Vinícius J. Cassol

Towards a quantitative approach for comparing crowds

In this paper, we propose a new model to quantitatively compare global flow characteristics of two crowds. The proposed approach explores a 4‐D histogram that contains information on the local velocity (speed and orientation) of each spatial position, and the comparison is made using histogram distances. The 4‐D histogram also allows the comparison of specific characteristics, such as distribution of orientations only, speed only, relative spatial occupancy only, and combinations of such features. Experimental results indicate that the proposed quantitative metric correlates with visual inspection. Copyright

Evaluating perceived trust from procedurally animated gaze

Adventure role playing games (RPGs) provide players with increasingly expansive worlds, compelling storylines, and meaningful fictional character interactions. Despite the fast-growing richness of these worlds, the majority of interactions between the player and non-player characters (NPCs) still remain scripted. In this paper we propose using an NPCs animations to reflect how they feel towards the player and as a proof of concept, investigate the potential for a straightforward gaze model to convey trust. Through two perceptual experiments, we find that viewers can distinguish between high and low trust animations, that viewers associate the gaze differences specifically with trust and not with an unrelated attitude (aggression), and that the effect can hold for different facial expressions and scene contexts, even when viewed by participants for a short (five second) clip length. With an additional experiment, we explore the extent that trust is uniquely conveyed over other attitudes associated with gaze, such as interest, unfriendliness, and admiration.

From their environment to their behavior: a procedural approach to model groups of virtual agents

Simulation of everyday situations from real life can be a very useful tool in entertainment applications and training systems. Such applications, as games or computer animated movies usually need to provide virtual environments populated with virtual autonomous agents. Commonly, the agents need to be able to evolve in their environment, avoiding collision with each other and obstacles, besides interacting with other characters in order to provide realistic simulations. We present a model to simulate coherent group behaviors based on procedural modeling and semantic environments. Our main focus is virtual environments and agents, present in the background of games or movies generated with few/without user intervention.

Procedural Hair Generation

This paper presents a hair generation technique for cartoon and Anime characters. The main goal of this work is to provide a procedural generation technique, robust enough in order to create different types of hairs. We propose a parametric model that can produce diversity of obtained results through parameters variation. The model is organized in three steps: firstly, the strands generation, followed by positioning them in the head of the characters. Finally, the third step is responsible for providing cartoon rendering, including common characteristics in cartoon shading, as the black edges and solid colors. Visual inspection of obtained results indicate that generated hairs are coherent with expected in terms of visual aspects.

Detection of Global and Local Motion Changes in Human Crowds

Crowds arise in a variety of situations, such as public concerts and sporting matches. In typical conditions, the crowd moves in an orderly manner, but panic situations may lead to catastrophic results. We propose a computer vision method to identify motion pattern changes in human crowds that can be related to an unusual event. The proposed approach can identify global changes, by evaluating 2D motion histograms in time, and also local effects, by identifying clusters that present similar spatial locations and velocity vectors. The method is tested both on publicly available data sets involving crowded scenarios and on synthetic data produced by a crowd simulation algorithm, which allows the creation of controlled environments with known motion patterns that are particularly suitable for multicamera scenarios.

Evaluating and Optimizing Evacuation Plans for Crowd Egress

Evacuation planning is an important and difficult task in building design. The proposed framework can identify optimal evacuation plans using decision points, which control the ratio of agents that select a particular route at a specific spatial location. The authors optimize these ratios to achieve the best evacuation based on a quantitatively validated metric for evacuation performance. This metric captures many of the important aspects of an evacuation: total evacuation time, average evacuation time, agent speed, and local agent density. The proposed approach was validated using a night club model that incorporates real data from an actual evacuation.

An Experience-Based Approach to Simulate Virtual Crowd Behaviors Under the Influence of Alcohol

Different models and tools are available in order to simulate homogeneous crowds taking into account different parameters. In this paper we present a framework to simulate heterogeneous crowds where we emphasize how the presence of alcohol into the simulated agents can make influence on the crowd behavior. Despite the obvious difficulty to evaluate results, this paper proposes a metric to compare the alcohol impact in crowd simulations and also taking as reference a real life scenario.

Simulation of Autonomous Agents using Terrain Reasoning

The representation of a real scenario through a 3D virtual environment populated by autonomous agents has several applications, such as games or animation movies. Such representation can be a difficult and complex task, since natural scenarios are composed by forms with a lot of details that should be reproduced realistically in the virtual world. Moreover, virtual people should understand and consider the virtual world for navigation similarly to real people. In this paper we describe a model to provide motion of groups of characters based on terrain reasoning in 3D virtual worlds. The model starts with the terrain generation process using height maps in a pseudo-infinite world, followed by the simulation of characters movement generated through an space colonization algorithm. Finally, the visualization of characters motion in the 3D terrain can be produced.

CrowdVis: a framework for real time crowd visualization

Crowd visualization is present mostly in digital games and computer animated movies, but are also observed in simulations and virtual reality applications. In crowd simulations we should represent the behavior of agents given different scenarios, and also, such simulations can be provided by different softwares and tools. This paper presents a framework for real time crowd visualization, that no programming knowledge and modeling skills are required from the users. Our main goal is to be able to visualize previously created crowd simulations in real time, combining rendering techniques and providing easy support for managing the scene and the virtual humans.

Crowd Analysis Based on Computer Vision

When evaluating or simulating crowd egress situations, there are several parameters that should be taken into account. For instance, the initial distribution of the people and/or local densities are important to assess possible hazardous events; tracking people or detecting main flows can be very useful to identify main escape routes or bottlenecks; finally, crowd behavior understanding methods can be used to identify collective behavior (e.g., groups) or unusual/abnormal events.