Stereoscopic static depth perception of enclosed 3D objects

Abstract

Figure 1: Lateral view of the positions where the enclosed object (simulated tumor) is located within the exterior object (brain). Back position (left), center position(middle), front position(right). Two rendering methods were used with different opacity values. AbstractDepth perception of semi-transparent virtual objects and the visualization of their spatial layout are crucial in many applications, in particular medical applications. Depth cues for opaque objects have been extensively studied, but this is not the case for stereoscopic semi-transparent objects, in particular in the case when one 3D object is enclosed within a larger exterior object.In this work we explored different stereoscopic rendering methods to analyze their impact on depth perception accuracy of an enclosed 3D object. Two experiments were performed: the first tested the hypotheses that depth perception is dependent on the color blending of objects (opacity -alpha) for each rendering method and that one of two rendering methods used is superior. The second experiment was performed to corroborate the results of the first experiment and to test an extra hypothesis: is depth perception improved if an auxiliary object that provides a relationship between the enclosed object and the exterior is used?The first rendering method used is simple alpha blending with Blinn-Phong shading model, where a segmented brain (exterior object) and a synthetic tumor (enclosed object) were blended. The second rendering method also uses Blinn-Phong, but the shading was modified to preserve silhouettes and to provide an illustrative rendering. Comparing both rendering methods, the brighter regions of the first rendering method will become more transparent in the second rendering method, thus preserving silhouette areas.The results show that depth perception accuracy of an enclosed object rendered with a stereoscopic system is dependent on opacity for some rendering methods (simple alpha blending), but this effect is less pronounced than the dependence on object position in relation to the exterior object. The illustrative rendering method is less dependent on opacity. The different rendering methods also perform slightly differently; an illustrative rendering method was * superior and the use of an auxiliary object seems to facilitate depth perception.

Effects of shading model and opacity on depth perception in optical see‐through augmented reality

“…Shading models used to render a virtual object has been proven to influence the depth perception 40–43 in computer graphics researches. Other research on the effect of shading focused on the shape perception 2,3 and spatial relationship between the virtual 3D objects 1,5,6 . These studies showed the importance of specular highlights for conveying depth information both along the surface of a virtual object and about the relative positioning of objects 44 .…”

“…The researches on opacity also focus on the depth perception in direct volume rendering (DVR) and the relationship between the virtual objects, as well as the spatial layout of occluded objects. The perceived depth order of semitransparent virtual objects rendered by means of volume rendering has been fully discussed in computer graphics 5,6 . Changing the opacity of the virtual objects in a stereo display can improve the depth perception of two enclosed 3D objects 5 .…”

“…The perceived depth order of semitransparent virtual objects rendered by means of volume rendering has been fully discussed in computer graphics 5,6 . Changing the opacity of the virtual objects in a stereo display can improve the depth perception of two enclosed 3D objects 5 . X‐ray visualization techniques created a semitransparent object overlaid on the ground plane 7,8 and helped with the spatial understanding.…”

“…Our current work focuses on the accuracy of users' depth perception of virtual objects under different rendering methods: shading models and transparency (opacity‐alpha). Shading provides important information about the surface shape to distinguish between convexities and concavities, which is widely researched in shape perception 2,3,4 or the position relationship 1,5,6 between two virtual objects in the virtual environment (VE). The results of these investigations showed that shading was a strong visual cue for users when perceiving a virtual object.…”

“…Occlusion only gives information about the order of the objects but cannot give more information about the depth. X‐ray visualization techniques 7–10 and color blending rendering method 5 are wildly used to create a transparency layout for spatial understanding in AR. If the virtual objects are transparent, users can see not only the virtual objects but also the real environment.…”