Virtual Reality Lensing for Surface Approximation in Feature-driven DVR

Abstract: We present a novel lens technique to support the identification of heterogeneous features in direct volume rendering (DVR) visualizations. In contrast to data-centric transfer function (TF) design, our image-driven approach enables users to specify target features directly within the visualization using deformable quadric surfaces. The lens leverages quadrics for their expressive yet simple parametrization, enabling users to sculpt feature approximations by composing multiple quadric lenses. By doing so, the lens offers greater versatility than traditional rigid-shape lenses for selecting and bringing into focus features with irregular geometry. We discuss the lens visualization and interaction design, advocating for bimanual spatial virtual reality (VR) input for reducing cognitive and physical strain. We also report findings from a pilot qualitative evaluation with a domain specialist using a public asteroid impact dataset. These insights not only shed light on the benefits and pitfalls of using deformable lenses but also suggest directions for future research.