ABSTRACT
Despite advanced developments in crowd simulation, little attention has been given to the questions of evaluation and comparison among crowds. A big challenge when comparing crowds is the definition of which metric(s) should be used to compare and characterize crowds. Since seminal papers pioneered by Reynolds back in 1987 [1], who simulated flocks of birds and schools of fishes, the evaluation of crowd simulation results is mostly based on visual inspection. This chapter presents some real case studies that have been compared with simulation data. In addition, a quantitative method to compare crowds is also presented.
