Crowd Characteristics Research Articles

Research on Improved Crowd Detection Based on YOLOv5

With the acceleration of the process of modern urbanization and the improvement of residents' material living standards, the flow of people in the public space is gradually becoming saturated. The monitoring equipment in public places records a huge amount of people flow information all the time, but due to the crowds tend to be dense and crowded. Traditional machine learning cannot make accurate and efficient identification of a large number of dense crowds, if the deep learning technology can be used to process the crowded crowd captured by the surveillance camera and accurately identify the number of people in public places, it provides an important guarantee for the flow of people in public areas and safety construction. However, for crowded targets with occlusions, the traditional target detection algorithm sometimes performs poorly. Based on the above background, this paper introduces an enhanced deep learning framework utilizing the YOLOv5 neural network for crowd detection research. aiming at the characteristics of dense and crowded crowds in public areas. By improving convolutional layer C3 in the backbone structure of YOLOv5 neural network and adding CBAM attention mechanism. Compared with the original YOLOv5, the improved model has increased the maximum F1 value of crowd recognition at near, middle and far distances. To sum up, the deep learning framework improved by YOLOv5 neural network proposed in this paper has significantly improved the recognition accuracy of crowded people in public areas.

The effect of building bottlenecks on crowd dynamics involving individuals with simulated disabilities

With the development of urbanization and the growth of population, there is a growing demand for safety in public building facilities. As one of the essential building components of urban architecture, bottlenecks have a significant impact on the evacuation efficiency of crowds. Furthermore, the heterogeneity of crowds also contributes to the complexity of crowd movement through bottlenecks, while aggravating the magnitude of congestion induced by bottlenecks. The objective of this paper is to explore the movement characteristics of heterogeneous crowds passing through a corridor with a bottleneck by conducting a controlled experiment. There were three variables in this experiment, namely the individual categories (i.e., able-bodied individuals, simulated individuals on crutches and simulated wheelchair users), bottleneck width (i.e., 1.2, 1.6 and 2.0 m) and proportion of simulated disabilities in crowds (i.e., 0 %, 5 % and 10 %). Then offset angle, passing efficiency, fundamental diagram, etc., were analyzed. In trials involving simulated individuals on crutches, a higher detouring degree is observed compared to trials involving simulated wheelchair users or mixed groups of two types of simulated disabilities. There is an increase in flow rate induced by increasing the bottleneck width and decreasing the proportion of simulated disabilities. The passing efficiency at the upstream of the bottleneck in all tests is primarily influenced by the bottleneck width, while by the type and proportion of simulated disabilities at the downstream or inside the bottleneck. The findings are intended to complement the dynamic theory of heterogeneous crowds at building bottlenecks, while providing a reference for congestion control of crowds at bottlenecks.

Hierarchical Inverse Distance Transformer for Enhanced Localization in Dense Crowds

Achieving precise individual localization within densely crowded scenes poses a significant challenge due to the intricate interplay of occlusions and varying density patterns. Traditional methods for crowd localization often rely on convolutional neural networks (CNNs) to generate density maps. However, these approaches are prone to inaccuracies stemming from the extensive overlaps inherent in dense populations. To overcome this challenge, our study introduces the Hierarchical Inverse Distance Transformer (HIDT), a novel framework that harnesses the multi-scale global receptive fields of Pyramid Vision Transformers. By adapting to the multi-scale characteristics of crowds, HIDT significantly enhances the accuracy of individual localization. Incorporating Focal Inverse Distance techniques, HIDT adeptly addresses issues related to scale variation and dense overlaps, prioritizing local small-scale features within the broader contextual understanding of the scene. Rigorous evaluation on standardized benchmarks has unequivocally validated the superiority of our approach. HIDT exhibits outstanding performance across various datasets. Notably, on the JHU-Crowd++ dataset, our method demonstrates significant improvements over the baseline, with MAE and MSE metrics decreasing from 66.6 and 253.6 to 59.1 and 243.5, respectively. Similarly, on the UCF-QNRF dataset, performance metrics increase from 89.0 and 153.5 to 83.6 and 138.7, highlighting the effectiveness and versatility of our approach.

A Dynamic Prediction Framework for Urban Public Space Vitality: From Hypothesis to Algorithm and Verification

Predicting and assessing the vitality of public urban spaces is crucial for effective urban design, aiming to prevent issues such as “ghost streets” and minimize resource wastage. However, existing assessment methods often lack temporal dynamics or heavily rely on historical big data, limiting their ability to accurately predict outcomes for unbuilt projects. To address these challenges, this study integrates previous methodologies with observations of crowd characteristics in public spaces. It introduces the crowd-frequency hypothesis and develops an algorithm to establish a time-dimensional urban vitality dynamic prediction model. Through a case study of the Rundle Mall neighborhood in Adelaide, Australia, the effectiveness of the prediction model was validated using on-site observation sampling and comparative verification. The prediction model framework allows for the determination of urban vitality within specific time ranges by directly inputting basic information, providing valuable support to urban planners and government officials during the design and decision-making processes. It offers a cost-effective approach to achieve sustainable urban vitality construction. Furthermore, machine learning techniques, specifically the decision tree model, were applied to case data to develop a set of preliminary algorithm tools, which enable output of reference urban vitality levels (high-medium-low).

On the ideal gas law for crowds with high pressure

Active particle systems, such as human crowds, are out of equilibrium posing a significant challenge in identifying a suitable equation of state. However, several previous observations suggest that a crowd’s speed distribution may conform to a two-dimensional Maxwell–Boltzmann distribution under certain yet-to-be-determined conditions. Our research uncovers that the divergence between the fluctuation velocity magnitude’s probability density function and its best 2D Maxwell–Boltzmann fit diminishes in a power-law fashion with decreasing collision time scale between individuals. These findings are robustly supported by both experimental data and simulations with diverse boundary conditions and force potentials. The equilibrium characteristics of crowds are interpreted through a canonical ensemble framework. Furthermore, we show that high pressure indicates equilibrium characteristics in human crowds. Remarkably, we reveal the ideal gas law for human crowds without resorting to any behavior assumption. We demonstrate the predictive capability of the ideal gas law on both observational and modeling data. Our research highlights a new pathway to explore and validate traditional thermodynamic quantities and laws in the setting of high-pressure human crowds, advancing our understanding of active matter systems.

Psychology of the Network Crowd as a Subject of Political Extremism

The purpose of the article is to describe the psychological characteristics of the network crowd and their use to disseminate the ideology of political extremism, as well as to propose recommendations on countering these technologies. Opening the resources of foreign social networks for its citizens, Russia has become a hostage to unspoken rules. The problem with stopping the wave of extremist content demonstrates a serious threat to ensuring information sovereignty and emphasizes the relevance and significance of studying this problem. The methods of interdisciplinary theoretical analysis, synthesis, comparison, generalization and conceptualization of the scientific ideas of the use of the psychological effects of the crowd in the network space, the content of the analysis of the posts of participants in telegram channels during protest actions were used. The article develops an author’s approach to understanding the phenomenon of the network crowd as a specific hybrid that has arisen on the basis of social networks, embodying elements of an organized nucleus and spontaneous periphery. The idea of the mechanism of control of an organized network crowd as a function from the composition of the content to create emotional resonance, the activities of groups that disperse information on social networks and infection leaders, involving participants in social networks with their emotional response, is justified. Conclusions are drawn about the conditions under which the network crowd turns into a threat. Recommendations are proposed to counteract the indication of participants in social networks by the ideology of extremism.

An enhanced multi-attribute crowd evacuation model incorporating the effects of environmental impact factors

We propose an enhanced floor field model (FFM) to analyze the behavioral characteristics of crowds with varying attributes proportions during evacuation. This model governs pedestrian movement through the Dynamic Floor Field (DFF) and the Static Floor Field (SFF). The DFF takes into account individual factors such as the gender, familiarity with the environment, and social relationships of evacuees, which influence safe evacuation. Concurrently, the SFF encapsulates the impact of environmental factors like obstacles, exits, and guidance effects. Subsequently, this refined FFM was applied and validated using a sports center evacuation scenario. The results demonstrated that the enhanced FFM accurately replicated evacuees' asymmetric behavior and queuing, and aligned well with other models when the number of evacuees fluctuated over time. In the absence of guidance, both environmental familiarity and gender emerged as primary factors influencing partial evacuation. Additionally, the gender of pedestrians significantly affected the overall evacuation. Notably, compared to pre-existing environmental information available to evacuees, the implementation of guidance to augment pedestrians' environmental familiarity resulted in a more efficient evacuation. The FFM model and these findings could be instrumental in simulating personnel evacuation and formulating emergency management strategies in crowded areas.

Experimental study on the movement characteristics of adversarial people in the corridor

With the continuous development of pedestrian dynamics, the movement characteristics of multiple types of people (elderly, children, disabled people, etc.) under various scenarios (flood, fire, special building structure, etc.) have been extensively and deeply studied. The relevant conclusions and movement rules are of great significance for guiding the design of building structures and crowd evacuation instructions. However, there is a lack of research on the movement of adversarial crowds with different movement purposes and strong interactions. The analysis of its movement characteristics and the movement modelling of interaction rules need to be further explored through controlled experiments. Therefore, this paper focuses on the study of the movement characteristics of the adversarial crowd and carries out a realistic experiment in which pedestrians in the corridor escape from attackers to complete the evacuation. The purpose was to address the following questions: the impact of attackers and corridor width on pedestrian movement. It was found that the attacker's movement pause would increase the desired velocity of the crowd, and thus the velocity-density relationship was different from the traditional fundamental diagram. The width of the corridor has a great impact on the movement speed and the collision probability of the crowd. Based on the results of the experiment, a model was established and simulation experiments were carried out by changing the width of the corridor. It is hoped that the results of this paper can provide help for emergency evacuation guidance for pedestrians in the future.

One-Shot Any-Scene Crowd Counting With Local-to-Global Guidance.

Due to different installation angles, heights, and positions of the camera installation in real-world scenes, it is difficult for crowd counting models to work in unseen surveillance scenes. In this paper, we are interested in accurate crowd counting based on the data collected by any surveillance camera, that is to count the crowd from any scene given only one annotated image from that scene. To this end, we firstly pose crowd counting as a one-shot learning task. Through the metric-learning, we propose a simple yet effective method that firstly estimates crowd characteristics and then transfers them to guide the model to count the crowd. Specifically, to fully capture these crowd characteristics of the target scene, we devise the Multi-Prototype Learner to learn the prototypes of foreground and density from the limited support image using the Expectation-Maximization algorithm. To learn the adaptation capability for any unseen scene, estimated multi prototypes are proposed to guide the crowd counting of query images in a local-to-global way. CNN is utilized to activate the local features. And transformer is introduced to correlate global features. Extensive experiments on three surveillance datasets suggest that our method outperforms the SOTA methods in the few-shot crowd counting.

APPLICATION OF CELL PHONE STATISTICS FOR ESTIMATING STRANDED PEOPLE BEHAVIOR AFTER SEVERE EARTHQUAKE IN TOKYO

Abstract. In this paper, we present a walking home simulation as anticipated after a large earthquake, and analyze people’s behaviors, walking and stopping, including the crowding of facilities by those unable to walk all the way home. For creating the necessary data for this simulation, we construct a method to estimate the spatiotemporal distribution of people with detailed individual information such as sex-age classification, and home location, by assembling population distribution data (Mobile Spatial Statistics and Person Trip survey data). The walking home simulation results verified significant variations in the crowding of facilities for stranded people due to differences in the day of the week and the time of the earthquake. Locations in Tokyo with insufficient numbers of facilities for stranded people were identified and some spatiotemporal characteristics of crowding, such as changes in crowding with time elapsed since the earthquake, were described.

Case Study of Bilibili's Marketing Techniques

In today's era, streaming media is becoming more and more prevalent and dominating the entertainment worldwide. Among them, streaming video platforms are the most popular. Take bilibili for example, it is a video platform that is very hot within China. Although bilibili has been in business for more than a decade and has a regular young user base and brand recognition, there are many issues behind the seemingly smooth sailing that cannot be ignored. These issues include bilibili's positioning, community, video content and finances, all of which have an impact on its operations. As a nascent streaming platform, bilibili has a variety of marketing models as well as marketing tools. This paper will analyze bilibili's marketing model, marketing tools and strategies using the SWOT and 4P models, and make decisions about the problems the company is experiencing. After analyzing according to various models, bilibili's marketing means is mainly to serve the Z-era crowd as the main target group, and for the characteristics of the Z-era crowd, users can access user-generated content on B station, and the platform also includes a close-knit community. These two locations will be used by B station in an effort to monetize and profit from the traffic. And bilibili attracts the love of fixed groups with its unique secondary yuan-oriented style, achieving a long-lasting emotional maintenance with core users.

Multi-level learning counting via pyramid vision transformer and CNN

Severe scale variation has become a challenging issue for hindering the improvement of accuracy in crowd counting task. To tackle the problem, we propose a Pyramid Transformer CNN Network (PTCNet), an effective combination of the transformer and the CNN, which possesses both the global receptive fields and the locality to deal with the severe scale variation problems and boost the prediction accuracy. Firstly, we utilize the pyramid vision transformer to extract multi-level global context information of the crowd, aiming at different head scales. And then, the multi-level information is fully fused in the multi-level feature aggregation module where detailed crowd characteristics from all feature spaces are preserved to be further processed. Finally, we design a multi-branch regression head to enrich the crowd features for strong representations and regress the density maps. Extensive experiments on challenging datasets with complex scenarios and multiple scales demonstrate the effectiveness of the our method. The proposed method achieves competitive results comparing with the state-of-the-art approaches and achieves state-of-the-art results(MAE:51.7, RMSE:79.6) on ShanghaiTech Part_A dataset.

Human–Computer Interaction and Participation in Software Crowdsourcing

Improvements in communication and networking technologies have transformed people’s lives and organizations’ activities. Web 2.0 innovation has provided a variety of hybridized applications and tools that have changed enterprises’ functional and communication processes. People use numerous platforms to broaden their social contacts, select items, execute duties, and learn new things. Context: Crowdsourcing is an internet-enabled problem-solving strategy that utilizes human–computer interaction to leverage the expertise of people to achieve business goals. In crowdsourcing approaches, three main entities work in collaboration to solve various problems. These entities are requestors (job providers), platforms, and online users. Tasks are announced by requestors on crowdsourcing platforms, and online users, after passing initial screening, are allowed to work on these tasks. Crowds participate to achieve various rewards. Motivation: Crowdsourcing is gaining importance as an alternate outsourcing approach in the software engineering industry. Crowdsourcing application development involves complicated tasks that vary considerably from the micro-tasks available on platforms such as Amazon Mechanical Turk. To obtain the tangible opportunities of crowdsourcing in the realm of software development, corporations should first grasp how this technique works, what problems occur, and what factors might influence community involvement and co-creation. Online communities have become more popular recently with the rise in crowdsourcing platforms. These communities concentrate on specific problems and help people with solving and managing these problems. Objectives: We set three main goals to research crowd interaction: (1) find the appropriate characteristics of social crowd utilized for effective software crowdsourcing, (2) highlight the motivation of a crowd for virtual tasks, and (3) evaluate primary participation reasons by assessing various crowds using Fuzzy AHP and TOPSIS method. Conclusion: We developed a decision support system to examine the appropriate reasons of crowd participation in crowdsourcing. Rewards and employments were evaluated as the primary motives of crowds for accomplishing tasks on crowdsourcing platforms, knowledge sharing was evaluated as the third reason, ranking was the fourth, competency was the fifth, socialization was sixth, and source of inspiration was the seventh.

Simulation investigation on crowd evacuation strategies for helping vulnerable pedestrians at different stages of egress

In emergencies, evacuations move pedestrians to safety. Extensive research has investigated strategies that can improve evacuation efficiency of pedestrian crowds in various contexts. However, most of this research considers pedestrians as homogeneous agents and the question of how vulnerable pedestrians can be supported in evacuations has not been convincingly resolved. Here, we focus not only on pedestrians with the same characteristics but also on vulnerable pedestrians defined by moving at a lower speed or being located further away from exits. We investigate strategies for helping these vulnerable pedestrians in three evacuation stages: the pre-evacuation stage where vulnerable pedestrians can respond to incidents quickly, the response stage where vulnerable pedestrians can be preferentially assigned to exits in the case of limited exit resources, and the evacuation phase where we place an obstacle in front of the exit. Our simulation results indicate that the effectiveness of these strategies depends on crowd characteristics and contexts (e.g., crowd size and pedestrian initial distribution), and different strategies may have other consequences, such as making pedestrians experience a higher accumulated force or walk longer distances. We found that only in some cases machine learning methods including kernel naïve Bayes and Linear SVM can predict which strategy will work best based on the initial distribution of pedestrians, suggesting that at least sometimes it may be possible to select strategies based on observed conditions. Our work highlights the importance of context and heterogeneous populations when developing strategies for evacuations and may thus be helpful for crowd safety and evacuation management.

Correlating viscosity and molecular crowding with fluorescent nanobeads and molecular probes: in vitro and in vivo.

In eukaryotes, intracellular physico-chemical properties like macromolecular crowding and cytoplasmic viscoelasticity influence key processes such as metabolic activities, molecular diffusion and protein folding. However, mapping crowding and viscoelasticity in living cells remains challenging. One approach uses passive rheology in which diffusion of exogenous fluorescent particles internalized in cells is tracked and physico-chemical properties inferred from derived mean square displacement relations. Recently, the crGE2.3 Förster resonance energy transfer biosensor was developed to quantify crowding in cells, though it is unclear how this readout depends on viscoelasticity and the molecular weight of the crowder. Here, we present correlative, multi-dimensional data to explore diffusion and molecular crowding characteristics of molecular crowding agents using super-resolved fluorescence microscopy and ensemble time-resolved spectroscopy. We firstly characterize in vitro and then apply these insights to live cells of budding yeast Saccharomyces cerevisiae. It is to our knowledge the first time this has been attempted. We demonstrate that these are usable both in vitro and in the case of endogenously expressed sensors in live cells. Finally, we present a method to internalize fluorescent beads as in situ viscoelasticity markers in the cytoplasm of live yeast cells and discuss limitations of this approach including impairment of cellular function.

Prediction model of crowd noise in large waiting halls.

Crowd noise is usually the primary noise in large waiting halls, and it is difficult to predict because it is influenced by several factors such as room acoustics and crowd characteristics. This study developed a crowd noise prediction model based on the superposition of direct and reverberant sound energy using the factors of the spatial layout of waiting halls, number and distribution of crowds, behavior ratio (ratio of vocal passengers to the total number of passengers), and average crowd sound power. To verify the model, on-site measurements were conducted in two large waiting halls to obtain the necessary input parameters. The crowd noise levels in one of the waiting halls were obtained from 1-s noise level data after excluding broadcast periods. A method for determining an individual's average sound power based on the model was also presented and found to be approximately 70.6 dB. Finally, the model was verified using measured data, and it showed that the model could accurately predict the average crowd noise level and changing trend of crowd noise in temporal and spatial dimensions with an average R-square of approximately 0.55 and average difference of approximately 1.1 dBA between the predicted and measured results.

Application analysis of highway traffic accident risk model based on geographically weighted negative binomial regression

The traditional generalized linear model (GLM) can not effectively analyze discrete road traffic accidents when analyzing road traffic accidents with spatial dependence and heterogeneity. Therefore, a risk analysis method of highway traffic accidents based on geographically weighted negative binomial regression model (GWNBR) is proposed. Using geographical weighted regression (GWR) model and negative binomial regression (NB) model, this paper makes a comparative analysis of highway traffic accidents in Xi’an, including local spatial geographical weighted Poisson regression (GWPR) model and two geographical weighted negative binomial regression (GWNBRg and GWNBR) models. The corresponding model bandwidth is determined, and the performance of the model is compared based on the data of traffic environment, road characteristics, crowd characteristics and road alignment. The experimental results show that compared with the single NB model, the proposed model can effectively reduce the interference of the spatial nonstationarity of the data, and can effectively extract the risk factors affecting the accident. The coefficients of GWNBRg model and GWNBR model are positive, which are better than GLM in the mean and likelihood of the residuals. The spatial autocorrelation of the residuals is significantly reduced, and the significance level is 5%, which reduces the spatial heterogeneity of the data. The over dispersion parameter value of GWNBRg model shows a downward trend from southwest to northeast in space, which can effectively reflect the spatial relationship between traffic flow and accident rate, indicating that GWNBR model has a good effect in traffic accident risk analysis of super discrete highway. Therefore, the application of geographical weighted negative binomial regression to highway traffic accident risk prediction has a good application effect, and can effectively reduce the probability of accidents as safety prevention and early warning.

Assess the satisfaction of Van Hien University students with the service quality of Ohao smart

The study presents an overview of the characteristics of the research sample, a general description of the sample's response results, and the results of testing the measurement scales, the research sample has reflected the characteristics of the research crowd. The results of the scale test using Cronbach's Alpha and the EFA factor analysis help us extract the components of leadership style affecting student satisfaction: (1) Ohao Smart design and landscape, (2) Ohao Smart responsiveness, (3) Ohao Smart support service quality, (4) Food quality /reliability, (5) Ohao Smart price. The scale of research concepts has reached convergent and discriminant validity through satisfying the conditions of EFA factor analysis and Pearson correlation analysis. The test results of linear regression show that all 5 components have a leadership style that positively impacts student satisfaction at a significant value of < 0.05 (95% significance level). The model test results show that the model is suitable, there is no violation of the test assumptions, and the results of testing the hypotheses are accepted.

Evaluation of road traffic noise exposure considering differential crowd characteristics

• A road traffic noise evaluation method for heterogeneous crowds is presented. • Heterogeneous crowd distribution, age and acoustic requirements are considered. • A noise response curve (NRC) is established. • The application verified the reasonableness and practicality of the method. The primary objective of this study is to develop an urban traffic noise exposure evaluation method for differential crowd characteristics, which takes crowd distribution, ages and acoustic requirements into consideration. First, road traffic noise data are simulated using a line-source model, and crowd distribution data are obtained by the decomposition of demographic data based on the different building attraction coefficients of the heterogeneous crowd. Then, the noise response curve (NRC) of the heterogeneous crowd is established with the normalization of subjective feelings of respondents and calibrating the response level of each age group in combination. Finally, with the help of the proposed noise evaluation indicators, and considering the noise function requirements (NFR), the traffic noise evaluation of differential crowd characteristics is realized. The method is applied to a typical scenario, and the results and analysis verified the reasonability and practicability of the method in urban traffic noise exposure evaluation.

A dual adaptive cellular automaton model based on a composite field and pedestrian heterogeneity

To better reveal the law of pedestrian flow and the mechanism of crowd accidents, this paper establishes a dual adaptive cellular automaton model (DACA) that integrates pedestrian heterogeneity, local density, target distance and asynchronous behaviour. The validity of the model is verified, and when the crowd density is greater than 3 persons/m2, the free movement of pedestrians is generally constrained by crowd density. In contrast, the critical density that affects the velocity of pedestrian movement is positively correlated with the complete free movement velocity (CFMV) of pedestrians. The steady-state density of the straight channel is negatively correlated with CFMV. A negative correlation between CFMV and the average crowd density at maximum flow is observed, and the time from highest density to steady density is independent of CFMV, but when the density is not the limit, the crowd dredging time is related to CFMV and the initial density of congestion. Finally, a case study of the start-up characteristics of a static crowd in a straight channel shows that the transmission speed of the start information under the limit density is greater than the moving speed of the crowd. The self-dredging process of crowds consumes additional time. The crowd distribution is sparse in the front and dense in the tail, and the crowd gathering risk is found to be directly proportional to the initial crowd density at the acceptable risk level. This study can effectively explore the laws of crowd movement and provide a theoretical reference for crowd safety management.