Lightweight Res-Connection Multi-Branch Network for Highly Accurate Crowd Counting and Localization

Crowd counting is a promising hotspot of computer vision involving crowd intelligence analysis,achieving tremendous success recently with the development of deep learning.However,there have been stillmany challenges including crowd multi-scale variations and high network complexity,etc.To tackle these issues,a lightweight Resconnection multi-branch network(LRMBNet)for highly accurate crowd counting and localization is proposed.Specifically,using improved ShuffleNet V2 as the backbone,a lightweight shallow extractor has been designed by employing the channel compression mechanism to reduce enormously the number of network parameters.A light multi-branch structure with different expansion rate convolutions is demonstrated to extract multi-scale features and enlarged receptive fields,where the information transmission and fusion of diverse scale features is enhanced via residual concatenation.In addition,a compound loss function is introduced for training themethod to improve global context information correlation.The proposed method is evaluated on the SHHA,SHHB,UCF-QNRF and UCF_CC_50 public datasets.The accuracy is better than those of many advanced approaches,while the number of parameters is smaller.The experimental results show that the proposed method achieves a good tradeoff between the complexity and accuracy of crowd counting,indicating a lightweight and high-precision method for crowd counting.

Deep Learning Based Efficient Crowd Counting System

Estimation of crowd count is becoming crucial nowadays,as it can help in security surveillance,crowd monitoring,and management for different events.It is challenging to determine the approximate crowd size from an image of the crowd’s density.Therefore in this research study,we proposed a multi-headed convolutional neural network architecture-based model for crowd counting,where we divided our proposed model into two main components:(i)the convolutional neural network,which extracts the feature across the whole image that is given to it as an input,and(ii)the multi-headed layers,which make it easier to evaluate density maps to estimate the number of people in the input image and determine their number in the crowd.We employed the available public benchmark crowd-counting datasets UCF CC 50 and ShanghaiTech parts A and B for model training and testing to validate the model’s performance.To analyze the results,we used two metrics Mean Absolute Error(MAE)and Mean Square Error(MSE),and compared the results of the proposed systems with the state-of-art models of crowd counting.The results show the superiority of the proposed system.

A multi-source information fusion layer counting method for penetration fuze based on TCN-LSTM

When employing penetration ammunition to strike multi-story buildings,the detection methods using acceleration sensors suffer from signal aliasing,while magnetic detection methods are susceptible to interference from ferromagnetic materials,thereby posing challenges in accurately determining the number of layers.To address this issue,this research proposes a layer counting method for penetration fuze that incorporates multi-source information fusion,utilizing both the temporal convolutional network(TCN)and the long short-term memory(LSTM)recurrent network.By leveraging the strengths of these two network structures,the method extracts temporal and high-dimensional features from the multi-source physical field during the penetration process,establishing a relationship between the multi-source physical field and the distance between the fuze and the target plate.A simulation model is developed to simulate the overload and magnetic field of a projectile penetrating multiple layers of target plates,capturing the multi-source physical field signals and their patterns during the penetration process.The analysis reveals that the proposed multi-source fusion layer counting method reduces errors by 60% and 50% compared to single overload layer counting and single magnetic anomaly signal layer counting,respectively.The model's predictive performance is evaluated under various operating conditions,including different ratios of added noise to random sample positions,penetration speeds,and spacing between target plates.The maximum errors in fuze penetration time predicted by the three modes are 0.08 ms,0.12 ms,and 0.16 ms,respectively,confirming the robustness of the proposed model.Moreover,the model's predictions indicate that the fitting degree for large interlayer spacings is superior to that for small interlayer spacings due to the influence of stress waves.

基于Counting Bloom Filter的海量网页快速去重研究

网页去重是从给定的大量的数据集合中检测出冗余的网页,然后将冗余的网页从该数据集合中去除的过程,其中基于同源网页的URL去重的研究已经取得了很大的发展,但是针对海量网页去重问题,目前还没有很好的解决方案,文章在基于MD5指纹库网页去重算法的基础上,结合Counting Bloom Filter算法的特性,提出了一种快速去重算法IMP-CBFilter。该算法通过减少I/O频繁操作,来提高海量网页去重的效率。实验表明,IMP-CBFilter算法的有效性。

分形曲线Box-counting维数的一种逼近算法

讨论Ｌ２范数意义下线性分形插值函数（ＬＦＩＦ）对给定函数图象的最佳逼近问题。利用［４］的结果得出了求最佳逼近点的方法，利用最佳产、给出了求给定函数国象Ｂｏｘ－ｃｏｕｎｔｉｎｇ维数近似值的方法。

Automated Counting of Rice Planthoppers in Paddy Fields Based on Image Processing

A quantitative survey of rice planthoppers in paddy fields is important to assess the population density and make forecasting decisions. Manual rice planthopper survey methods in paddy fields are time-consuming, fatiguing and tedious. This paper describes a handheld device for easily capturing planthopper images on rice stems and an automatic method for counting rice planthoppers based on image processing. The handheld device consists of a digital camera with WiFi, a smartphone and an extrendable pole. The surveyor can use the smartphone to control the camera, which is fixed on the front of the pole by WiFi, and to photograph planthoppers on rice stems. For the counting of planthoppers on rice stems, we adopt three layers of detection that involve the following：（a） the first layer of detection is an AdaBoost classifier based on Haar features;（b） the second layer of detection is a support vector machine（SVM） classifier based on histogram of oriented gradient（HOG） features;（c） the third layer of detection is the threshold judgment of the three features. We use this method to detect and count whiteback planthoppers（Sogatella furcifera） on rice plant images and achieve an 85.2% detection rate and a 9.6% false detection rate. The method is easy, rapid and accurate for the assessment of the population density of rice planthoppers in paddy fields.

基于Counting Bloom Filter的海量网页快速去重研究

网页去重是从给定的大量的数据集合中检测出冗余的网页,然后将冗余的网页从该数据集合中去除的过程,可以有效地减少检索和存储的压力。其中基于同源网页的URL去重方法、基于网页结构和特征的抽取指纹方法和基于网页内容的聚类方法的研究都已经取得了很大的发展,但是针对海量网页去重问题,上述三种方法,目前还是很难解决网页去重的时间和空间问题,本文在基于MD5指纹库网页去重算法的基础上,结合Counting Bloom filter算法的特性,提出一个节省空间的大规模数据表示和快速去重策略,实现了一种快速去重算法IMP-CM Filter,大大降低了网页去重算法的时间复杂度和空间复杂度。该算法通过减少I/0频繁操作,来提高海量网页去重的效率。最后通过实验表明,IMP-CM Filter算法的有效性。

Recent progress of SiC UV single photon counting avalanche photodiodes

4H-SiC single photon counting avalanche photodiodes(SPADs)are prior devices for weak ultraviolet(UV)signal detection with the advantages of small size,low leakage current,high avalanche multiplication gain,and high quantum efficiency,which benefit from the large bandgap energy,high carrier drift velocity and excellent physical stability of 4 H-SiC semiconductor material.UV detectors are widely used in many key applications,such as missile plume detection,corona discharge,UV astronomy,and biological and chemical agent detection.In this paper,we will describe basic concepts and review recent results on device design,process development,and basic characterizations of 4 H-SiC avalanche photodiodes.Several promising device structures and uniformity of avalanche multiplication are discussed,which are important for achieving high performance of 4 HSiC UV SPADs.

Preparation of the Ag_2O_2-PbO_2 Modified Electrode and Its Application towards Escherichia coli Fast Counting in Water

A novel nano crystalline Ag2O2-PbO2 film chemically modified electrode (CME) was prepared and the CME was characterized by X-ray diffractometer (XRD) and atomic force microscope (AFM). By chronoamperometry, the nano Ag2O2-PbO2 CME was used as bioelectro- chemical sensor to determine the population of Escherichia coli (E. coli) in water. Compared with conventional methods, it is found that the technique we used is fast and convenient in counting E. coli.

基于Counting Bloom Filter的DNS异常检测

鉴于失败的DNS查询(failed DNS query)能提供恶意网络活动的证据,以DNS查询失败的数据为切入口,提出一种轻量级的基于Counting Bloom Filter的DNS异常检测方法。该方法使用带语义特征的可逆哈希函数对被查询的域名及发起查询的IP进行快速的聚类和还原。实验结果证明该方法能以较少的空间占用和较快的计算速度有效识别出DNS流量中的异常,适用于僵尸网络、分布式拒绝服务(DDoS)攻击等异常检测的前期筛选和后期验证。

Evolution of the Late Cretaceous Yongfeng-Chongren Basin in Jiangxi Province,southeast China: insights from sedimentary facies analysis and pebble counting

The Gan-Hang Belt in Southeast China is characterized by several igneous and siliciclastic basins associated with crustal extension during Late Mesozoic. The sedimentary evolution of the red basins is still poorly understood. In this study, sedimentary fades analysis and pebble counting were performed on outcrop sections of the Late Cretaceous Guifeng Group in the Yongfeng-Chongren Basin in central Jiangxi Province. Thirty-five conglomerate outcrops were chosen to measure pebble lithology, size, roundness, weathering degree and preferred orientation. Results show that gravels are mostly fine to coarse pebbles and comprise dominantly quartzites, metamorphic rocks, granitoids and sandstones. Rose diagrams based on imbricated pebbles indicate variable paleocurrent directions. Combining with typical sedimentary structures and vertical successions, we suggest that the Guifeng Group were deposited in alluvial fan, river and playa lake depositional systems. The proposed depositional model indicates that the Hekou Formation represents the start-up stage of the faulted basin, accompanied by sedimentation in alluvial fan and braided river environments. Then this basin turned into a stable expansion stage during the deposition of the Tangbian Formation. Except for minor coarse sediments at the basin margin, the other area is covered with fine-grained sediments of lake and river environments. The Lianhe Formation, however, is once again featured by conglomerates, suggesting a probable tectonic event. Therefore, the study region possibly suffered two tectonic events represented by the conglomerates of the Hekou and Lianhe formations in the context of the crustal extension in Southeast China.

Photon-counting chirped amplitude modulation lidar system using superconducting nanowire single-photon detector at 1550-nm wavelength

We demonstrate a photon-counting chirped amplitude modulation （CAM） light detection and ranging （lidar） system incorporating a superconducting nanowire single-photon detector （SNSPD） and operated at a wavelength of 1550 nm. The distance accuracy of the lidar system was determined by the CAM bandwidth and signal-to-noise ratio （SNR） of an intermediate frequency （IF） signal. Owing to a short dead time （10 ns） and negligible dark count rate （70 Hz） of the SNSPD, the obtained IF signal attained an SNR of 42 dB and the direct distance accuracy was improved to 3 mm when the modulation bandwidth of the CAM signal was 240 MHz and the modulation period was 1 ms.

基于Counting Bloom Filter的流抽样算法研究

为适应高速网络环境并实现对网络流量的准确测量,提出一种将计数型布隆过滤器结构与基于报文的流抽样技术相结合的网络流等概率抽样算法。利用4 bit的Counter向量识别是否有新流出现,通过实时调整抽样频率弥补新流判定中的错误率,从而对网络流进行等概率抽样并获取较真实的网络流分布情况。实验结果表明,该算法的测量结果与网络流真实值较接近,且具有可扩展性,可以满足当前复杂多变的高速网络环境下的流量测量需求。

Filtering Method of Aero⁃engine Load Spectrum Based on Rain Flow Counting

A filtering method of aero-engine load spectrum based on the rain flow counting is proposed in this paper.Firstly,the original load spectrum is counted through the rain flow method to get the peak and valley values.Then,some data points in the original load spectrum are added between the peak and valley values.Finally,the filtering spectrum is obtained.The proposed method can reflect the path information of the original load spectrum.In addition,it can also eliminate the noise in the signal and improve the efficiency of signal processing,which is of practical significance for the research of aero-engine.

A Deep-CNN Crowd Counting Model for Enforcing Social Distancing during COVID19 Pandemic: Application to Saudi Arabia’s Public Places

With the emergence of the COVID19 virus in late 2019 and the declaration that the virus is a worldwide pandemic,health organizations and governments have begun to implement severe health precautions to reduce the spread of the virus and preserve human lives.The enforcement of social distancing at work environments and public areas is one of these obligatory precautions.Crowd management is one of the effective measures for social distancing.By reducing the social contacts of individuals,the spread of the disease will be immensely reduced.In this paper,a model for crowd counting in public places of high and low densities is proposed.The model works under various scene conditions and with no prior knowledge.A Deep CNN model(DCNN)is built based on convolutional neural network(CNN)structure with small kernel size and two fronts.To increase the efficiency of the model,a convolutional neural network(CNN)as the front-end and a multi-column layer with Dilated Convolution as the back-end were chosen.Also,the proposed method accepts images of arbitrary sizes/scales as inputs from different cameras.To evaluate the proposed model,a dataset was created from images of Saudi people with traditional and non-traditional Saudi outfits.The model was also trained and tested on some existing datasets.Compared to current counting methods,the results show that the proposed model has significantly improved efficiency and reduced the error rate.We achieve the lowest MAE by 67%,32%.and 15.63%and lowest MSE by around 47%,15%and 8.1%than M-CNN,Cascaded-MTL,and CSRNet respectively.

Accelerating Packet Classification with Counting Bloom Filters for Virtual Open Flow Switching

The growing trend of network virtualization results in a widespread adoption of virtual switches in virtualized environments. However, virtual switching is confronted with great performance challenges regarding packet classification especially in Open Flow-based software defined networks. This paper first takes an insight into packet classification in virtual Open Flow switching, and points out that its performance bottleneck is dominated by flow table traversals of multiple failed mask probing for each arrived packet. Then we are motivated to propose an efficient packet classification algorithm based on counting bloom filters. In particular, counting bloom filters are applied to predict the failures of flow table lookups with great possibilities, and bypass flow table traversals for failed mask probing. Finally, our proposed packet classification algorithm is evaluated with real network traffic traces by experiments. The experimental results indicate that our proposed algorithm outperforms the classical one in Open v Switch in terms of average search length, and contributes to promote virtual Open Flow switching performance.

Modeling the photon counting and photoelectron counting characteristics of quanta image sensors

A signal chain model of single-bit and multi-bit quanta image sensors(QISs)is established.Based on the proposed model,the photoresponse characteristics and signal error rates of QISs are investigated,and the effects of bit depth,quantum efficiency,dark current,and read noise on them are analyzed.When the signal error rates towards photons and photoelectrons counting are lower than 0.01,the high accuracy photon and photoelectron counting exposure ranges are determined.Furthermore,an optimization method of integration time to ensure that the QIS works in these high accuracy exposure ranges is presented.The trade-offs between pixel area,the mean value of incident photons,and integration time under different illuminance level are analyzed.For the 3-bit QIS with 0.16 e-/s dark current and 0.21 e-r.m.s.read noise,when the illuminance level and pixel area are 1 lux and 1.21μm^(2),or 10000 lux and 0.21μm^(2),the recommended integration time is 8.8 to 30 ms,or 10 to21.3μs,respectively.The proposed method can guide the design and operation of single-bit and multi-bit QISs.

The Effects of Fetal Movement Counting on Mental Health of Mother in Third Trimester A Randomized Controlled Trial

Anxiety is one of the psychological problems in pregnant women that sometimes takes the form of pathological and affects the mental health of mother. The aim of this study was to determine the effects of fetal movement counting on mental health of mother. In a randomized-controlled trial, 208 nulliparous women were randomly divided into two groups. At 28th weeks, both groups completed the GHQ-28. Then the intervention group started to count fetal movements from 28th to 37th weeks of gestation and the control group received routine prenatal care. Again, both groups completed the questionnaire at 37 weeks＇ gestation and the results were compared. Analysis was performed by SPSS and a P value 〈 0.05 was considered significant. The mean scores of mental health of mothers in 28th and 37th of pregnancy was respectively 23.52 ± 10.23 and 21.09 ± 10.12 in the intervention group and the difference was significant （P = 0.025）. The mean in the control group was 23.69 ± 9.43 and 23.88± 8.60 respectively, and the difference was not significant （P = 0.52）. In comparing the mean scores between the two groups, it was found that the difference was not significant at 28th weeks of gestation （P = 0.37）, but it was significant in 37th week （P = 0.002） and the counting of fetal movements could improve the mental health of mothers compared to control group. The women who had fetal movements counting at weeks 28 to 37 Of gestation reported better mental health than the control group. The mother renorted concerns about decreased fetal movement was similar in the two grouns.

A Statistical Analysis of Intensities Estimation on the Modeling of Non-Life Insurance Claim Counting Process

This study presents an estimation approach to non-life insurance claim counts relating to a specified time. The objective of this study is to estimate the parameters in non-life insurance claim counting process, including the homogeneous Poisson process (HPP) and the non-homogeneous Poisson process (NHPP) with a bell-shaped intensity. We use the estimating function, the zero mean martingale (ZMM) as a procedure of parameter estimation in the insurance claim counting process. Then, Λ(t) , the compensator of is proposed for the number of claims in the time interval . We present situations through a simulation study of both processes on the time interval . Some examples of the situations in the simulation study are depicted by a sample path relating to its compensator Λ(t). In addition, an example of the claim counting process illustrates the result of the compensator estimate misspecification.

A counting method for complex overlapping erythrocytes-based microscopic imaging

Red blood cell(RBC)counting is a standard medical test that can help diagnose various conditions and diseases.Manual counting of blood cells is highly tedious and time consuming.However,new methods for counting blood cells are customary employing both electronic and computer-assisted techniques.Image segmentation is a classical task in most image processing applications which can be used to count blood cells in a microscopic image.In this research work,an approach for erythrocytes counting is proposed.We employed a classification before counting and a new segmentation idea was implemented on the complex overlapping clusters in a microscopic smear image.Experimental results show that the proposed method is of higher counting accuracy and it performs much better than most counting algorithms existed in the situation of three or more RBCs overlapping complexly into a group.The average total erythrocytes counting accuracy of the proposed method reaches 92.9%.