Infocarve: A Framework for Volume Visualization on Commodity Augmented Reality Displays

摘要

We present a framework for interactive real-time visualization of three-dimensional volume data on commodity augmented reality (AR) displays. In particular, we address the problem of seamlessly blending internal anatomy data sets with real-world objects. One key challenge, particularly relevant to this scenario, is conveying the correct sense of relative depths of virtual and real world objects. To address this issue, we exploit information captured by a depth sensor to build a mask which is used as a weighting parameter to correctly combine rendered volume images with real imagery in a depth-preserving manner. Results obtained on prototype AR hardware devices indicate improvements to relative depth perception. Furthermore, we address performance challenges and provide solutions to ensure that the framework is applicable to a range of different AR devices, many of whom have limited computational and graphical rendering capabilities.