Computer Engineering and Applications ›› 2012, Vol. 48 ›› Issue (19): 209-214.

Previous Articles     Next Articles

Research of virtual driver’s visual attention model

REN Xiaoming1, XUE Qing1, YANG Ji2   

  1. 1.Department of Equipment Commanding and Management, Academy of Armored Force Engineering, Beijing 100072, China
    2.Department of Traffic and Transportation, Southwest Jiaotong University, Chengdu 611731, China
  • Online:2012-07-01 Published:2012-06-27

虚拟驾驶员视觉注意力模型研究

任晓明1,薛  青1,杨  冀2   

  1. 1.装甲兵工程学院 装备指挥与管理系,北京 100072
    2.西南交通大学 交通运输系,成都 611731

Abstract: For improving the fidelity of driver’s perceptual simulation, research of attention model is the key of driving behavior. This paper Introduces the cognitive models of attention, establishes the virtual driver’s visual attention model based on the double processing theory, and divides the visual attention to the pre-attention and attention concentration. In pre-attention, it gets the feature map of figure according to the phase spectrum information to complete the attention procedure from bottom to top. The paper Utilizes the RBF(Radial Basis Function) network to combine the procedures from bottom to top and from top to bottom. In attention concentration procedure, it completes the effects of attention focus. Simulating the visual attention model, the model is validated, and the foundation for virtual driver’s visual perception model is provided.

Key words: attention, visual perception, pre-attention, phase spectrum of figure, attention concentration

摘要: 为提高驾驶员视觉感知行为仿真的逼真性,注意力模型成为驾驶行为研究的重点。介绍了注意力的认知模型,根据注意力的双加工理论,建立了虚拟驾驶员视觉注意力模型,将视觉注意过程分为预注意和集中注意两个阶段,在预注意阶段,基于图像的相位谱信息建立了图像的特征显著图,实现注意的自下而上过程。使用RBF网络实现了预注意阶段的自下而上和自上而下两个过程的结合。在集中注意阶段,实现了注意力的聚焦效应。对视觉注意力模型进行了仿真实验,验证了模型的有效性,为虚拟驾驶员视觉感知行为建模与仿真提供了基础。

关键词: 注意力, 视觉感知, 预注意, 图像相位谱, 集中注意