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

In recent years, much progress has been made in LiDAR-based 3D objectdetection mainly due to advances in detector architecture designs andavailability of large-scale LiDAR datasets. Existing 3D object detectors tendto perform well on the point cloud regions closer to the LiDAR sensor asopposed to on regions that are farther away. In this paper, we investigate thisproblem from the data perspective instead of detector architecture design. Weobserve that there is a learning bias in detection models towards the denseobjects near the sensor and show that the detection performance can be improvedby simply manipulating the input point cloud density at different distanceranges without modifying the detector architecture and without dataaugmentation. We propose a model-free point cloud density adjustmentpre-processing mechanism that uses iterative MCMC optimization to estimateoptimal parameters for altering the point density at different distance ranges.We conduct experiments using four state-of-the-art LiDAR 3D object detectors ontwo public LiDAR datasets, namely Waymo and ONCE. Our results demonstrate thatour range-based point cloud density manipulation technique can improve theperformance of the existing detectors, which in turn could potentially inspirefuture detector designs.

Improving LiDAR 3D Object Detection via Range-based Point Cloud Density Optimization