AI Paper Atlas

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

Find n' Propagate: Open-Vocabulary 3D Object Detection in Urban Environments

Djamahl Etchegaray, Zi Helen Huang, Tatsuya Harada, Yadan Luo

ECCV 2024 · Detection and segmentation › 3D detection and driving · approach: Depth-based and label-free 3D detection · 18 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 point-dropout density simulator on top of GT sampling (ablation)Find n' Propagate (Transfusion) · nuScenes Setting 1AP_N: 21.14 -> 30.00Improved
add translation noise only to pasted objects (ablation)Geometry simulator · nuScenes Setting 1AP_N: AP_N 21.14 -> 19.41; AR_N 50.59 -> 56.40Mixed
add rotation variance on top of translation variance (and dropout) (ablation)Geometry simulator · nuScenes Setting 1AP_N: 19.41 -> 33.65Improved
combine density ranking with multi-view alignment in the Box Oracle (ablation)Greedy Box Oracle · nuScenes Setting 3AP_N: 15.97 -> 16.72Improved
add density ranking to alignment-only scoring (ablation)Greedy Box Oracle · nuScenes Setting 3AP_N: 15.05 -> 16.72Improved
replace FGR rectangular-prior box regression with Greedy Box Seekerbottom-up proposals · KITTI val pedestrian/cyclistAP_3D: Cyclist Moderate 3.28 -> 8.02; Pedestrian Moderate 4.51 -> 14.29Improved
replace OpenSight proposals with Greedy Box Seeker boxesnovel box proposals · nuScenes Setting 3AP_N: 5.80 -> 16.72Improved
add self-training with Box Seeker pseudo-labels instead of top-down self-trainTop-down Self-train (CLIP, Transfusion) · nuScenes Setting 1AP_N: 6.47 -> 33.65Improved
self-training with GLIP labels on top-down proposalsTop-down Projection (GLIP) · nuScenes Setting 1AP_N: 4.53 -> 3.79Hurt
add self-training to CLIP top-down projectionTop-down Projection (CLIP) · nuScenes Setting 1AP_N: 3.06 -> 6.47Improved
use logit fusion with base detector in top-down projectionTop-down Projection (GLIP, Transfusion) · nuScenes Setting 1mAP: 24.23 -> 30.64Improved
swap Transfusion for Centerpoint detector in top-down projectionTop-down Projection · nuScenes Setting 1mAPHurt

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