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

3D single object tracking with LiDAR points is an important task in thecomputer vision field. Previous methods usually adopt the matching-based ormotion-centric paradigms to estimate the current target status. However, theformer is sensitive to the similar distractors and the sparseness of pointcloud due to relying on appearance matching, while the latter usually focuseson short-term motion clues (eg. two frames) and ignores the long-term motionpattern of target. To address these issues, we propose a mixed paradigm withtwo stages, named MTM-Tracker, which combines motion modeling with featurematching into a single network. Specifically, in the first stage, we exploitthe continuous historical boxes as motion prior and propose an encoder-decoderstructure to locate target coarsely. Then, in the second stage, we introduce afeature interaction module to extract motion-aware features from consecutivepoint clouds and match them to refine target movement as well as regress othertarget states. Extensive experiments validate that our paradigm achievescompetitive performance on large-scale datasets (70.9% in KITTI and 51.70% inNuScenes). The code will be open soon athttps://github.com/LeoZhiheng/MTM-Tracker.git.

Motion-to-Matching: A Mixed Paradigm for 3D Single Object Tracking