基于改进粒子群算法的体绘制传递函数设计

为降低体绘制过程中人机交互的复杂性,提出一种体绘制传递函数的自动设计方法.该方法把对传递函数的抽象评价转变为对绘制图像的显式评价,然后将传递函数的设计转变为一个多参数优化问题,并使用改进的粒子群算法进行自动寻优.图像的评价使用图像信息熵、差分熵、边界熵和主观评价的融合方法.针对粒子群算法易于陷入局部最优的缺点,结合遗传算法的思想对粒子群算法进行改进.该方法在体绘制应用中,具有更好的全局搜索能力和更高的收敛速度.实验结果表明,在一般体绘制应用中,本文的方法可以在1.0～2.0min内完成传递函数设计,实现用户满意的体绘制效果.

Modified PSO method for automating transfer function designing in volume rendering

To reduce the complexity of human computer interaction in volume rendering, this paper introduces an automated approach for transfer function designing in volume rendering. This approach transfers the abstract evaluation of a transfer function into the explicit evaluation of its rendering image, and then transfers the designing of a transfer function into a multi parameter optimization problem. The image quality is assessed by combining image information entropy, differential entropy, boundary entropy, and humans subjective evaluation. Optimizing process utilizes an improved PSO (Particle Swarm Optimization) method which is strengthened by a genetic algorithm to avoid falling into the local optimum. The results of tests show that this modified PSO algorithm has a better global searching ability and efficiency in the application of volume rendering. The experimental results demonstrate that the proposed approach is able to design high quality transfer functions according to the humans perspective in 1 2 minutes for common cases．