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

Single Object Tracking in LiDAR point cloud is one of the most essentialparts of environmental perception, in which small objects are inevitable inreal-world scenarios and will bring a significant barrier to the accuratelocation. However, the existing methods concentrate more on exploring universalarchitectures for common categories and overlook the challenges that smallobjects have long been thorny due to the relative deficiency of foregroundpoints and a low tolerance for disturbances. To this end, we propose a Siamesenetwork-based method for small object tracking in the LiDAR point cloud, whichis composed of the target-awareness prototype mining (TAPM) module and theregional grid subdivision (RGS) module. The TAPM module adopts thereconstruction mechanism of the masked decoder to learn the prototype in thefeature space, aiming to highlight the presence of foreground points that willfacilitate the subsequent location of small objects. Through the aboveprototype is capable of accentuating the small object of interest, thepositioning deviation in feature maps still leads to high tracking errors. Toalleviate this issue, the RGS module is proposed to recover the fine-grainedfeatures of the search region based on ViT and pixel shuffle layers. Inaddition, apart from the normal settings, we elaborately design a scalingexperiment to evaluate the robustness of the different trackers on smallobjects. Extensive experiments on KITTI and nuScenes demonstrate that ourmethod can effectively improve the tracking performance of small targetswithout affecting normal-sized objects.

Small Object Tracking in LiDAR Point Cloud: Learning the Target-awareness Prototype and Fine-grained Search Region