AI Paper Atlas

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

FoundObj: Self-supervised Foundation Models as Rewards for Label-free 3D Object Segmentation

Zihui Zhang, Zhixuan Sun, Yafei YANG, Jinxi Li, Jiahao Chen, Bo Yang

ICML 2026 · Detection and segmentation › 3D detection and driving · approach: 3D objectness and scene segmentation · 0 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 geometric reward module (ablation)FoundObj · ScanNet valAP: 19.5 -> 24.2Improved
add semantic reward module (ablation)FoundObj · ScanNet valAP: 15.3 -> 24.2Improved
DBSCAN radius 0.05 instead of 0.02 (ablation)FoundObj · ScanNet valAP: 21.9 -> 24.2Improved
DBSCAN radius 0.05 instead of 0.1 (ablation)FoundObj · ScanNet valAP: 22.5 -> 24.2Improved
point ratio 30% instead of 20% (ablation)FoundObj · ScanNet valAP: 22.1 -> 24.2Improved
neighbor threshold 0.1 instead of 0.2 (ablation)FoundObj · ScanNet valAP: 21.4 -> 24.2Improved
cost bank size 20 instead of 10 (ablation)FoundObj · ScanNet valAP: 20.9 -> 24.2Improved
cost bank size 20 instead of 40 (ablation)FoundObj · ScanNet valAP: 21.1 -> 24.2Improved
neighbor threshold 0.1 instead of 0.05 (ablation)FoundObj · ScanNet valAP: 23.0 -> 24.2Improved
use RL agent with TRELLIS+DINO features instead of UnScene3D clustering (ablation)FoundObj · ScanNet val, DINO+TRELLISAP: 15.3 -> 24.2Improved
use RL agent instead of Part2Object on DINO-only features (ablation)FoundObj · ScanNet val, DINO-onlyAP: 19.6 -> 19.5No clear effect
clean pseudo masks using ground-truth IoU filtering (ablation)Mask3D training · ScanNet valAP: 24.2 -> 37.0Improved

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