Research map / Detection and segmentation / Open-set and few-shot detection
Inference Fusion with Associative Semantics for Unseen Object Detection
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 target-centric multi-associative semantic transfer to the single-association model (ablation) | DPIF on MSCOCO · MSCOCO ZSD | mAP: 17.39 -> 19.82 | Helped |
| Add target-centric multi-associative semantic transfer to the single-association model (ablation) | DPIF on MSCOCO · MSCOCO ZSD | recall@100: 58.19 -> 55.73 | Hurt |
| Use two-stage training instead of end-to-end training (ablation) | DPIF-S on MSCOCO · MSCOCO ZSD | mAP: 14.87 -> 17.39 | Helped |
| Use two-stage training instead of end-to-end training (ablation) | DPIF-S on MSCOCO · MSCOCO ZSD | recall@100: 55.67 -> 58.19 | Helped |
| Replace k-means concept association with target-centric multi-association (ablation) | DPIF on MSCOCO · MSCOCO ZSD | mAP: 16.58 -> 19.82 | Helped |
| Replace target-centric single association with target-centric multi-association (ablation) | DPIF on MSCOCO · MSCOCO ZSD | mAP: 17.39 -> 19.82 | Helped |
| Increase association top-K from 1 to 5 (ablation) | DPIF-M on MSCOCO · MSCOCO ZSD | mAP: 17.15 -> 19.82 | Helped |
| Increase association top-K from 5 to 10 (ablation) | DPIF-M on MSCOCO · MSCOCO ZSD | mAP: 19.82 -> 19.24 | Hurt |
| Add associative semantic transfer and fuse its predictions with visual-semantic transfer (ablation) | DPIF on MSCOCO · MSCOCO ZSD | mAP: 12.40 -> 19.82 | Helped |
| Add associative semantic transfer and fuse its predictions with visual-semantic transfer (ablation) | DPIF on MSCOCO · MSCOCO ZSD | recall@100: 43.56 -> 55.73 | Helped |
| Use DPIF-M rather than the transductive TL-ZSD baseline | MSCOCO ZSD · MSCOCO ZSD | mAP: 48.15; 55.73 | Helped |
| Use DPIF instead of VSA-ZSD | Pascal VOC GZSD · Pascal VOC GZSD | mAPt: 62.1; 62.26 | Mixed |
Closest papers by meaning
- Hybrid Knowledge Routed Modules for Large-scale Object Detection (NeurIPS 2018 · similarity 0.88)
- Semantic-Aligned Fusion Transformer for One-Shot Object Detection (CVPR 2022 · similarity 0.88)
- Reasoning-RCNN: Unifying Adaptive Global Reasoning Into Large-Scale Object Detection (CVPR 2019 · similarity 0.88)
- Heuristic-inspired Reasoning Priors Facilitate Data-Efficient Referring Object Detection (CVPR 2026 · similarity 0.88)
- Weakly Supervised Knowledge Transfer with Probabilistic Logical Reasoning for Object Detection (ICLR 2023 · similarity 0.87)
- SEEN-DA: SEmantic ENtropy guided Domain-aware Attention for Domain Adaptive Object Detection (CVPR 2025 · similarity 0.86)
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.