Research map / Detection and segmentation / 3D detection and driving
Unlocking 3D Affordance Segmentation with 2D Semantic Knowledge
Yu Huang, Zelin Peng, Changsong Wen, Xiaokang Yang, Wei Shen
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add affinity alignment loss (Laff) to reconstruction-only pretraining (ablation) | CMAT pretraining · PIADv2 (Seen) | aIoU: 39.27 -> 44.13 | Improved |
| add feature diversity loss (KoLeo) to Lrec + Laff (ablation) | CMAT pretraining · PIADv2 (Seen) | aIoU: 44.13 -> 44.88 | Improved |
| use DINOv3 teacher instead of DINOv2 (ablation) | CMAT pretraining · PIADv2 (Seen) | aIoU: 43.26 -> 44.88 | Improved |
| use CMAT-pretrained backbone instead of standard PointMAE weights (ablation) | Lightweight affordance segmentor · PIADv2 (Seen) | aIoU: 38.16 -> 44.88 | Improved |
| use CMAT-pretrained backbone instead of standard PointMAE weights (ablation) | Lightweight affordance segmentor · PIADv2 (Unseen) | aIoU: 22.07 -> 27.40 | Improved |
| use CMAT-pretrained backbone instead of standard PointMAE weights (ablation) | Lightweight affordance segmentor · PIAD (Unseen) | aIoU: 7.8 -> 8.4 | Improved |
| use CMAT-pretrained backbone instead of standard PointMAE weights (ablation) | Lightweight affordance segmentor · PIAD (Seen) | aIoU: 19.9 -> 21.4 | Improved |
| use CMAT-pretrained backbone instead of standard PointMAE weights (ablation) | Lightweight affordance segmentor · PIADv2 (Unseen) | MAE: 0.142 -> 0.127 | Improved |
| use proposed method vs best prior (GREAT) | affordance segmentation · PIADv2 (Seen) | MAE: 0.067 -> 0.071 | Hurt |
| use proposed method vs LASO | text-prompted affordance segmentation · LASO dataset (Seen) | MAE: 0.093 -> 0.096 | Mixed |
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.
- GREAT — GREAT: Geometry-Intention Collaborative Inference for Open-Vocabulary 3D Object Affordance Grounding (CVPR 2025) · better on 2 of 3 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
- 3D-AffordanceLLM: Harnessing Large Language Models for Open-Vocabulary Affordance Detection in 3D Worlds (ICLR 2025 · similarity 0.90)
- SeqAfford: Sequential 3D Affordance Reasoning via Multimodal Large Language Model (CVPR 2025 · similarity 0.90)
- UniFunc3D: Unified Active Spatial-Temporal Grounding for 3D Affordance Segmentation (NeurIPS 2026 · similarity 0.90)
- GEAL: Generalizable 3D Affordance Learning with Cross-Modal Consistency (CVPR 2025 · similarity 0.90)
- LASO: Language-guided Affordance Segmentation on 3D Object (CVPR 2024 · similarity 0.88)
- Learning 2D Invariant Affordance Knowledge for 3D Affordance Grounding (AAAI 2025 · similarity 0.88)
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.
