English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

The inherent ambiguity in ground-truth annotations of 3D bounding boxes,caused by occlusions, signal missing, or manual annotation errors, can confusedeep 3D object detectors during training, thus deteriorating detectionaccuracy. However, existing methods overlook such issues to some extent andtreat the labels as deterministic. In this paper, we formulate the labeluncertainty problem as the diversity of potentially plausible bounding boxes ofobjects. Then, we propose GLENet, a generative framework adapted fromconditional variational autoencoders, to model the one-to-many relationshipbetween a typical 3D object and its potential ground-truth bounding boxes withlatent variables. The label uncertainty generated by GLENet is a plug-and-playmodule and can be conveniently integrated into existing deep 3D detectors tobuild probabilistic detectors and supervise the learning of the localizationuncertainty. Besides, we propose an uncertainty-aware quality estimatorarchitecture in probabilistic detectors to guide the training of the IoU-branchwith predicted localization uncertainty. We incorporate the proposed methodsinto various popular base 3D detectors and demonstrate significant andconsistent performance gains on both KITTI and Waymo benchmark datasets.Especially, the proposed GLENet-VR outperforms all published LiDAR-basedapproaches by a large margin and achieves the top rank among single-modalmethods on the challenging KITTI test set. The source code and pre-trainedmodels are publicly available at \url{https://github.com/Eaphan/GLENet}.

GLENet: Boosting 3D Object Detectors with Generative Label Uncertainty Estimation