neural computer models
social cultural sphere
The article describes the possibilities of using neuro models of massive social and cultural scenes in the field (especially in movies). In this paper, the focus is on computer simulation of crowd scenes in the sociocultural sphere, which defines the basic requirements for her - visual plausibility to the end viewer. To model the mass of social and cultural scene in the area of the main thing is its manageability and predictability of joint actions of virtual characters as a single mass. In this article the idea of the phase structure of the masses (the crowd). The term «phase» we mean the kind of steady state mass (the crowd), which is described in the classification of the masses. Indeed, as Article task is to create a model of the mass scenes with a high degree of ordering. Visually, the crowd may look orderly, but in reality it is subject to strict rules and laws. In this paper, to simulate a large number of virtual characters used multiagent approach. He proved to be effective for solving the problem of simulating a large number of virtual characters. The main problem is the management of individual agents and crowd control at large. In our case it comes to the classification problem. You must choose the action that the agent must make in time for the recruitment of a changing environment. A basic method was chosen artificial neural network (NN). But despite its flexibility, in itself a neural network has a weak formalized management and labor often in training (ranging from the complexity of the problem). To address this shortcoming as control functions in a neural network to use the rules of fuzzy logic. Thus, by setting rules based on fuzzy logic is very simple and effective «training of artificial neural network with a teacher». Most applications do not require mass scenes photorealistic visualization, doing graphics primitives for primary interpretation of results. At the same socio-cultural sphere, this approach to visualization is inadequate, because model must have a high degree of likelihood, estimated spectator. Using the results obtained in this work the author began to develop computer models of crowd scenes for use in digital three-dimensional reconstructions of monuments of culture on the project vizerra.