Mapprior: Bird's-eye view perception with generative models
Zhu, Xiyue
Loading…
Permalink
https://hdl.handle.net/2142/121554
Description
Title
Mapprior: Bird's-eye view perception with generative models
Author(s)
Zhu, Xiyue
Issue Date
2023-07-21
Director of Research (if dissertation) or Advisor (if thesis)
Wang, Shenlong
Kindratenko, Volodymyr
Department of Study
Electrical & Computer Eng
Discipline
Electrical & Computer Engr
Degree Granting Institution
University of Illinois at Urbana-Champaign
Degree Name
M.S.
Degree Level
Thesis
Keyword(s)
Bev Perception
Bev Map Segmentation
Generative Models
Vqgan
Autonomous Driving
Language
eng
Abstract
This thesis presents a novel generative methodology in bird’s-eye view map segmentation. Despite tremendous advancements in bird’s-eye view (BEV) perception, existing models fall short in generating realistic and coherent semantic map layouts, and they fail to account for uncertainties arising from partial sensor information (such as occlusion or limited coverage). This thesis introduces MapPrior, a novel BEV perception framework that combines a traditional discriminative BEV perception model with a learned generative model for semantic map layouts. Our MapPrior delivers predictions with better accuracy, realism, and uncertainty awareness. Evaluated on the large-scale nuScenes benchmark, it establishes a new state-of-the-art mean IoU score, with significantly improved MMD and ECE scores, for both cameras- and LiDAR-based BEV perception models. Furthermore, our method can be used to perpetually generate layouts with unconditional sampling.
Use this login method if you
don't
have an
@illinois.edu
email address.
(Oops, I do have one)
IDEALS migrated to a new platform on June 23, 2022. If you created
your account prior to this date, you will have to reset your password
using the forgot-password link below.
