Zhipeng Bao

Jilin University, Carnegie Mellon University

Papers

Total Citations

H-Index

About

Zhipeng Bao is a researcher whose work bridges classical robotics and modern deep learning, with key contributions in visual simultaneous localization and mapping (VSLAM), visual odometry (VO), and neural scene understanding. His 2021 paper on stereo visual odometry pose correction introduced an unsupervised deep learning framework that overcomes the limitations of classical VO systems, achieving more robust ego-motion estimation for positioning and navigation—a critical advance for autonomous robots. Building on this, his 2023 work on scene-property synthesis with neural radiance fields (NeRF) moves beyond RGB to enable comprehensive 3D scene understanding, integrating geometric and semantic properties through a novel generative approach rather than traditional discriminative models. This work, with its potential to transform robot perception, has already garnered significant attention in the computer vision community. With over 20 citations across his most-cited papers, Bao is establishing himself as a rising voice in the fusion of unsupervised learning and 3D vision, offering elegant solutions to long-standing challenges in autonomous navigation and scene interpretation.