The Role of Camera Convergence in Stereoscopic Video See-through Augmented Reality Displays

Abstract

In the realm of wearable augmented reality (AR) systems, stereoscopic video see-through displays raise issues related to the user's perception of the three-dimensional space. This paper seeks to put forward few considerations regarding the perceptual artefacts common to standard stereoscopic video seethrough displays with fixed camera convergence. Among the possible perceptual artefacts, the most significant one relates to diplopia arising from reduced stereo overlaps and too large screen disparities. Two state-of-the-art solutions are reviewed. The first one suggests a dynamic change, via software, of the virtual camera convergence, whereas the second one suggests a matched hardware/software solution based on a series of predefined focus/vergence configurations. Potentialities and limits of both the solutions are outlined so as to provide the AR community, a yardstick for developing new stereoscopic video see-through systems suitable for different working distances.

Joint effects of depth-aiding augmentations and viewing positions on the quality of experience in augmented telepresence

“…This decrease of accuracy may adversely impact depth estimation and positioning tasks. Cutolo et al [51] suggested that HMD optics may be the cause for spatial perception errors, and that camera convergence improves near-distance 3D object positioning [54]. However, these publications did not address the QoE of keystone artefacts from converged cameras, or the influence of camera FoV.…”

Perspective Preserving Solution for Quasi-Orthoscopic Video See-Through HMDs

“…This value leads to a visual acuity of about half of the average visual acuity in human vision (visual acuity = d r −1 ). The approximated formula of the depth resolution dZ at a distance Z from the observer can be retrieved as follows [18,36,37]:…”

“…This angle of convergence should be established based on assumptions made on the average working distance. In such configuration, for preserving a natural visual perception of the space (i.e., the conditions of orthostereoscopy) and reduce stereoscopic distortions as keystone distortion and depth-plane curvature, theoretically also the two displays should be physically converged of the same angle [11,18]. Yet, this requirement cannot be fulfilled from a practical standpoint and this has implications on the ability of the stereoscopic system to recovering natural depth perception.…”

Ambiguity-Free Optical–Inertial Tracking for Augmented Reality Headsets

“…• the presence of poorly calibrated tracking cameras; • the anthropomorphic geometry of the head-anchored stereo rig [13][14][15] (e.g., the short baseline of the stereo cameras, which should be ≈ to the user's interpupillary distance, and the limited camera focal length that should comply with orthostereoscopic viewing conditions); • the presence of inaccuracies in the feature detection that may lead to numerical instability and tracking ambiguities particularly for those tracking strategies that rely on a reduced number of feature points; • the limited frame rate of the tracking cameras typically mounted over AR headsets (60 Hz at most); • the presence of noise due to head movements affecting the quality of the tracking; • the presence of occlusions on the line-of-sight between the user's wearing the AR headset and the target scene; • the latency typical of purely optical tracking methods that results in misregistration between virtual content and real world in optical see-through (OST) headsets or in delayed perceptions of the reality in video see-through (VST) HMDs [16].…”