Exploring Legibility of Augmented Reality X-Ray
Exploring Legibility of Augmented Reality X-Ray
复制标题
探索增强现实 X 射线的清晰度
DOI：
10.1007/s11042-015-2954-1
复制
发表时间：
2015
影响因子：
3.6
通讯作者：
H.
中科院分区：
文献类型：
--
作者：
Santos;M.;Souza;I.;Yamamoto;G.;Taketomi;T.;Sandor;C.;and Kato;H.
虚拟对象可以使用增强现实（AR）在真实的对象内部可视化。这种可视化被称为AR X射线，因为它给人的印象是透过真实的物体。在标准AR中，虚拟信息覆盖在真实的世界之上。为了将虚拟对象定位在对象内部，AR X射线需要用真实的对象的视觉上重要的区域部分地遮挡虚拟对象。实际上，与完全未被遮挡时相比，虚拟对象变得较难辨认。易读性是AR X射线各种应用的重要考虑因素。在这项研究中，我们探讨了两种实现的AR X射线，即基于边缘和基于显着性的易读性。在我们的第一个实验中，我们探索了可容忍的遮挡量，以舒适地区分小的虚拟对象。在我们的第二个实验中，我们比较了基于边缘和基于显着性的AR X射线方法时，可视化虚拟对象内的各种真实的对象。此外，我们对这两种方法的易读性进行了基准测试。从我们的实验中，我们观察到用户对这两种方法的适量遮挡线索有不同的偏好。基于边缘和基于显著性的方法产生的部分遮挡需要根据光照条件和遮挡对象的纹理复杂度进行调整。在大多数情况下，用户使用基于显著性的AR X射线比使用基于边缘的AR X射线更快地识别对象。这项研究的见解可以直接应用于AR X射线应用的开发。
Virtual objects can be visualized inside real objects using augmented reality (AR). This visualization is called AR X-ray because it gives the impression of seeing through the real object. In standard AR, virtual information is overlaid on top of the real world. To position a virtual object inside an object, AR X-ray requires partially occluding the virtual object with visually important regions of the real object. In effect, the virtual object becomes less legible compared to when it is completely unoccluded. Legibility is an important consideration for various applications of AR X-ray. In this research, we explored legibility in two implementations of AR X-ray, namely, edge-based and saliency-based. In our first experiment, we explored on the tolerable amounts of occlusion to comfortably distinguish small virtual objects. In our second experiment, we compared edge-based and saliency-based AR X-ray methods when visualizing virtual objects inside various real objects. Moreover, we benchmarked the legibility of these two methods against alpha blending. From our experiments, we observed that users have varied preferences for proper amounts of occlusion cues for both methods. The partial occlusions generated by the edge-based and saliency-based methods need to be adjusted depending on the lighting condition and the texture complexity of the occluding object. In most cases, users identify objects faster with saliency-based AR X-ray than with edge-based AR X-ray. Insights from this research can be directly applied to the development of AR X-ray applications.