Research map / Vision-language and multimodal / Human and embodied modeling
Diverse Image Annotation
Check 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 |
|---|---|---|---|
| Use semantic-path-constrained DPP sampling rather than top-k quality ranking (ablation) | DPP-S · ESP Game, 3 tags | F1-sp: 29.34 -> 33.43 | Helped |
| Use semantic-path-constrained DPP sampling rather than top-k quality ranking (ablation) | DPP-S · IAPRTC-12, 3 tags | F1-sp: 25.32 -> 30.13 | Helped |
| Use semantic-path-constrained DPP sampling rather than top-k quality ranking (ablation) | DPP-S · ESP Game, 5 tags | F1-sp: 37.04 -> 35.96 | Hurt |
| Use semantic-path-constrained DPP sampling rather than top-k quality ranking (ablation) | DPP-S · IAPRTC-12, 5 tags | F1-sp: 34.56 -> 34.23 | Hurt |
| Use the tag similarity matrix S rather than identity during learning (ablation) | DPP top-k · ESP Game, 3 tags | F1-sp: 29.02 -> 29.34 | Helped |
| Use the tag similarity matrix S rather than identity during learning (ablation) | DPP top-k · ESP Game, 5 tags | F1-sp: 36.77 -> 37.04 | Helped |
| Use the tag similarity matrix S rather than identity during learning (ablation) | DPP top-k · IAPRTC-12, 3 tags | F1-sp: 25.3 -> 25.32 | Helped |
| Use the tag similarity matrix S rather than identity during learning (ablation) | DPP top-k · IAPRTC-12, 5 tags | F1-sp: 34.35 -> 34.56 | Helped |
| Evaluate tag subsets with semantic-path F1-sp rather than conventional F1 | DIA annotation evaluation · ESP Game, 3 tags | consistency with human judgments: 31.81% -> 68.19% | Helped |
| Evaluate tag subsets with semantic-path F1-sp rather than conventional F1 | DIA annotation evaluation · IAPRTC-12, 3 tags | consistency with human judgments: 19.59% -> 64.62% | Helped |
| Use DPP-S-sampling instead of ML-MG | image annotation · ESP Game, 3 tags | human preference wins: 250 vs 187 | Helped |
| Use DPP-S-sampling instead of DPP-S-topk | image annotation · ESP Game, 3 tags | human preference wins: 445 vs 47 | Helped |
Closest papers by meaning
- Tagging Like Humans: Diverse and Distinct Image Annotation (CVPR 2018 · similarity 0.92)
- Joint Multi-View Representation Learning and Image Tagging (AAAI 2016 · similarity 0.89)
- Separating Self-Expression and Visual Content in Hashtag Supervision (CVPR 2018 · similarity 0.88)
- Doomed to Re-Annotate, Forever: The ImageNet Story (NeurIPS 2026 · similarity 0.85)
- MOFI: Learning Image Representations from Noisy Entity Annotated Images (ICLR 2024 · similarity 0.84)
- Image Generation Diversity Issues and How to Tame Them (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.