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

In this paper, we propose a new paradigm, named Historical Object Prediction(HoP) for multi-view 3D detection to leverage temporal information moreeffectively. The HoP approach is straightforward: given the current timestampt, we generate a pseudo Bird's-Eye View (BEV) feature of timestamp t-k from itsadjacent frames and utilize this feature to predict the object set at timestampt-k. Our approach is motivated by the observation that enforcing the detectorto capture both the spatial location and temporal motion of objects occurringat historical timestamps can lead to more accurate BEV feature learning. First,we elaborately design short-term and long-term temporal decoders, which cangenerate the pseudo BEV feature for timestamp t-k without the involvement ofits corresponding camera images. Second, an additional object decoder isflexibly attached to predict the object targets using the generated pseudo BEVfeature. Note that we only perform HoP during training, thus the proposedmethod does not introduce extra overheads during inference. As a plug-and-playapproach, HoP can be easily incorporated into state-of-the-art BEV detectionframeworks, including BEVFormer and BEVDet series. Furthermore, the auxiliaryHoP approach is complementary to prevalent temporal modeling methods, leadingto significant performance gains. Extensive experiments are conducted toevaluate the effectiveness of the proposed HoP on the nuScenes dataset. Wechoose the representative methods, including BEVFormer and BEVDet4D-Depth toevaluate our method. Surprisingly, HoP achieves 68.5% NDS and 62.4% mAP withViT-L on nuScenes test, outperforming all the 3D object detectors on theleaderboard. Codes will be available at https://github.com/Sense-X/HoP.

Temporal Enhanced Training of Multi-view 3D Object Detector via Historical Object Prediction