Mono3D++: Monocular 3D Vehicle Detection with Two-Scale 3D Hypotheses and Task Priors
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.
Cite
Text
He and Soatto. "Mono3D++: Monocular 3D Vehicle Detection with Two-Scale 3D Hypotheses and Task Priors." AAAI Conference on Artificial Intelligence, 2019. doi:10.1609/AAAI.V33I01.33018409Markdown
[He and Soatto. "Mono3D++: Monocular 3D Vehicle Detection with Two-Scale 3D Hypotheses and Task Priors." AAAI Conference on Artificial Intelligence, 2019.](https://mlanthology.org/aaai/2019/he2019aaai-mono/) doi:10.1609/AAAI.V33I01.33018409BibTeX
@inproceedings{he2019aaai-mono,
title = {{Mono3D++: Monocular 3D Vehicle Detection with Two-Scale 3D Hypotheses and Task Priors}},
author = {He, Tong and Soatto, Stefano},
booktitle = {AAAI Conference on Artificial Intelligence},
year = {2019},
pages = {8409-8416},
doi = {10.1609/AAAI.V33I01.33018409},
url = {https://mlanthology.org/aaai/2019/he2019aaai-mono/}
}