Purple fringing, a persistent artifact caused by Longitudinal Chromatic Aberration (LCA) in camera lenses, has long degraded the clarity and realism of digital imaging. Traditional solutions rely on complex and expensive apochromatic (APO) lens hardware and the extraction of handcrafted features, ignoring the data-driven approach. To fill this gap, we introduce DCA-LUT, the first deep learning framework for purple fringing removal. Inspired by the physical root of the problem-the spatial misalignment of RGB color channels due to lens dispersion, we introduce a novel Chromatic-Aware Coordinate Transformation (CA-CT) module, learning an image-adaptive color space to decouple and isolate fringing into a dedicated dimension. This targeted separation allows the network to learn a precise "purple fringe channel," which then guides the accurate restoration of the luminance channel. The final color correction is performed by a learned 5D Look-Up Table (5D LUT), enabling efficient and powerful non-linear color mapping. To enable robust training and fair evaluation, we constructed a large-scale synthetic purple fringing dataset (PF-Synth). Extensive experiments in synthetic and real-world datasets demonstrate that our method achieves state-of-the-art performance in purple fringing removal.

Purple flare, a diffuse chromatic aberration artifact commonly found around highlight areas, severely degrades the tone transition and color of the image. Existing traditional methods are based on hand-crafted features, which lack flexibility and rely entirely on fixed priors, while the scarcity of paired training data critically hampers deep learning. To address this issue, we propose a novel network built upon decoupled HSV Look-Up Tables (LUTs). The method aims to simplify color correction by adjusting the Hue (H), Saturation (S), and Value (V) components independently. This approach resolves the inherent color coupling problems in traditional methods. Our model adopts a two-stage architecture: First, a Chroma-Aware Spectral Tokenizer (CAST) converts the input image from RGB space to HSV space and independently encodes the Hue (H) and Value (V) channels into a set of semantic tokens describing the Purple flare status; second, the HSV-LUT module takes these tokens as input and dynamically generates independent correction curves (1D-LUTs) for the three channels H, S, and V. To effectively train and validate our model, we built the first large-scale purple flare dataset with diverse scenes. We also proposed new metrics and a loss function specifically designed for this task. Extensive experiments demonstrate that our model not only significantly outperforms existing methods in visual effects but also achieves state-of-the-art performance on all quantitative metrics.

The integration of physical and cyber networks can be a double-edged dynamic. While it accelerates the dissemination of positive information, it also intensifies the diffusion of negative information. A clear understanding of the stochastic dynamics of competitive information dissemination is essential. It enables effective control strategies to curb the spread of negative information and maintain social stability. However, designing and controlling the dissemination of competitive information in cyber–physical integrated networks (CPINs) is challenging due to competition and stochasticity. Moreover, the heterogeneity of CPINs further aggravates this problem. To address this, we propose a competitive information dissemination method to capture and control the stochastic dynamics in CPINs. Specifically, we build a CPIN model to characterize the heterogeneity between physical and cyber networks. Furthermore, we develop a temporal point process-based competitive information dissemination model (TPP-CIDM) that captures the stochastic evolution of both positive and negative information. This dissemination model quantifies the stochastic dynamics by computing the probability distribution of the sizes of competitive information, reducing biases inherent in deterministic solutions. Finally, we design an event-driven optimal control (EOC) strategy to dynamically modulate the intervention intensity. The intervention optimization problem is formulated to maximize utility under cost constraints, and a heuristic solution is provided. Numerical simulations on both synthetic and real-world networks demonstrate the effectiveness of the proposed method.

Although Segment Anything Model (SAM) have recently achieved remarkable progress, their generalization capability in infrared small target detection remains limited due to the inherently high noise levels in infrared imagery. To preserve the generalization and noise suppression ability of the model, we propose an method called NSDSAM, a noise-suppression-driven approach that enhances SAM at both internal and external levels. Internally, we develop a Hybrid Adapter for suppressing the noise of feature maps, consisting of an MLP adapter and a self-attention adapter. The self-attention adapter first performs entropy-aware reconstruction of features from noisy inputs and employs a gating mechanism for soft-attention fusion, mitigating SAM’s sensitivity to noise. Externally, we design a Spatial-Frequency hybrid Module (SFHM) that jointly processes spatial and frequency domains to overcome the self-attention model’s bias toward low-frequency components, further strengthening the suppression of background clutter and noise. Extensive experiments on multiple infrared datasets demonstrate that the proposed method achieves state-of-the-art (SOTA) performance in infrared small target detection. The project code is available upon acceptance.

Group features have significant effects on pedestrian movement and constitute a focal point in pedestrian trajectory prediction research. In reality, pedestrians within a group exhibit notable consistency features due to their compact spatial positions, close destinations, and factors such as coordination within the group. In contrast, owing to the dispersed destinations among groups and the lack of coordination, there are significant differences in velocity and direction between the groups, leading to strong conflicts. However, existing pedestrian trajectory prediction models based on group features lack sufficient quantification of both within-group and between-group features. To address this problem, we propose Group-PTP, a novel pedestrian trajectory prediction model based on group features. Specifically, we first propose a group graph attention network-based group features aggregation method (Group-GAT). By quantifying and aggregating the intra-consistency and inter-conflict features exhibited by the groups, our method can better capture the features and interactions both within and between groups. Second, we propose a group multi-feature information representation model that fuses captured group aggregate features, pedestrian coordinates, surrounding pedestrian features, and obstacle features through fusion concatenation. Finally, we propose a multi-feature temporal convolutional network (MF-TCN) that embeds the impact weights of multi-feature information into pedestrian coordinates to obtain feature outputs and conducts temporal operations on feature outputs to predict future trajectories. The experimental results demonstrate that our proposed Group-PTP achieves state-of-the-art performance on several different trajectory prediction benchmarks.

With the rapid development of AR/VR technologies, achieving natural and seamless human-scene interactions has emerged as a critical challenge in computer vision. Existing methods suffer from low model placement accuracy and unnatural scene interactions. Therefore, we propose a framework called human-scene interaction with geometric and physical constraints (GP-HSI), which places a given pose of a 3D human model in an appropriate position within a 3D scene by establishing geometric and physical constraints, while ensuring interactive fidelity between the human and the scene. Specifically, first, we propose a pose-guided human contact semantic generation method, which generates human semantic labels by classifying the given human poses. Second, we propose a geometrically and semantically constrained human model placement method, which determines the optimal position of the human model in the scene by constraining the geometric proximity and semantic consistency between models. Third, we propose an inverse kinematics-based pose adjustment method, which finds the target human-scene interaction points by constructing a heterogeneous kinematic tree and solves the rotation matrix of human joints to obtain a physically plausible optimal human pose. At last, we develop an interactive system to visualize the generated human-scene interaction. The results of qualitative and quantitative experiments show that our approach is able to place human models at appropriate locations in the scene and generate plausible interactions.

Trust plays a crucial role in crowdsourcing Internet of Things (IoT), as it can be used to select trustworthy participants to improve the quality of crowdsourced services and strengthen system security. While traditional research has focused on micro-level aspects, including trust computation and propagation, a comprehensive macro-level trust analysis remains underexplored. In this paper, we propose a macroscopic trust ecosystem analysis framework for crowdsourcing IoT services. We first construct a Trust Ecosystem Model (TEM), where trust clusters serve as an abstraction to capture and quantify overall trust characteristics based on their size and structure. To analyze the dynamic evolution of TEM, we propose a Percolation-based Trust Ecosystem Analysis Model (P-TEAM), which maps the formation of trust clusters to a joint site-bond percolation process. Thus, the study of TEM evolution can be reframed into an investigation of how trust clusters evolve as users' trust attributes change. Through P-TEAM, we identify the critical thresholds associated with trust attributes that trigger trust phase transitions in crowdsourcing IoT services, which act as key metrics for evaluating the ecosystem's robustness macroscopically. Finally, we further evaluate the trust ecosystem beyond these thresholds by calculating the proportions of trusted giant components. We validate our approach on directed networks, using both synthetic and real-world datasets. The experimental results further substantiate our findings and provide valuable insights into constructing a healthy and sustainable trust ecosystem for crowdsourcing IoT services. Codes are available athttps://github.com/sd-sclab/MacroTrust.

In emergencies, negative emotion has a significant impact on decision-making during crowd evacuation. Psychological studies suggest that negative emotion in decision-making has a double-edged sword effect. Excessive negative emotion has adverse impacts, such as causing crowd chaos and congestion. Conversely, moderate negative emotion has a positive effect by speeding up crowd movement. However, current researches mainly focus on one aspect which is how to reduce the negative effects of negative emotion on crowd evacuation, while overlooking the benefits of negative emotion. How to fully explore the double-edged sword effect of negative emotion and regulate negative emotion to improve the efficiency of crowd evacuation is still an open issue. To achieve this, we propose the NegEmotion model which considers the positive impact of negative emotion on crowd evacuation, and regulates crowd emotion by controlling knowledge spreading according to Siminov’s psychological principle. In this model, the knowledge spreading network (KSN) and the stress emotional contagion network (SECN) are constructed. Based on these networks, we study the evolution process of knowledge spreading and stress emotional contagion, respectively. Next, we formulate the emotional regulation as an optimization problem to maximize the efficiency of crowd evacuation. Then, a heuristic algorithm is used to solve for the optimal emotional regulation strategy. Finally, a crowd simulation system is implemented to verify the effectiveness of our NegEmotion model. The experimental results show that our method is effective to improve the efficiency of crowd evacuation.

Robot-assisted interventions are often provided for crowd evacuation to prevent accidents and keep crowds safety, which is currently a hot topic of research. However, the existing work has ignored the impact of crowd motion chaos which is a fundamental factor determining the effective evacuation. To assist the crowd evacuation more efficiently and reduce the intervention costs, robots need to identify destructive individuals based on the randomly evolved chaos and intervene in them specifically. To this end, we propose a stochastic optimal intervention method to provide the optimal strategies for robot-assisted crowd evacuation. First, the definition of chaos is introduced to quantify the chaotic degree of crowd movement. Then, we study the stochastic evolution of chaos by exploring the chaos state transition with marked temporal point process (MTPP) and stochastic differential equations (SDEs). Next, the robot-assisted intervention is formulated as a stochastic optimal intervention problem that seeks to maximize the intervention utility. Fortunately, a closed-form solution is obtained by using the Bellman’s principle of optimality and solving the derived Hamilton-Jacobi-Bellman (HJB) equation. At last, we develop an event-driven crowd motion system to simulate crowd evacuation process. The simulation results show that our method can prevent the occurrence of chaos effectively, and improve the efficiency of crowd evacuation significantly. The proposed method is expected to provide guidance for emergency management.

Crowd evacuation simulation can provide guidance for public emergencies or mass casualties which is easy to cause crowding and stampede. Deep reinforcement learning is an effective method for the path planning of crowd evacuation simulation which can reduce the dependence on data and has strong generalization. However, current methods of deep reinforcement learning ignore the consideration of panic emotion which could result in weak authenticity since the panic emotion has a significant impact on crowd evacuation. To address this problem, we propose a novel deep reinforcement learning for the panic crowd evacuation. First, a quadrant based crowd network is constructed to guide the movement of agents according to the change of panic degree in different quadrants. Second, a method of quantifying panic degree is proposed. Through the mean field equation, the dynamic evolution of agent states is realized, and the panic degree in different regions is quantified. Finally, a panic crowd evacuation model based on multi-agent deep deterministic policy gradient (MADDPG) was established. The computated panic degree is introduced into the reward function to recommend a path for multiagent to avoid panic. Experimental results show the effectiveness of our method.

The global spread of COVID-19 causes great losses to human society. Accurate calculation of the scale of epidemic spread is of great significance for the implementation of corresponding epidemic prevention measures. However, the existing method ignores the group formed by social relations of the population, which reduces the accuracy of the epidemic spread number calculation. In this paper, we propose an epidemic model based on intra- and inter-group interactions. Firstly, we construct a dual network model of epidemic spread based on intra- and inter-group interactions. The network describes how epidemics spread intra- and inter-group. To capture the intergroup influences, we construct a model for social mobility to calculate the inter-group spread rate. Secondly, we propose a computational model for the epidemic spread. We calculate the infection probability of groups in the upper layer network by using a continuous-time Markov chain (CTMC). We describe a dynamic evolution of the intra-group infection in the underlying network based on the mean field equation. And the number of infections in the population is calculated by integrating intra- and inter-group effects. Finally, we implement an epidemic spread simulation system to visualize the spread process. The experimental results show that the model can analyze the epidemic spread process more accurately.

There are two types of research on abnormal crowd behavior detection: micro modeling methods (e.g., crowd pose) and macro modeling methods (e.g., optical flow methods). In crowd evacuation, however, micro methods fail to solve the occlusion problem, while macro methods fail to address issues such as poor real-time performance and non-adaptive motion target detection. To address these issues, we propose a method that combines macro and micro methods to detect abnormal crowd behavior. Firstly, we extract the two-dimensional poses of moving people using the human pose estimation algorithm (OpenPose) and get the corresponding micro motion features. Secondly, we create an optical flow map of the video using the dense optical flow algorithm (Farneback) and a scene activity map by comparing the optical flow sizes between two consecutive frames. Then we derive the entropy change curve, which represents the macro motion features. Finally, we fuse micro and macro motion features and put the fused hybrid feature vectors into our classification network to train our model, and then find the transition time from micro to macro abnormal states in crowd evacuation videos. The experimental findings demonstrate that our proposed method can more accurately and quickly identify the abnormal behaviors that appear in evacuation scenes.

Studying the spread process of epidemics in the crowd is a significant topic, which can help to take measures in advance to avoid epidemic spread and reduce the losses of life and property of the general public. However, this is a challenging problem that how to capture the spatio-temporal characteristics of the epidemic spreading. In this paper, we propose a risk region-based prediction model to characterize the spread feature. First, we propose a risk region-based spread model to describe the spatial relationship between individuals by dividing risk regions and assigning different spread characteristics to different risk regions. Second, we develop a spatio-temporal point process-based infection intensity quantification method to describe the variation of epidemic spread characteristics over time in different risk regions in the form of intensity. Then, we propose a long short-term memory (LSTM) based intensity function prediction method to solve the corresponding intensity and predict the spread process by the spatio-temporal point process (STPP). Finally, we implement an epidemic spreading simulation platform to verify our method and visualize the experimental results. The proposed model is expected to provide guidance for predicting the spread of epidemics.

Numerous cameras deployed in venues can collect video data, which can be analyzed to help evacuate people in emergency situations. Most surveillance videos can provide data support for pedestrian trajectory prediction. However, data-driven prediction method do not consider the impact of personalization on pedestrian trajectories, which results in ignoring individual personalization. How to accurately analyze the differences among pedestrians and yet accurately predict pedestrian trajectories is a challenging problem. To solve the above problems, we propose an singular value decomposition (SVD) based pedestrian trajectory prediction method, which introduces matrix decomposition to pedestrian trajectory prediction for the first time. Thus, This method analyzes the impact of pedestrian personalization on motion trajectories by mining the interaction patterns between pedestrians and environment. Firstly, we propose an information collection method based on environmental semantics, which can extract scene information from historical video data to construct an environmental information matrix. Secondly, we propose an SVD-based individual environmental feature preference method, which uses singular value decomposition methods to mine the data and analyze personalized pedestrian motion patterns to construct a personalized individual preference matrix. Finally, we built an personalized trajectory prediction method to predict pedestrian movement trajectories. The experimental results show that the method can not only analyze the effect of personalization on pedestrian movement but also accurately predict the pedestrian movement trajectory.

In crowd evacuation scenarios, it is an effective way to ensure evacuation safety by recognizing real crowd emotions and then taking measures such as emotional infection to reduce crowd panic. However, factors such as distance, exposure, angle, and occlusion can lead to incomplete face expression information collection, thus failing to accurately identify crowd emotions. Therefore, it is still a very challenging problem to recognize individual emotions in the case of incomplete face expression information collection, and thus accurately identify crowd emotions during evacuation. To solve this problem, we propose a spatial-temporal consistency-based crowd emotion recognition method to accurately identify the real emotions of the crowd. First, for video frames that can capture the complete facial expression information, we use the residual network to accurately identify the individual emotion values in each frame. For video frames that cannot capture the complete facial expression information, we propose an individual emotion calculation model based on spatial-temporal consistency to calculate the individual emotion values in each frame. Second, we define the crowd panic level and obtain the real crowd emotion by calculation. Finally, we implement an end-to-end crowd panic emotion recognition system to verify our method. The experimental results show that the method can accurately calculate the crowd panic level, which is important for guiding crowd evacuation.

Thyroid nodule is a common clinical disease, although most nodules are benign, the incidence rate of thyroid cancer has risen rapidly in recent years. Even though many methods have achieved automated thyroid nodule segmentation based on deep learning, these methods are based on supervised learning and require a large amount of labeled data for training. However, the labeling work must be carried out by professional doctors, which results in a small number of datasets and difficulty in labeling. To address this problem, this paper proposes a semi-supervised thyroid nodule segmentation model via joint consistency learning and co-training. This model includes two branches: consistency learning and co-training. In the consistency learning branch, based on consistent regularization, the teacher model guides the student model to optimize. In order to make the teacher model more stable, we design a co-training framework to further optimize the teacher model. In co-training branch, the teacher model and TransUNet extract different representations of the same sample and teach each other to prevent consistent but incorrect predictions between the teacher model and the student model. This semi-supervised model can learns useful feature representations from unlabeled data, and effectively trains the model with a small amount of labeled data, reducing the dependence on labeled data during the model training process.

Collaborative crowdsensing (CCS) requires the recruited team to collaborate closely to complete sensing tasks with high quality of service (QoS). The team recruitment of CCS is mainly influenced by the subjective willingness of participants and the objective trust evaluation of the sensing platform; that is, the higher the subjective mutual willingness to work together and the objective mutual trust among participants, the more efficiency with which the CCS tasks will be achieved. However, the existing research lacks comprehensive consideration of mutual willingness and mutual trust among recruited participants. This results in poor QoS. To address this problem, we propose a novel team recruitment method for CCS that jointly considers the willingness and trust to recruit optimal teams. First, we build a graph convolutional network‐based willingness‐trust network (GCN‐WTN) model for CCS to obtain mutual willingness and trust among participants more accurately. Second, we propose a willingness and trust‐based team recruitment (WT‐TR) method to recruit the optimal teams for CCS. This method introduces the consensus and similarity constraints into the willingness and trust networks to better meet the collaboration needs of CCS. Finally, we implement a recruitment simulation platform for CCS to simulate the team recruitment process and validate the effectiveness of our proposed method. The experimental results show that the teams recruited by the proposed method can significantly improve QoS for CCS.

Traditional crowd evacuation simulation methods focus on the analysis of individual behavior in the crowd, but ignore the analysis of potential group properties in the process of crowd motion, which reduces the visual realism of the crowd evacuation simulation process. During crowd motion, crowds cause unconscious self-organization due to common destinations and social relationships, thus forming groups. Considering the influence of group properties on crowd motion is important to improve the visual realism of crowd motion simulation. To address this problem, we propose a crowd motion modeling method for group properties analysis (GPA-CMM). First, we build a data-driven group properties quantification (DGPQ) model to describe the characteristics of group motion accurately. In the model, we divide the crowd into several groups, extract motion properties of the crowd, and quantify the features of intra-group stability and inter-group conflict. Then, in order to analyze their influences of group properties, we build a stability and conflict based crowd motion analysis (SC-CMA) model. Finally, we implement a crowd simulation system based on SC-CMA, visualize the results of the theoretical analysis in a graphical way. The experimental results show that the method can simulate crowd motion more realistically.

As crowd sensing tasks become more complex, it is increasingly needed for participants to form teams to collaborate interactively in order to complete tasks more efficiently. It is valuable to fully understand user behavior, the organizational principles of crowd collaboration teams and incorporate them into the user recruitment process. In fact, there is a two-way process of subjective and objective selection for building a team in crowd sensing. From the participant’s perspective, the willingness of users to participate in the task is a subjective factor; from the platform’s perspective, the trust relationship between users is an objective factor. However, existing work has not well recognized the value of including collaborative teams in the crowd sensing recruitment process and has not considered the willingness and trust relationships between participants simultaneously, which results in poor quality of service (QoS) among recruited users. To address this problem, we propose a willingness and trust-based collaborative team recruitment method (WT-CTRM). First, we construct a willingness network (WN) and a trust network (TN) to describe the interaction of willingness and trust among the collaborated users respectively. Based on these networks, we then predict the unknown willingness and trust relationships between users for collaboration by using graph convolutional networks (GCNs). Finally, we build collaborative teams for the user recruitment process of crowd sensing based on the consensus constraints of the predicted willingness network and trust network. Simulation results show that the recruitment scheme significantly improves the QoS in task scenarios with collaborative requirements.

With the continuous urban expansion, the safety of public places has attracted more and more attention. Once an accident occurs and people distrust their surroundings, it will cause congestion, trampling and other events. Individual trust value is particularly important in crowd evacuation. However, it is very difficult to evaluate the individual credibility with existing methods. We propose a blockchain-based crowd-sensing trust management mechanism (B-CSTM) to address this problem. First, we provide a message credibility evaluation mechanism that can evaluate the credibility of messages sent by individuals on spot. Second, we developed a trust calculation method that uses a distributed Evacuation Perception Unit (EPU) to calculate individuals’ trust values by perception scores. Finally, we build a blockchain-based trust management mechanism model that uses blockchain to store and query trust values. Through experimental analysis, we visualize the trust results in the evacuation scenario. It is shown that the mechanism is feasible for the management of trust values (collection, calculation and storage, etc.) in crowd evacuation.

The surveillance videos captured by cameras deployed in large-scale places can provide help for path planning of crowd evacuation in emergencies. However, most surveillance videos are not allowed to be shared in order to avoid privacy disclosure. Thus, it is difficult to obtain the video data of the whole scenario for path planning which greatly reduces effectiveness of path planning. How to provide global path planning for crowd evacuation while ensuring privacy protection is a challenging problem. To solve this problem, we propose a federated learning based path planning method for crowd evacuation. In this method, the potential field model are introduced into the federated learning framework to provide the privacy protection since only the potential field information is aggregated between the camera terminal and the central server, instead of the surveillance videos. First, we construct the local potential field for the surveillance scene of each camera by extracting the scene information and the crowds information. Second, the central server periodically performs the global potential field information aggregation and integrates these information to plan the optimal global path dynamically for crowd evacuation. Finally, we implemente a simulation platform to simulate the process of crowd movement and visualize the dynamic change of the potential energy field. Experimental results prove that the proposed method can provide privacy-preserving path planning for multi-camera evacuation scenarios.

Understanding the emotional contagion process in the crowd will help to take measures in advance to avoid the large-scale spread of negative emotions in emergencies and reduce the loss of lives and properties. Studying the phase transition phenomenon is fundamental to analyzing and evaluating the crowd emotional contagion. However, it is a challenging issue since most people participate in both the physical and cyber networks at the same time. In this article, we focus on the emotional contagion in physical-cyber integrated networks from the phase transition perspective. To achieve this, we first construct a physical-cyber integrated network model to describe the interactions between physical and cyber networks. Second, we build an emotional contagion model to capture the characteristics of emotional contagion in the physical and cyber integrated networks accurately. Finally, we study the phase transition phenomenon of emotional contagion and identify the critical threshold by mapping the emotional contagion to the joint site/bond percolation model. Numerical simulations and experiments further support and enrich our conclusions. The proposed method is expected to provide guidance for controlling emotional contagion in emergencies.

Crowd emotional contagion is affected by physical cyberspace and virtual cyberspace at the same time. In addition, the individual personality also plays an important role in crowd emotional contagion. However, few existing works have investigated the two factors jointly. To solve this problem, we propose a personalized virtual and physical cyberspace-based emotional contagion model (PVP-ECM) to simulate the process of crowd emotional contagion. First, we construct an individual emotion model and propose the personalized emotional contagion rules to consider the influences of the individual personality on crowd emotional contagion coupling the virtual and physical cyberspaces. Second, we use mean field theory to derive an evolution process and obtain mean field equations for the PVP-ECM. Then, we solve the PVP-ECM numerically using the finite difference method. Third, we construct a personalized BA scale-free network for the PVP-ECM simulation to further verify the stability of the model. Finally, we perform the PVP-ECM simulation and implement a simulation system to visualize the results of our theoretical analysis. The experimental results show that our approach can simulate the process of crowd emotional contagion more realistically. The proposed method can provide guidance for coping with the personalized crowd emotional contagion of public emergencies.

Negative emotion will spread among the crowd, and make people irrational. Intervention can reduce crowd panic by isolating the source of transmission or spreading positive emotions, which is an effective measure to ensure the safety of the crowd. However, it is still a very challenging problem that how to use appropriate intervention to maximize the influence to infected individuals while minimizing the cost of the intervention. To this end, we present an intervention optimization method for crowd emotional contagion (IO-CEC) to adaptively intervene in negative emotional contagion. First, we build a stochastic events based emotional contagion model (SEEC) to capture the dynamic evolution characteristics of crowd emotional contagion. In this model, we combine the SIS infection model and the Markov stochastic process to study the dynamic change of the number of infected individuals. Second, we propose a Hawkes process based intervention model (HIM) to describe the intensity of intervention. By using the Hawkes stochastic point process, two intervention intensity functions are defined to quantify the infection rate and recovery rate. Then, in order to improve the efficiency of intervention, we formulate the intervention optimization as a problem of maximizing the utility and use a heuristic algorithm to solve it. Finally, we implement a crowd emotional contagion simulation system to visualize the results of emotional intervention. Experimental results show that our method is effective and has great significance for guiding realistic emotional intervention.

Data-driven crowd evacuation learning methods are often used to enhance the realism of crowd simulation. However, the learning results of traditional methods cannot adapt to the dynamic changes of the simple scene, and thus have the disadvantage of poor generalization. To solve this problem, we propose a data-driven crowd evacuation framework based on hierarchical deep reinforcement learning. The framework consists of: a macro-control layer with path programming function and a micro control layer with collision avoidance function. In this paper, a path programming method combining data-driven and deep reinforcement learning is proposed in the macro-control layer. The method combines the pedestrian motion attributes in the video with the DDPG algorithm to learn the pedestrian track in the video from a macro perspective. In the micro-control layer, the track sequence learned in the macro-control layer is used as the motion target to learn the collision-free motion velocity of individuals using the multiple agent deep reinforcement learning method. When the scene changes, the micro-control layer adaptively adjusts the motion speed without the need for the macro-control layer to repeat the path programming learning. The experimental results demonstrate that the proposed hierarchical crowd evacuation framework can not only simulate the real crowd movement behavior and improve the simulation fidelity, but also flexibly adapt to the dynamic changes of the simple scene and enhance the generalization.

In emergency evacuations, crowds often become congested and stampeded because of extreme panic, resulting in injuries and fatalities. Safety officers can spread positive emotions and reduce crowd panic by issuing information or appeasement, which is an effective way to ensure evacuation safety. However, how to deploy safety officers appropriately and maximize positive emotional contagion to cover the largest number of chaotic individuals is still a challenging problem. To solve this problem, we propose an IoT-based positive emotional contagion (IoT-PEC) method for crowd evacuation. First, we employ IoT technologies to capture the attributes of crowd movement, such as position and velocity. Based on these attributes, the crowd chaos can be quantified by calculating the macroscopic and microscopic entropies. Second, we propose an entropy-based anisotropic emotional contagion model (E-AECM) to achieve the nonuniform positive emotional contagion considering the influence of chaos on the propagation rate. Third, we formulate the problem of safety officer placement as an optimization problem to maximize the positive emotional contagion on the chaos individuals. Finally, we implement a simulation system for crowd evacuation to visualize the results of positive emotion contagion. The proposed method can provide guidance for emergency response management.

In a crisis situation, negative emotions often spread among the crowd, and they have adverse impacts on human decisions, resulting in stampedes and crushes. Safety officers are often dispatched to scenes of emergencies because their positive emotions can calm the crowd down and avoid serious accidents. However, how to utilize the positive emotional contagion to maximize the “calm-down” effect remains a challenging problem in crowd evacuation. In this paper, we present an approach for optimizing positive emotional contagion in crowd evacuation. First, a computational model of positive emotional contagion is proposed to describe how safety officers calm a crowd down. To capture important influential factors for positive emotional contagion, such as the trust relationships among the individuals involved in a crisis situation and the variations of emotional contagion speed, we construct a trust-based emotional contagion network (Trust-ECN) and a heterogeneous emotional contagion speed computation model (HECS-CM). Based on these models, the emotional contagion process can be analyzed in a parametric way, and the infection probability for each individual in a given time window can be computed analytically with a continuous-time Markov chain (CTMC). Second, a maximization problem of emotional contagion is formulated. Since this optimization problem is NP-hard, an artificial bee colony optimized emotional contagion (ABCEC) algorithm is used to solve for the optimal positions of safety officers. We demonstrate the effectiveness of our method on both synthetic and real-world data at different scales. Finally, we implement a crowd simulation system to visualize the results of our theoretical analysis in a graphical manner. The proposed method can provide guidance for emergency response management.

Summary Small social groups based on kinship or friendships are ubiquitous in human crowds. Therefore, the effect of social groups on crowd evacuations and that of crowd evacuations on social groups must be investigated. To simulate the group phenomenon when an emergency occurs, we propose an improved social force model that takes into account the social group relationship among the population, and based on our proposed model, a novel grouping algorithm predicated on non‐uniform binary grid partitioning is put forward. The approach initially maps the individuals into the plane space, and then it adopts top‐down binary grid partitioning iteratively until the divided grid contains only the individuals with relations; then, the values of the relation and density of the non‐empty grid cells are calculated, and the grids are sorted according to these values. After sorting, selecting, merging, and forming the core grids, the other grids are merged to the core grids. We have compared the algorithm with the hierarchical classification algorithm and the grid‐based algorithm. The results show that the accuracy, speed, and scalability are all advantages. We also establish a simulation platform to illustrate the proposed grouping algorithm and the improved social force model for crowd evacuation simulation.

Summary Fluid simulation is an important research topic since it is widely used in virtual training, education, computer games, and digital entertainment. Generating fluid simulation results with high degree of visual realism is challenging in these applications. We present a visualization approach for fluid simulation in this paper. Our method allows students to understand the complex natural phenomena so as to solve the fluid engineering problems intuitively. In this study, an SPH (Smoothed Particle Hydrodynamics)‐based multi‐resolution method is proposed for fluid simulation. The method can keep the accuracy while improve the efficiency significantly. We construct a multi‐resolution fluid model from bottom to top by combining fluid computation and surface reconstruction. We first compute multi‐resolution particles with a novel adaptive SPH model in the bottom layer. Next, surface particles are extracted and pre‐processed to provide samples for surface extraction. Then, the multi‐resolution surface reconstruction model is built in the top layer. Meshes from different reconstruction resolutions are merged to obtain the final fluid surface. Experimental results show that our approach can effectively represent multi‐scale fluid details and efficiently produce visual‐pleasing fluid simulation results.

ABSTRACT We present a rigid‐motion‐inspired method for animating liquid characters in this paper. Our method allows an animator to control the motion of liquid characters with motion capture data that is widely used in rigid body animation. It animates the most visual interesting part of liquid character, that is, to preserve character's shape as well as produce enough liquid details. To this end, we build a two‐layer model to represent the character by two coaxial layers: the rigid kernel and the liquid shell. Different control paradigms are used for the two layers instead of applying homogeneous force that is common in previous approaches. By embedding the control algorithm to the Navier–Stokes equations, we compute the fluid velocity that drives the motion of the liquid character. Results show that the method is easy and intuitive to use while incurring little additional cost.Copyright © 2013 John Wiley & Sons, Ltd.