度检测。

基于FrameNet框架关系的文本蕴含识别

文本蕴含识别是处理自然语言中广泛存在的同义异形现象的一种有效途径。该文基于FrameNet中框架及框架之间的八种关系,结合WordNet中词汇间的语义关系,提出了一种文本蕴含识别方法。在给定文本T和假设H中词元激起的框架基础上,该方法利用深度优先搜索,在FrameNet框架关系图中,查询T和H中框架之间的上下位关系;再使用WordNet中语义关系比较二者的框架元素是否一致或相似。实验对RTE2007中50个文本对进行了测试,达到了76.6%的准确率,略高于RTE2007评测的最优结果。

Framenet核心框架元素间关系研究

本文首先阐述了框架元素的类型:核心框架元素、外围框架元素和额外主题框架元素,然后再详细论述了Framenet中3种典型的核心框架元素关系:核心集关系、必要关系和排斥关系,并总结出继承关系的框架间核心框架元素关系的变化规律。

链接请求的运行机制及其对.NET Framework分布式系统安全性的影响

.NET安全体系是采用全新的模式建立的,更多地考虑了基于.NET的应用系统在安全性和执行效率上的统一,它提出了基于Web的分布式应用系统在安全核心技术上新的设计思想。影响.NETFramework应用系统安全性的因素有很多,其中由链接请求引起的影响系统安全性的潜在错误相当多。本文分析了链接请求可能带来的系统安全性隐患,并就此提出了一些建议。

浅谈FrameNet框架汉译时的译者术语意识

汉语框架网的建设需要参照FrameNet,因而FrameNet框架汉译是其中必要的一步。由于FrameNet框架会涉及众多"概念名称",所以在FrameNet框架汉译时,译者术语意识通过提供一定程度上的理论指导来加强控制汉语框架网的框架质量。

基于改进U-Net的火龙果采摘图像分割和姿态估计方法

为了实现火龙果采收自动化作业,提出一种基于改进U-Net的火龙果图像分割和姿态估计方法。首先,在U-Net模型的跳跃连接(编码器与解码器部分特征图进行的连接操作)中引入通道和空间注意力机制模块(Concurrent spatial and channel squeeze and channel excitation,SCSE),同时将SCSE模块集成到残差模块(Double residual block,DRB)中,在增强网络提取有效特征能力的同时提高网络的收敛速度,得到一种基于注意力残差U-Net的火龙果图像分割网络。通过该网络分割出果实及其附生枝条的掩膜图像,利用图像处理技术和相机成像模型拟合出果实及其附生枝条的轮廓、果实质心、果实最小外接矩形框和三维边界框,进而结合果实及其附生枝条的位置关系进行火龙果三维姿态估计,并在火龙果种植园中获得一个测试集,以评价该算法的性能,最后在自然果园环境下进行实地采摘试验。试验结果表明,火龙果果实图像分割平均交并比(mIoU)和平均像素准确率(mPA)分别达到86.69%和93.89%,三维姿态估计平均误差为8.8°,火龙果采摘机器人在果园环境下的采摘成功率为86.7%,平均采摘时间为22.3 s。满足火龙果机械化作业要求。

Effects of material of metallic frame on the penetration resistances of ceramic-metal hybrid structures

The effects of metallic material on the penetration resistances of ceramic-metal hybrid structures against vertical long-rod tungsten projectiles were studied by artillery-launched experiments and numerical simulation.Hybrid structures with rectangular cores in transverse orthogonal arrangement and slidefitting ceramic inserts of zirconia toughened alumina prisms were fabricated with titanium alloy TC4(Ti6 Al4 V),AISI 4340 steel and 7075 aluminum alloy panels,respectively.The results showed that the hybrid structure of Ti6A14V exhibited the highest penetration resistance,followed by that of 7075 aluminum alloy with the same area density.The penetration resistance of the hybrid structure of AISI4340 steel was the lowest.The underlying mechanisms showed that the metallic material of a ceramicmetal hybrid structure can directly affect its energy absorption from the impact projectile,which further affects its penetration resistance.Different metallic frames exhibited different failure characteristics,resulting in different constraint conditions or support conditions for ceramic prisms.The high penetration resistance of the Ti6Al4V hybrid structure was due to its stronger back support to ceramic prisms as compared with that of AISI 4340 steel hybrid structure,and better constraint condition for ceramic prisms by metallic webs as compared with that of 7075 aluminum alloy hybrid structure.The results of mass efficiency and thickness efficiency showed that the Ti6Al4V hybrid structure has advantages in reducing both the thickness and the mass of protective structure.In addition,because the ceramic-metal hybrid structures in the present work were heterogeneous,impact position has slight influence on their penetration resistances.

基于改进RetinaNet的马拉松号码簿检测方法研究

针对自然场景中马拉松运动员号码簿的扭曲形变与多目标混合等复杂情况,提出一种基于改进RetinaNet的号码簿检测网络模型,以提高号码簿检测精度。该模型以RetinaNet为基础网络结构,在原有候选框生成策略中引入角度参数,通过生成不同角度、大小的候选框来匹配水平及倾斜目标。模型中针对旋转检测框在角度回归中引起的边界问题,采用环形平滑标签方法将角度回归问题转化为分类问题,通过对窗口函数及窗口半径的设置,使得目标区域定位更精确。实验结果表明,在收集到的真实马拉松数据集上,相比原始RetinaNet,改进后的模型在简单及复杂背景下的检测精度均有所提升。在相同数据标注的条件下,与具有多角度的文本检测模型CTPN、EAST模型相比,改进后模型对于倾斜号码簿的检测性能更优。

一种基于模板匹配的芯片Frame图像分割算法

针对背景复杂、边界模糊以及芯片相连等特征的芯片Frame图像,提出了一种基于模板匹配的芯片Frame图像分割算法。首先,对整幅芯片Frame图像预分割出多个区域模块;然后,基于区域模块图像采取模板匹配算法匹配单芯片图像;最后,通过合并单芯片的重叠匹配框并记录合并框的坐标信息的方式分割出单芯片图像。实验结果表明:选取合适的模板和阈值,能使该算法的分割准确率达到100%,并且比不基于区域模块匹配的分割算法节省了至少45.76%的分割时间,满足芯片Frame高精度和高速度的分割需求,为芯片图像分割算法的研究提供了一种新思路。