Research map / Detection and segmentation / CNN features
PhaseWin Search Framework Enable Efficient Object-Level Interpretation
Zihan Gu, Ruoyu Chen, Junchi Zhang, Yue Hu, Hua Zhang, Xiaochun Cao
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 |
|---|---|---|---|
| replace greedy VPS-50 search with PhaseWin-50 | VPS attribution (Grounding DINO) · MS COCO detection | MEC_ave: 2548.8 -> 536.8 | Improved |
| replace greedy VPS-50 search with PhaseWin-50 | VPS attribution (Grounding DINO) · MS COCO detection | insertion AUC: 0.5195 -> 0.4785 | Hurt |
| replace greedy VPS-50 search with PhaseWin-50 | VPS attribution (Grounding DINO) · MS COCO detection | A-C ratio: 2.04 -> 8.92 | Improved |
| replace greedy VPS-50 with PhaseWin-50 | VPS attribution (Grounding DINO) · RefCOCO REC | MEC_ave: 2290.6 -> 630.1 | Improved |
| replace greedy VPS-50 with PhaseWin-50 | VPS attribution (Grounding DINO) · LVIS V1 rare | MEC_ave: 2544.6 -> 465.9 | Improved |
| replace greedy VPS with PhaseWin on Florence-2 | VPS attribution (Florence-2) · MS COCO detection | insertion AUC: 0.7678 -> 0.7615 | No clear effect |
| replace greedy VPS with PhaseWin on Florence-2 | VPS attribution (Florence-2) · RefCOCO REC | insertion AUC: 0.8301 -> 0.8312 | No clear effect |
| replace greedy VPS with PhaseWin on Florence-2 | VPS attribution (Florence-2) · MS COCO detection | A-C ratio: 2.98 -> 3.49 | Improved |
| replace greedy VPS-100 with PhaseWin-100 on failure cases | VPS attribution (Grounding DINO) · RefCOCO incorrect grounding | insertion AUC: 0.4981 -> 0.5047 | No clear effect |
| replace greedy VPS-100 with PhaseWin-100 | VPS attribution (Grounding DINO) · MS COCO misclassified | insertion AUC: 0.3357 -> 0.3018 | Hurt |
| replace greedy VPS-100 with PhaseWin-100 | VPS attribution (Grounding DINO) · MS COCO undetected | insertion AUC: 0.2102 -> 0.2156 | No clear effect |
| replace greedy VPS-100 with PhaseWin-100 | VPS attribution (Grounding DINO) · MS COCO undetected | MEC_ave: 10100 -> 2160.2 | Improved |
Closest papers by meaning
- Interpreting Object-level Foundation Models via Visual Precision Search (CVPR 2025 · similarity 0.89)
- Less is More: Fewer Interpretable Region via Submodular Subset Selection (ICLR 2024 · similarity 0.86)
- Object-Centric Pretraining via Target Encoder Bootstrapping (ICLR 2025 · similarity 0.85)
- M^3-VOS: Multi-Phase, Multi-Transition, and Multi-Scenery Video Object Segmentation (CVPR 2025 · similarity 0.84)
- BOOTPLACE: Bootstrapped Object Placement with Detection Transformers (CVPR 2025 · similarity 0.84)
- Cross-Modal Distillation for 2D/3D Multi-Object Discovery from 2D Motion (CVPR 2025 · similarity 0.84)
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.
