基于虚拟现实的人机交互视觉目标感知仿真

Visual Goals Perception in Human-Computer Interaction Based on Virtual Reality

人机交互是人与计算机或其它外部设备之间的信息交换过程。在人机交互过程中,由于网络环境复杂以及不稳定等因素,直接影响视觉目标的感知效果。为解决上述问题,提出基于虚拟现实的人机交互视觉目标感知方法。利用虚拟现实技术建立安全的人机交互网络环境,并在此环境中截取视觉目标二维图像,采用Gabor小波变换法,提取视觉目标的特征向量;采用最小二乘支持向量机,建立人机交互视觉目标感知模型,通过粒子群算法确定模型相关参数;将提取的目标特征向量输入该模型中,完成人机交互视觉目标的感知。实验结果表明,上述研究方法的平均感知精度在95%~100%之间,高于其它两种方法的测试结果,研究方法的目标感知速度和感知效果也均更优。

Human-computer interaction is the process of information exchange between people and computers or other external devices.In the process of human-computer interaction,the complexity and instability of network can directly affect the perception effect of visual objects.Therefore,a method of sensing visual objects in the process of human-computer interaction is proposed based on virtual reality.Firstly,virtual reality technology was used to build a safe environment for human-computer interaction network,and then 2D image of visual object was captured in this environment.Secondly,Gabor wavelet transform method was used to extract the feature vector of visual object.Thirdly,least square support vector machine was used to construct a model of sensing visual object in human-computer interaction,and then relevant parameters of the model were determined by particle swarm optimization.Finally,the extracted target feature vectors were input into the model,thus completing the perception of visual objects.Experimental results prove that the average perceptive accuracy of the proposed method is between 95%and 100%,which is higher than the results of the other methods.In addition,its perception speed and perception effect are also better than others.