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 Plus annual

Presents the access of premium content as premium feature

Premium Content

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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

We propose a novel semi-supervised active learning (SSAL) framework formonocular 3D object detection with LiDAR guidance (MonoLiG), which leveragesall modalities of collected data during model development. We utilize LiDAR toguide the data selection and training of monocular 3D detectors withoutintroducing any overhead in the inference phase. During training, we leveragethe LiDAR teacher, monocular student cross-modal framework from semi-supervisedlearning to distill information from unlabeled data as pseudo-labels. To handlethe differences in sensor characteristics, we propose a data noise-basedweighting mechanism to reduce the effect of propagating noise from LiDARmodality to monocular. For selecting which samples to label to improve themodel performance, we propose a sensor consistency-based selection score thatis also coherent with the training objective. Extensive experimental results onKITTI and Waymo datasets verify the effectiveness of our proposed framework. Inparticular, our selection strategy consistently outperforms state-of-the-artactive learning baselines, yielding up to 17% better saving rate in labelingcosts. Our training strategy attains the top place in KITTI 3D andbirds-eye-view (BEV) monocular object detection official benchmarks byimproving the BEV Average Precision (AP) by 2.02.

Monocular 3D Object Detection with LiDAR Guided Semi Supervised Active Learning