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

Multi-view radar-camera fused 3D object detection provides a fartherdetection range and more helpful features for autonomous driving, especiallyunder adverse weather. The current radar-camera fusion methods deliver kinds ofdesigns to fuse radar information with camera data. However, these fusionapproaches usually adopt the straightforward concatenation operation betweenmulti-modal features, which ignores the semantic alignment with radar featuresand sufficient correlations across modals. In this paper, we present MVFusion,a novel Multi-View radar-camera Fusion method to achieve semantic-aligned radarfeatures and enhance the cross-modal information interaction. To achieve so, weinject the semantic alignment into the radar features via the semantic-alignedradar encoder (SARE) to produce image-guided radar features. Then, we proposethe radar-guided fusion transformer (RGFT) to fuse our radar and image featuresto strengthen the two modals' correlation from the global scope via thecross-attention mechanism. Extensive experiments show that MVFusion achievesstate-of-the-art performance (51.7% NDS and 45.3% mAP) on the nuScenes dataset.We shall release our code and trained networks upon publication.

MVFusion: Multi-View 3D Object Detection with Semantic-aligned Radar and Camera Fusion