Abstract

Augmented reality (AR) is one of the essential technologies for the fourth industrial revolution. In AR, virtual objects are embedded in the physical reality and observers can freely interact with virtual objects as well as physical objects. In order to create a realistic and immersive AR environment, three-dimensional (3D) depth perception of virtual objects must be errorless or at least with very little error. However, unlike virtual reality, 3D depth perception of virtual objects in AR has been less studied and the literature shows mixed results as to whether 3D depth perception of virtual objects is overestimated or underestimated. We reviewed 3D depth perception of virtual objects in AR. We discuss in depth how the 3D cues that the visual system uses to calculate or extract 3D egocentric distance (the distance between an observer and an object) may differ in AR. Specifically, binocular disparity and vergence might be the most important cues that cause 3D depth perception errors in AR. We introduce a new display technology, which might provide more immersive AR by resolving some problems in AR such as vergence-accommodation mismatch.