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

LiDAR-based 3D object detection is an indispensable task in advancedautonomous driving systems. Though impressive detection results have beenachieved by superior 3D detectors, they suffer from significant performancedegeneration when facing unseen domains, such as different LiDARconfigurations, different cities, and weather conditions. The mainstreamapproaches tend to solve these challenges by leveraging unsupervised domainadaptation (UDA) techniques. However, these UDA solutions just yieldunsatisfactory 3D detection results when there is a severe domain shift, e.g.,from Waymo (64-beam) to nuScenes (32-beam). To address this, we present a novelSemi-Supervised Domain Adaptation method for 3D object detection (SSDA3D),where only a few labeled target data is available, yet can significantlyimprove the adaptation performance. In particular, our SSDA3D includes anInter-domain Adaptation stage and an Intra-domain Generalization stage. In thefirst stage, an Inter-domain Point-CutMix module is presented to efficientlyalign the point cloud distribution across domains. The Point-CutMix generatesmixed samples of an intermediate domain, thus encouraging to learndomain-invariant knowledge. Then, in the second stage, we further enhance themodel for better generalization on the unlabeled target set. This is achievedby exploring Intra-domain Point-MixUp in semi-supervised learning, whichessentially regularizes the pseudo label distribution. Experiments from Waymoto nuScenes show that, with only 10% labeled target data, our SSDA3D cansurpass the fully-supervised oracle model with 100% target label. Our code isavailable at https://github.com/yinjunbo/SSDA3D.

SSDA3D: Semi-supervised Domain Adaptation for 3D Object Detection from Point Cloud