Point cloud visibility is a crucial attribute for 3D tasks as it links the visible object points to a given viewpoint. In this paper, we address the problem of point cloud visibility for monocular vehicle 6D pose estimation. To this end, a network, dubbed Mono6D++, is introduced which jointly predicts vehicle poses and the associated points visibility. Our method mainly consists of: 1) a multi-model feature extraction module and 2) a fusion unit for learning the pose- and visibility-specific representations. Consequently, the proposed method significantly outperforms the baseline approaches. Mono6D++ is capable of handling heavily occluded, truncated and/or appearance-ambiguous vehicles.
Yangxintong, L, Ducastel, O, Royen, RD & Munteanu, A 2023, Mono6D++: Learning Point Cloud Visibility for 3D Prior-based Vehicle 6D Pose Estimation. in 2023 11th European Workshop on Visual Information Processing, EUVIP 2023 - Proceedings. Proceedings - European Workshop on Visual Information Processing, EUVIP, IEEE, pp. 1-6, 2023 11th European Workshop on Visual
Information Processing (EUVIP), Gjovik, Norway, 11/09/23. https://doi.org/10.1109/EUVIP58404.2023.10323075
Yangxintong, L., Ducastel, O., Royen, R. D., & Munteanu, A. (2023). Mono6D++: Learning Point Cloud Visibility for 3D Prior-based Vehicle 6D Pose Estimation. In 2023 11th European Workshop on Visual Information Processing, EUVIP 2023 - Proceedings (pp. 1-6). (Proceedings - European Workshop on Visual Information Processing, EUVIP). IEEE. https://doi.org/10.1109/EUVIP58404.2023.10323075
@inproceedings{558ad4361a1b4df49af79a355853c1ce,
title = "Mono6D++: Learning Point Cloud Visibility for 3D Prior-based Vehicle 6D Pose Estimation",
abstract = "Point cloud visibility is a crucial attribute for 3D tasks as it links the visible object points to a given viewpoint. In this paper, we address the problem of point cloud visibility for monocular vehicle 6D pose estimation. To this end, a network, dubbed Mono6D++, is introduced which jointly predicts vehicle poses and the associated points visibility. Our method mainly consists of: 1) a multi-model feature extraction module and 2) a fusion unit for learning the pose- and visibility-specific representations. Consequently, the proposed method significantly outperforms the baseline approaches. Mono6D++ is capable of handling heavily occluded, truncated and/or appearance-ambiguous vehicles.",
author = "Lyu Yangxintong and Olivier Ducastel and Royen, {Remco Donovan} and Adrian Munteanu",
note = "Funding Information: V. ACKNOWLEDGEMENT The authors would like to thank for the financial support provided by Innoviris (TORRES, SPECTRE) and by the Fonds Wetenschappelijk Onderzoek (FWO) - 1S89420N. Publisher Copyright: {\textcopyright} 2023 IEEE.; 2023 11th European Workshop on Visual<br/>Information Processing (EUVIP), EUVIP ; Conference date: 11-09-2023 Through 14-09-2023",
year = "2023",
month = sep,
day = "11",
doi = "10.1109/EUVIP58404.2023.10323075",
language = "English",
series = "Proceedings - European Workshop on Visual Information Processing, EUVIP",
publisher = "IEEE",
pages = "1--6",
booktitle = "2023 11th European Workshop on Visual Information Processing, EUVIP 2023 - Proceedings",
}