【摘要】 在三维体数据的可视化中,信息的获取不仅与传输函数的设置有关,而且与用户观察的视点有关。提出一种新的应用直接体绘制来可视化体数据时最佳视点的选择算法,通过计算梯度模在图像上的分布,寻找梯度模分布最为均匀且投影面积最大的视点方向,同时,在视域球上离散采样得到初始视点墒分布后,使用自适应步长的梯度下降法来优化全局最佳视点方向。实验结果证实了所提出方法能够有效地找到最佳视点。更多还原

【Abstract】 In the visualization of 3D volume data,the information gained is not only dependent on the setting of transfer function,but also on the viewpoint of users to project.This paper presents a new algorithm to select the optimal viewpoints when visualizing volume data using direct volume rendering,which searches the viewpoint direction in most even gradient magnitude distribution with a largest projection area through the calculation of the distribution of gradient magnitude on the image.After a coarse view entropy distribution is obtained from discretized sampling on the view sphere,a gradient steepest ascent method with adaptive step size is used to optimize the global optimal viewpoint.Several examples of volume data experiment show that the method can effectively find the optimal viewpoint.更多还原