SLAMANTIC - Leveraging Semantics to Improve VSLAM in Dynamic Environments

Matthias Schorghuber, Daniel Steininger, Yohann Cabon, Martin Humenberger, Margrit Gelautz; Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2019, pp. 0-0

In this paper, we tackle the challenge for VSLAM of handling non-static environments. We propose to include semantic information obtained by deep learning methods in the traditional geometric pipeline. Specifically, we compute a confidence measure for each map point as a function of its semantic class (car, person, building, etc.) and its detection consistency over time. The confidence is then applied to guide the usage of each point in the mapping and localization stage. Points with high confidence are used to verify points with low confidence in order to select the final set of points for pose computation and mapping. Furthermore, we can handle map points whose state may change between static and dynamic (a car can be parked or in motion). Evaluating our method on public datasets, we show that it can successfully solve challenging situations in dynamic environments which cause state-of-the-art baseline VSLAM algorithms to fail and that it maintains performance on static scenes. Code is available at github.com/mthz/slamantic