One paper of Robotics and Visual Intelligence (RVI) lab is accepted to IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), which is one of the top journals in AI/CS/EE (IF:16.389)!
We propose a new linear RGB-D simultaneous localization and mapping (L-SLAM) formulation by utilizing planar features of the structured environments. The key idea is to understand a given structured scene and exploit its structural regularities such as the Manhattan and Atlanta worlds. We evaluate L-SLAM on a synthetic dataset and RGB-D benchmarks, demonstrating comparable performance to other state-of-the-art SLAM methods without using expensive nonlinear optimization. We assess the accuracy of L-SLAM on a practical application of augmented reality.
“Linear RGB-D SLAM for Structured Environments” by Kyungdon Joo, Pyojin Kim, Martial Hebert, In So Kweon and Hyoun Jin Kim
Note: This work is a collaboration with Sookmyung Women’s University, Carnegie Mellon University (CMU), KAIST, and SNU.