Gerhard Mentges

Universität Hamburg, Hamburg University of Technology

Papers

Total Citations

H-Index

About

Gerhard Mentges is a researcher whose work sits at the intersection of computer vision, robotics, and 3D reconstruction, with a particular focus on leveraging line segments—rather than the more commonly used points—for spatial understanding. His major contributions center on developing novel multi-view algorithms for surface reconstruction from matched line segments, a challenging problem because lines provide less direct geometric constraints than points. In his most cited work, "Maximally informative surface reconstruction from lines" (2014, 4 citations), Mentges proposed a method that forms plane hypotheses from non-collinear, sufficiently coplanar segment pairs, enabling robust surface generation for applications in robotic mapping and image-based rendering. He extended this line of research in "Surface reconstruction from image space adjacency of lines using breadth-first plane search" (2016, 3 citations), where he integrated his reconstruction pipeline with Line-SLAM to project 3D segments into keyframes, improving efficiency and accuracy. Though his citation counts are modest, Mentges’ work is notable for tackling a fundamental, underexplored problem in 3D vision—reconstructing surfaces from lines—paving the way for more robust mapping systems in environments where traditional point-based methods fail.