David Lovi

Papers

Total Citations

H-Index

About

David Lovi is a researcher whose work sits at the intersection of computer vision, real-time 3D reconstruction, and robotic teleoperation. His most impactful contribution is his 2011 thesis on "Incremental free-space carving for real-time 3D reconstruction," which has garnered 35 citations. This work tackled the challenging problem of automatically reconstructing general 3D scenes from monocular video in real time, offering a dynamic alternative to the slow, offline batch methods prevalent in multi-view stereo literature. Lovi’s approach enabled practical, on-the-fly modeling without the need for highly calibrated settings. In parallel, he made significant strides in improving human-robot interaction. His 2010 paper on "Predictive display for mobile manipulators in unknown environments" (11 citations) pioneered a vision-based system to compensate for communication latency, allowing operators to teleoperate robots effectively despite delayed visual feedback. This work, along with a related study on predictive displays from computer vision models (4 citations), directly addressed the critical problem of operator dissociation caused by even sub-second delays. Lovi’s research provides foundational techniques for making real-time 3D vision practical and for enhancing the usability of remote robotic systems in unknown environments.