Motion-Aware Ghosted Views for Single Layer Occlusions in Augmented Reality

摘要

Functional realism focuses on helping users with tasks execution through an enhanced perception of the augmented scene. This work applies a common visualization technique, Ghosting, to improve depth perception in Augmented Reality scenes. Computer Vision and Image Processing techniques are used to extract natural features from a real scene, which will guide the assignment of transparency to each pixel of the virtual object, and provide the ghosting effect while blending the virtual object into the real scene. A moving object in a real scene catches users' attention. So, it is expected that natural and important visual information of the scene does not get occluded when the moving object passes over it. Because of that, the main contribution of this work is the inclusion of a motion detection technique to the scene feature analysis step of the Ghosting technique pipeline. A qualitative evaluation of the results achieved shows that the case studies of this work, in indoor and outdoor environments, using the proposed technique led to a better depth perception of the augmented scene, preserving the most relevant information for visual attention.