近年來,3D電腦動畫的應用如雨後春筍,在電影、電玩、廣告方面已然是不可或缺的一項元素,其中,以3D動畫進行群體特效的模擬更是造就浩大場面的重要功臣之一,然而,對於大量角色的動作編輯一直是項歷日曠久的製作過程,因此在動畫領域有所謂「行為動畫」(Behavior Animation)的研究,目的便是用以縮短大量角色的動作編輯過程,本文亦繼承此一目的,旨在提出一套基於情緒與群眾理論的行為模型,作為群體模擬時之行為決策依據。 本研究所提出之群體行為模型之設計,乃從情緒心理學及群眾學相關理論進行分析整理,進而架構出一模擬群體行為之模型;我們除了視情緒為影響群眾行為的主要因素,更同時加入理性認知方面的考量;由於群眾形式包含多種類型,本文根據群眾學之群聚理論(Convergence Theory),選擇最常見的「偶集群眾」類型(Casual Crowd)作為主要情境模擬,同時系統架構結合現存三個相關模型理論之優點,包括:產生情緒的OCC Model、行為決策的BDI Model以及心理學的情緒評估理論(Emotion Appraisal Theory),我們亦將角色加入了個性差異之因素,並設定相關參數以供調整,以便應用於不同情境。 在模型設計完成後,我們以Maya平台進行實際的群體行為動畫模擬,實驗結果顯示,此基於情緒心理學與群眾理論的模型設計,確實有助於簡化現實中複雜的人群互動且較符合實際人群聚集活動的行為表現。
Recently there has been an increased interest in using a large number of virtual characters in 3D Animation. However, editing an animation with a large number of different characters is a time-consuming procedure. The technique of “Behavior Animation” provides some automatic solutions to help animators to deal with various types of group behavior. In this study we confine ourselves to the group behavior with emotional convergence. First we propose a model of “Casual Crowd” according to convergence theory of crowd psychology. Our study focuses on how emotion affects on group behavior. We not only consider the influence of each individual’s emotion but also take into account the global factors such as the distribution of convergence centers in the environment and their corresponding crowd sizes. Following the line of existing models such as: “BDI Model”, “OCC Model” and “Emotional Appraisal Theory”, we propose a novel computation model for group behavior. The experiment results show that our model can effectively synthesize very natural animations of group behavior with emotional convergence.