- 學位論文
群體情緒感染行為之電腦動畫
Computer Animation of Group Behavior with Emotional Contagion
摘要
在動畫與電影中，群體特效之應用已久，它不但能製作出許多大場景，也能為製片人節省不少臨時演員的經費。但由於傳統手工動畫的製程曠日費時，加上電腦科技不斷進步，現今利用電腦技術快速地產生群體行為動畫已愈來愈普遍。近年來無論在電影、動畫或遊戲界，對群體特效的需求皆與日俱增，是故愈來愈多相關研究，開始探討各類型的群體行為之電腦動畫演算法。 本研究旨在以電腦動畫模擬群體行為中，特殊的情緖感染行為現象，並且自心理學領域探索相關知識，了解情緒感染的理論與成因。所謂的情緒感染，乃是個體受他人影響，改變自己的情緒與行為，最終造成群體裡情緒與行為的傳遞現象。James 與Keike於2000年提出他們的情緒感染模型，本研究參考其模型，並結合Jonathan Gratch與Stacy Marsella的情緒模型，提出一組適用於模擬群體情緒感染行為的模型與流程，並且分析模擬中可控制的參數。此外，本研究借助細胞自動機(Cellular Automata)的運算模式，建構出群體中單一個體受其他鄰居個體影響的規則，並且將這些規則以互動式介面之方法呈現，使動畫製作者可利用介面調整參數，以便模擬出更為豐富與多元的群體動畫效果。 在模型建立後，本研究將它實際應用在幾個情境中，並實作於Maya平台上。根據實例所呈現的效果可以發現，本研究所提之模型與流程，的確能夠製作具群體情緒感染行為之動畫效果。此外，我們發現細胞自動機的架構有助於簡化群體中複雜的傳遞行為，而為了使模擬結果不會過於制式化且更逼近現實，我們將心理學相關知識納入考慮，並建立一些合理的變數，以增加系統的多樣性。
並列摘要
Group animation technique has been used by producers of animations and movies for a long time because it can reduce the expenditure of temporary actors. However, making all these group effects by hand takes animators a lot of time and effort. As a result, some animation software try to provide solutions for group behavior animation, and many researchers are working hard to make this kind of simulation more effective and real. We focus on simulating group behavior with emotional contagion for animation. In order to have real results, we use psychology as our base knowledge of how group behave in emotional contagion situations. Emotional contagion is a special phenomenon where people tend to have similar emotion and behaviors as others beside them. In 2000, James and Keike proposed their model of emotional contagion. We combined their model with Gratch and Marsella’s computational emotion model as our new emotional model for simulating group behavior with emotional contagion. Besides, we use cellular automata to build the relationship of agents. By adjusting CA rules, different kinds of results will come out. We also analyze the parameters users could change to have various simulation results. After building our emotional contagion model, we implement some situations in Alias Maya for testing. The testing results show that our model could actually produce animation of group behavior with emotional contagion. By using cellular automata, we reduce the complexion of group simulation. But because of the reasonable parameters we add in the simulation based on psychology, the results can be more realistic.