AI Paper Atlas

Research map / Detection and segmentation / 3D detection and driving

FSD V2: Improving Fully Sparse 3D Object Detection With Virtual Voxels

Lue Fan, Wang Feng, Naiyan Wang, Zhaoxiang Zhang

TPAMI 2024 · Detection and segmentation › 3D detection and driving · approach: Depth-based and label-free 3D detection · 62 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 virtual voxel mixer on real voxels (ablation)FSDv2-Baseline · nuScenes valmAP: 60.9 -> 62.2Improved
add virtual voxelization (virtual voxels from voted centers) (ablation)FSDv2 + Mixer · nuScenes valmAP: 62.2 -> 62.7Improved
use weighted centroids for virtual voxel assignment instead of geometric centers (ablation)FSDv2 · Argoverse 2 valmAP: 34.3 -> 36.7Improved
use weighted centroids for assignment instead of geometric centers (ablation)FSDv2 · Argoverse 2 valBollard AP: 34.1 -> 51.4Improved
add virtual voxel encoder (VVE) (ablation)FSDv2 · nuScenes valmAP: 62.1 -> 62.7Improved
add real voxels as mixer input (ablation)FSDv2 · nuScenes valmAP: 60.6 -> 62.0Improved
use multi-scale real voxels as mixer input (ablation)FSDv2 · nuScenes valmAP: 62.0 -> 62.7Improved
scale center voting offsets from 0 to 1 (full virtual voxel effect) (ablation)FSDv2 · nuScenes valBarrier AP: 66.8 -> 72.7Improved
treat only nearest virtual voxel as positive instead of all in box (ablation)FSDv2 · Argoverse 2 valmAP: 36.2 -> 35.0Hurt
enlarge GT boxes by 0.5m for assignment (ablation)FSDv2 w/o VA · Argoverse 2 valmAP: 34.3 -> 33.5Hurt
use weighted centroid with alpha=0.0 (ablation)FSDv2 · Argoverse 2 valmAP: 36.2 -> 36.9Improved
vary virtual voxel size from (0.4,0.4,0.4) to (0.3,0.3,0.4) (ablation)FSDv2 · nuScenes valmAP: 62.0 -> 61.7No clear effect

Later papers that compared with this method

2 later accepted papers report FSDv2 as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 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