平面图像立体化研究

将平面图像转换成立体图像很有意义。Hou等人(2002)提出了一种平面图像立体化方法,称为Hou方法。在这个方法中Hou等人使用随机变量等参数控制转换过程,但没有讨论这些参数对立体效果的影响。该文利用Hou方法在标准的计算机监视器上对含有心理深度暗示的平面图像进行了立体化,给出了评价转换后的立体效果的定量指标,并讨论了各参数对立体效果的影响。实验结果表明:当随机变量矩阵中的每一个随机变量都服从同一种均匀分布时,随机变量的取值对立体效果没有多大的影响 当监视器屏幕与观察者的距离为一个合适的值(例如1m)时,图像子块的个数取不同的值对立体效果没有多大的影响。-Three-dimensional image plane Research

The establishment of body image plane image conversion makes sense. Hou et al (2002) proposed a method for three-dimensional image plane, called Hou methods. Hou et random variables and other parameters used to control the conversion process in this method, but did not discuss the effects of these parameters on the three-dimensional effect. In this paper, the use of Hou method on a standard computer monitor on a flat image containing the psychological depth cues were three-dimensional, gives a quantitative index evaluation converted three-dimensional effect, and the influence of various parameters on the three-dimensional effect. The results showed that: when the random variables each of the matrix are subject to a random variable with a uniform distribution, the value of the random variable is not much influence on the three-dimensional effect　monitor screen when the distance between the observer as an appropriate value (for example 1m), the nu