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

Object detection on Lidar point cloud data is a promising technology forautonomous driving and robotics which has seen a significant rise inperformance and accuracy during recent years. Particularly uncertaintyestimation is a crucial component for down-stream tasks and deep neuralnetworks remain error-prone even for predictions with high confidence.Previously proposed methods for quantifying prediction uncertainty tend toalter the training scheme of the detector or rely on prediction sampling whichresults in vastly increased inference time. In order to address these twoissues, we propose LidarMetaDetect (LMD), a light-weight post-processing schemefor prediction quality estimation. Our method can easily be added to anypre-trained Lidar object detector without altering anything about the basemodel and is purely based on post-processing, therefore, only leading to anegligible computational overhead. Our experiments show a significant increaseof statistical reliability in separating true from false predictions. Wepropose and evaluate an additional application of our method leading to thedetection of annotation errors. Explicit samples and a conservative count ofannotation error proposals indicates the viability of our method forlarge-scale datasets like KITTI and nuScenes. On the widely-used nuScenes testdataset, 43 out of the top 100 proposals of our method indicate, in fact,erroneous annotations.

LMD: Light-weight Prediction Quality Estimation for Object Detection in Lidar Point Clouds