Mono3D++: Monocular 3D Vehicle Detection with Two-Scale 3D Hypotheses and Task Priors

Authors

  • Tong He University of California, Los Angeles
  • Stefano Soatto University of California, Los Angeles

DOI:

https://doi.org/10.1609/aaai.v33i01.33018409

Abstract

We present a method to infer 3D pose and shape of vehicles from a single image. To tackle this ill-posed problem, we optimize two-scale projection consistency between the generated 3D hypotheses and their 2D pseudo-measurements. Specifically, we use a morphable wireframe model to generate a fine-scaled representation of vehicle shape and pose. To reduce its sensitivity to 2D landmarks, we jointly model the 3D bounding box as a coarse representation which improves robustness. We also integrate three task priors, including unsupervised monocular depth, a ground plane constraint as well as vehicle shape priors, with forward projection errors into an overall energy function.

Downloads

Published

2019-07-17

How to Cite

He, T., & Soatto, S. (2019). Mono3D++: Monocular 3D Vehicle Detection with Two-Scale 3D Hypotheses and Task Priors. Proceedings of the AAAI Conference on Artificial Intelligence, 33(01), 8409-8416. https://doi.org/10.1609/aaai.v33i01.33018409

Issue

Section

AAAI Technical Track: Vision