David Johnson

Papers

Total Citations

156

H-Index

About

David Johnson is a leading researcher in robotic perception, with a focus on enabling autonomous manipulation in complex, unstructured environments. His most impactful work, “SegICP: Integrated deep semantic segmentation and pose estimation” (2017, 151 citations), addresses a critical bottleneck in robotics: the need for fast, reliable object detection and pose estimation in realistic, cluttered scenes. By fusing deep semantic segmentation with iterative closest point registration, Johnson’s SegICP framework dramatically improves both the speed and robustness of robotic perception, a contribution recognized by the manipulation competitions that inspired it. He further advanced this line of research with “SegICP-DSR: Dense Semantic Scene Reconstruction and Registration” (2017), achieving millimeter-level pose accuracy and demonstrating successful object identification in real-time. While this follow-up work has garnered fewer citations, it showcases Johnson’s commitment to pushing the boundaries of precision in dense semantic mapping. His integrated approach—combining deep learning with geometric registration—has become a foundational reference for researchers building next-generation robotic manipulation systems, cementing his reputation as a key innovator in the field.