Paul Baumstarck

Papers

Total Citations

222

H-Index

About

Paul Baumstarck is a roboticist whose work sits at the intersection of computer vision and autonomous systems, pioneering methods that help machines perceive and interact with the physical world. His research focuses on three key areas: multi-modal sensor fusion, biologically inspired vision architectures, and high-performance computing for real-time object detection. In his seminal 2008 paper, "Integrating Visual and Range Data for Robotic Object Detection" (88 citations), Baumstarck demonstrated that combining standard RGB imagery with depth data from range sensors dramatically improves detection accuracy in cluttered environments—a foundational insight for modern RGB-D perception systems. His 2007 work on "Peripheral-foveal vision for real-time object recognition and tracking" (82 citations) introduced a computational model mimicking the human visual system's fovea-periphery structure, achieving remarkable tracking speeds while maintaining recognition fidelity. Perhaps most presciently, his 2009 paper "Scalable learning for object detection with GPU hardware" (52 citations) was among the first to leverage GPU parallelism for vision tasks, accelerating detection of common indoor objects like mugs and staplers from minutes to milliseconds. Baumstarck's contributions remain highly cited for bridging biological vision principles with practical, computationally efficient robotic systems.