User:Hm356

Hasan Mehmedagic

Computer Graphics /198:428/

Fall 2016

Prof. Mubbasir Kapadia

This user page was made as an assignment for CS:428 - Computer Graphics at Rutgers University.

History and Motivation Behind Crowd Analysis and Management
With the rise of population growth and worldwide urbanization, crowd analysis and management has become a more critical role, leading to some technical and social disciplines to pay heavy attention to it. Population growth has led to many global security and monitoring issues. The more crowded an area or environment are, the harder it is to distinguish individuals, and the more difficult it is to manage them. Providing safety in public areas has also become a major concern. By better understanding the procedures and research behind this phenomenon, we can prepare proper management of people in crowded areas that can lead to safety in the case that a need for an evacuation arises.

Theory of Crowd Analysis
Being critical about the rising population has led to discussions about security and monitoring complications. When the magnitude of crowds increases, the need for more security and monitoring rises with it. Public facilities need a way to determine ways to safely manage these crowds that fluctuates rapidly throughout the day. Significant research has been done to understand the way crowds moves in order to predict where areas of conflicts may occur. This research is done by analyzing data from crowds, and then proceeding to create models of similar situations using software. Many models that simulate crowd behavior exists, with some stating "macroscopic models like network-based models or fluid-dynamics models as well as microscopic models like e.g. the Social Force Model or Cellular Automata."

Surveillance systems have begun to integrate computer vision algorithms, however, these systems have a weakness, in which they are inefficient at attempting process dense crowds. Object detection, tracking, and even event detection degrades significantly as the density of a crowd in a scene increases. Tracking an individual in a dense crowd is a complicated task due to factors largely based around interactions between each individual within the crowd. To counteract some of these factors, such as tracking algorithms (which uses collective patterns), are learned from specific scenes that are likely locations of individual agents. By continuing to apply this integration, it has shown improved accuracy in tracking individuals.

Examples of Real World Application of Crowd Analysis
There are many ways which crowd analysis has been used in real world applications today.

Studying the crowd as a phenomenon
Seed Projects at Stanford University developed a computational model incorporating human behavior in environments with emergency exits.

Micro-simulations based on the impact of police or military forces on crowds is being developed by The Crowd-MAGS Project, which is funded by GEOIDE and the Canadian Network of Centers of Excellence in Geometrics.

Studying the crowd quantitatively
The EPSRC from the UK funded projects which measured possibly dangerous situations based on crowd motion and density.

Some of the projects funded by the EU is PRISMATICA and ADVISOR, which study and research the management of public transportation routes by using CCTV cameras

Theory of Crowd Management
The improper management of crowds can lead to disaster. Access to public facilities as well as safety goals for pedestrians can be met when proper management of pedestrians is put into effect. Large public gatherings are held all over the world for events, such as sports, political events, protests, religious events, parades and holidays. These events are held in public facilities where crowd stampede injuries and fatalities may be caused due to inadequate crowd management. Systems which are crafted to support egress for massive crowds are a necessity in case of emergency. Without these systems in play, situations can arise on a large scale whereby injury or even death may occur to both pedestrians and bystanders. Techniques, which determine optimal paths to safe zones for pedestrians, have been introduced.

Examples of Real World Application of Crowd Management
Oasys Software has developed MassMotion, which is a pedestrian simulation and crowd analysis tool that helps design teams develop crowd management solutions by testing different scenarios.