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

Vehicle-to-Everything (V2X) collaborative perception has recently gainedsignificant attention due to its capability to enhance scene understanding byintegrating information from various agents, e.g., vehicles, andinfrastructure. However, current works often treat the information from eachagent equally, ignoring the inherent domain gap caused by the utilization ofdifferent LiDAR sensors of each agent, thus leading to suboptimal performance.In this paper, we propose DI-V2X, that aims to learn Domain-Invariantrepresentations through a new distillation framework to mitigate the domaindiscrepancy in the context of V2X 3D object detection. DI-V2X comprises threeessential components: a domain-mixing instance augmentation (DMA) module, aprogressive domain-invariant distillation (PDD) module, and a domain-adaptivefusion (DAF) module. Specifically, DMA builds a domain-mixing 3D instance bankfor the teacher and student models during training, resulting in aligned datarepresentation. Next, PDD encourages the student models from different domainsto gradually learn a domain-invariant feature representation towards theteacher, where the overlapping regions between agents are employed as guidanceto facilitate the distillation process. Furthermore, DAF closes the domain gapbetween the students by incorporating calibration-aware domain-adaptiveattention. Extensive experiments on the challenging DAIR-V2X and V2XSetbenchmark datasets demonstrate DI-V2X achieves remarkable performance,outperforming all the previous V2X models. Code is available athttps://github.com/Serenos/DI-V2X

DI-V2X: Learning Domain-Invariant Representation for Vehicle-Infrastructure Collaborative 3D Object Detection