Scipediacontent: Scipediacontent moved page Draft Content 759917905 to Ross et al 2020a

2021-02-03T15:25:57Z

Scipediacontent moved page Draft Content 759917905 to Ross et al 2020a

Scipediacontent: Created page with " == Abstract == We propose DOPS, a fast single-stage 3D object detection method for LIDAR data. Previous methods often make domain-specific design decisions, for example proj..."

2021-02-03T15:25:54Z

Created page with " == Abstract == We propose DOPS, a fast single-stage 3D object detection method for LIDAR data. Previous methods often make domain-specific design decisions, for example proj..."

New page

== Abstract ==

We propose DOPS, a fast single-stage 3D object detection method for LIDAR data. Previous methods often make domain-specific design decisions, for example projecting points into a bird-eye view image in autonomous driving scenarios. In contrast, we propose a general-purpose method that works on both indoor and outdoor scenes. The core novelty of our method is a fast, single-pass architecture that both detects objects in 3D and estimates their shapes. 3D bounding box parameters are estimated in one pass for every point, aggregated through graph convolutions, and fed into a branch of the network that predicts latent codes representing the shape of each detected object. The latent shape space and shape decoder are learned on a synthetic dataset and then used as supervision for the end-to-end training of the 3D object detection pipeline. Thus our model is able to extract shapes without access to ground-truth shape information in the target dataset. During experiments, we find that our proposed method achieves state-of-the-art results by ~5% on object detection in ScanNet scenes, and it gets top results by 3.4% in the Waymo Open Dataset, while reproducing the shapes of detected cars.

Comment: To appear in CVPR 2020

== Original document ==

The different versions of the original document can be found in:

* [http://arxiv.org/abs/2004.01170 http://arxiv.org/abs/2004.01170]

* [http://arxiv.org/pdf/2004.01170 http://arxiv.org/pdf/2004.01170]

* [http://xplorestaging.ieee.org/ielx7/9142308/9156271/09156417.pdf?arnumber=9156417 http://xplorestaging.ieee.org/ielx7/9142308/9156271/09156417.pdf?arnumber=9156417],
: [http://dx.doi.org/10.1109/cvpr42600.2020.01193 http://dx.doi.org/10.1109/cvpr42600.2020.01193]

* [https://dblp.uni-trier.de/db/conf/cvpr/cvpr2020.html#NajibiLKLRFPRDF20 https://dblp.uni-trier.de/db/conf/cvpr/cvpr2020.html#NajibiLKLRFPRDF20],
: [https://openaccess.thecvf.com/content_CVPR_2020/papers/Najibi_DOPS_Learning_to_Detect_3D_Objects_and_Predict_Their_3D_CVPR_2020_paper.pdf https://openaccess.thecvf.com/content_CVPR_2020/papers/Najibi_DOPS_Learning_to_Detect_3D_Objects_and_Predict_Their_3D_CVPR_2020_paper.pdf],
: [https://openaccess.thecvf.com/content_CVPR_2020/html/Najibi_DOPS_Learning_to_Detect_3D_Objects_and_Predict_Their_3D_CVPR_2020_paper.html https://openaccess.thecvf.com/content_CVPR_2020/html/Najibi_DOPS_Learning_to_Detect_3D_Objects_and_Predict_Their_3D_CVPR_2020_paper.html],
: [https://arxiv.org/abs/2004.01170 https://arxiv.org/abs/2004.01170],
: [https://arxiv.org/pdf/2004.01170 https://arxiv.org/pdf/2004.01170],
: [http://www.arxiv-vanity.com/papers/2004.01170 http://www.arxiv-vanity.com/papers/2004.01170],
: [https://doi.org/10.1109/CVPR42600.2020.01193 https://doi.org/10.1109/CVPR42600.2020.01193],
: [https://academic.microsoft.com/#/detail/3035709245 https://academic.microsoft.com/#/detail/3035709245]

← Older revision	Revision as of 15:25, 3 February 2021
(No difference)

Ross et al 2020a - Revision history

Scipediacontent: Scipediacontent moved page Draft Content 759917905 to Ross et al 2020a

Scipediacontent: Created page with " == Abstract == We propose DOPS, a fast single-stage 3D object detection method for LIDAR data. Previous methods often make domain-specific design decisions, for example proj..."