AI Paper Atlas

Research map / Point clouds / Point cloud classification

3D Single-object Tracking in Point Clouds with High Temporal Variation

Qiao Wu, Kun Sun, Pei An, Mathieu Salzmann, Yanning Zhang, Jiaqi Yang

ECCV 2024 · Point clouds › Point cloud classification · approach: Point cloud detection and tracking · 10 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add observation angle memory bank (OM) to relative-pose-aware memory (ablation)HVTrack · KITTI-HV 5 intervals (mean)Success: 45.8 -> 46.6Improved
add Base-Expansion cross-attention (BEA) instead of vanilla cross-attention (ablation)HVTrack · KITTI-HV 5 intervals (mean)Success: 46.0 -> 46.6Improved
add Contextual Point Guided Self-Attention (CPA) instead of vanilla self-attention (ablation)HVTrack · KITTI-HV 5 intervals (mean)Success: 45.1 -> 46.6Improved
add CPA on large objects (Van) (ablation)HVTrack · KITTI-HV 5 intervals (Van)Success: 28.1 -> 28.7 Success; Van precision 33.5 -> 32.4Mixed
use only expansion branch vs only base branch in BEA (ablation)HVTrack · KITTI-HV 5 intervals (mean)Success: 46.0 -> 46.4 mean; Car 60.3 -> 60.0Mixed
increase test-time memory size from 1 to 6 frames (ablation)HVTrack · KITTI-HV 5 intervals (mean)Success: 43.6 -> 46.6Improved
increase memory size beyond 6 frames (ablation)HVTrack · KITTI-HV 5 intervals (mean)Success: 46.6 -> 46.3Hurt
replace DGCNN backbone with PointNet++ (ablation)HVTrack · KITTI-HV 5 intervals (mean)Success: 46.6 -> 47.3Improved
use HVTrack instead of CXTrack with same backbone and RPNKITTI-HV 5 intervals · KITTI-HV 5 intervals (mean)Success: 35.3 -> 46.6Improved
enlarge search area offset for baselines at 5 frame intervals (ablation)BAT · KITTI-HV 5 intervals (Car)Success: 16.62 (20 m) -> 44.13 (4 m)Mixed
apply HVTrack on regular KITTI vs CXTrackHVTrack · KITTI (regular, mean)Success: 67.5 -> 65.5Hurt
train HVTrack on intervals 1,2,3,5,10 together (ablation)HVTrack · KITTI-HV Car, test interval 5Success: 21.2 -> 63.9Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy